(updated 6.1.2023)

6.7.2023 4:52PM - I'm still exploring the new logic: "Fatima only hates Israel, not the Jewish people. Therefore, her speech was not a hate speech." Aristotle told us that if A hates B, not C, A is still a hater. Perhaps a noble hater, but a hater. Beyond me. (reposted:

6.7.2023 10:03am - Who said the conflict is "complex"? It is as clear as the path of the sun across the sky! (reposted:

6.6.2023 11:30pm - (1/2) Everything dwarfs when we think about that momentous endeavor, D-day: 7,000 Allied ships disembarking 130,000 men. I am trying to put myself in the shoes of one of those brave men -- I can't. All I remember was my parents looking over a newspaper with a huge headline: (reposted:
6.6.2023 11:30pm - (2/2) "The Invasion Started". From the looks in their faces I could tell it was something they had expected for a long time, something that would change the course of history -- it did.

6.6.2023 3:56pm - @DovWaxman , I'm just checking Mehdi's logic here: "Hatred against Israel is not the same as hatred against the Jewish people" yet Fatima's speech, in which she ONLY expressed hatred against Israel, was NOT a hate speech. Did I get it right? Sounds like Mehdi Hasan's logic. (reposted:

6.6.2023 3:03pm - (Replying to @AndrewPessin @ActForIsrael and 2 others) Read how Abe Foxman describes the way they got him to agree to participate: "It's about the special relationshipd between American Jews and Israel".

6.6.2023 12:32pm - I truly feel for the Center for Israel Studies at UCLA for being the first in the nation to be duped into this stomach-turning film. Someone should have warned them in time, but who could imagine such shrewd deception. @ActForIsrael , @Ostrov_A @AaronBandler @AndrewPessin (reposted:

6.6.2023 11:59am - (Replying to @ShaharAzani @DavidSuissaJJ and 13 others) Your shock of "a film about Israel featuring the trifecta of Chomsky, Beinart and Cornel West! " is second only to my embarrassment of seeing my campus, UCLA, being the first to fall for this deception. Will our students ever forgive us?

6.6.2023 6:13am - @11Drtlee11 , we need you at UCLA. Our Vice Chancellor of EDI is leaving, and a search committee is currently seeking qualified candidates. You seem able to lift this office to its charter. (reposted:

6.6.2023 2:08am - Two changes are indisputable when speaking about the impacts of the Six-Day war (June 5, 1967) (1) Religious sites became accessible to all. (2) Israel is forced to defend itself from an area it would rather stay out of. (reposted:

6.6.2023 12:45am - What "exogenous source of variation" do we imagine when we say that "the moon causes the tide to rise" ?

6.6.2023 12:22am - (1/2) Join me in welcoming the past week's yield of CI papers: My eyes were attracted to: Park etal, Measure Theoretic Foundations of Causality which foundation-minded readers would find interesting. I'm not convinced though that the
6.6.2023 12:22am - (1/2) weaknesses of SCMs, as noted in this paper (mostly in cyclic systems) justify abandoning their beauty in favor of the Causal Spaces approach. Readers may convince me otherwise.

6.5.2023 5:25pm - Great! Oh, God, If I could only paint an ameba, I would beat him by 3.8 billion years. Or, as Lord Balfour use to say: Sing me a Jebusite song, and I'll give you Jerusalem. (reposted:

6.5.2023 4:11am - I must add that, IMO, anthropomorphic metaphors should be welcome as primitives when we consider reductions. Such metaphors are stored deep in our neural architecture, which assures consistency and transparency in model construction. (reposted:

6.4.2023 11:48pm - Also notable is the @NYT way of describing how poor Fatima "became the subject of nearly nonstop, consistently furious international TABLOID coverage." You see, only "tabloids" would criticize an angel like Fatima, not a serious newspaper like the NYT. (reposted:

6.4.2023 11:30pm - (Replying to @DAGophile @DmitriGallow and @vlad_surdea) 6.6.2023 6:13am - @11Drtlee11 , we need you at UCLA. Our Vice Chancellor of EDI is leaving, and a search committee is currently seeking qualified candidates. You seem able to lift this office to its charter. (reposted:

6.6.2023 5:51am - (Replying to @tkaiser_science and @smueller) Smart move. I do not know of any other approach; the critics are not producing answers to the research question asked: P(benefit)=? because they are scared to use counterfactual notation to express this quantity, which is counterfactual in its very being.

6.6.2023 5:29am - (Replying to @tkaiser_science) Yes. And after discovering that we are "dangerously misguided" if not worse, we have written an appendix to mitigate the charges: My comment on Dawid's DT approach illuminate the controversy: @smueller may have additional comments

6.6.2023 5:04am - (Replying to @Dkarbowitz) A year after "Our New Marranos", Peter Beinart inspired me to write "Jews of Discomfort": A tormented ex-Jew, he is still unable to forgive his mother for making him feel so inadequate. (posted: (posted:

6.6.2023 2:32am - This may seem like other anti-Israel propaganda films we have seen in the past. It is not. It's coming to the Jewish Study Center on your campus. Its Jewish cast gives it the necessary cover to go under the radar, as it did at UCLA, (scheduled for June 7) @AaronBandler (reposted:

6.6.2023 2:08am - Two changes are indisputable when speaking about the impacts of the Six-Day war (June 5, 1967) (1) Religious sites became accessible to all. (2) Israel is forced to defend itself from an area it would rather stay out of. (reposted:

6.6.2023 1:59am - (Replying to @jeffrey_bowers and @People4Bernie) Very difficult, but not impossible, as argued in "The war of Return" (posted:

6.6.2023 1:29am - Some colleagues complained that I shouldn't use the term "so-called Jewish," lest it deepens divisions among Jews. I beg to differ, as explained in the paper "Our New Maranos" I don't believe anti-Zionists should be given equal voice in Jewish education. (reposted:

6.6.2023 12:45am - What "exogenous source of variation" do we imagine when we say that "the moon causes the tide to rise" ?

6.6.2023 12:22am - (1/2) Join me in welcoming the past week's yield of CI papers: My eyes were attracted to: Park etal, Measure Theoretic Foundations of Causality which foundation-minded readers would find interesting. I'm not convinced though that the
6.6.2023 12:22am - (1/2) weaknesses of SCMs, as noted in this paper (mostly in cyclic systems) justify abandoning their beauty in favor of the Causal Spaces approach. Readers may convince me otherwise.

6.5.2023 5:25pm - Great! Oh, God, If I could only paint an ameba, I would beat him by 3.8 billion years. Or, as Lord Balfour use to say: Sing me a Jebusite song, and I'll give you Jerusalem. (reposted: 6.6.2023 12:55am - (Replying to @Dkarbowitz @AaronBandler and @deborahlipstadt) I don't think politics has anything to do with my expectation that @deborahlipstadt sees the drawbacks of the original version of the Initiative and takes corrective measures.

6.6.2023 12:46am - (Replying to @eliasbareinboim @DAGophile and @vlad_surdea) I agree.

6.6.2023 12:45am - What "exogenous source of variation" do we imagine when we say that "the moon causes the tide to rise" ?

6.6.2023 12:35am - I think about it differently. I think it is features 1-5 of DAGs that give people the sensation that they understand the model better. So, 6 would be a corollary.

6.6.2023 12:27am - (Replying to @soboleffspaces) This follows from Validity, which stands for the capacity to look at a formal specification and decide whether it captures your scientific understanding of the problem.

6.6.2023 12:22am - (1/2) Join me in welcoming the past week's yield of CI papers: My eyes were attracted to: Park etal, Measure Theoretic Foundations of Causality which foundation-minded readers would find interesting. I'm not convinced though that the
6.6.2023 12:22am - (1/2) weaknesses of SCMs, as noted in this paper (mostly in cyclic systems) justify abandoning their beauty in favor of the Causal Spaces approach. Readers may convince me otherwise.

6.5.2023 6:01pm - (Replying to @mikopeled) By this stretch of imagination and causal speculation I'd argue that it was actually @mikopeled who "made it possible", by giving unconditional support to Israel enemies and her delegitimizers, thus forcing IDF to continue and defend Israel from areas it'd rather stay out of.

6.5.2023 5:25pm - Great! Oh, God, If I could only paint an ameba, I would beat him by 3.8 billion years. Or, as Lord Balfour use to say: Sing me a Jebusite song, and I'll give you Jerusalem. (reposted:

6.5.2023 4:11am - I must add that, IMO, anthropomorphic metaphors should be welcome as primitives when we consider reductions. Such metaphors are stored deep in our neural architecture, which assures consistency and transparency in model construction. (reposted:

6.5.2023 4:04am - (1/2) DAGs are not sacred, but they are more that just "an alternative way of making assumptions" They provide a way of making assumptions that has 5 advantages, allowing us to: 1. check consistency, 2. check redundancy 3. check completeness, 4. check validity, 5. check testability (reposted:
6.5.2023 4:04am - (2/2) To see how these capabilities compare with with a famous alternative called Potential Outcomes (PO), see:, where a simple problem is solved both ways and compared side by side. No, Imbens is wrong saying that the choice is just a matter of "taste".

6.4.2023 11:48pm - Also notable is the @NYT way of describing how poor Fatima "became the subject of nearly nonstop, consistently furious international TABLOID coverage." You see, only "tabloids" would criticize an angel like Fatima, not a serious newspaper like the NYT. (reposted:

6.4.2023 11:30pm - (Replying to @DAGophile @DmitriGallow and @vlad_surdea) The "listening" metaphor was sprinkled throughout my writings but was explicitly called "society of listening agents" in It was further refined in, after Dawid objected to its anthropomorphic flavor.

6.4.2023 6:00pm - Congratulations, Bernard-Henri Levy @BHL upon receiving the 2023 Daniel Pearl Award for Courage and Integrity in Journalism. A true champion of humanity, your life, writings and journeys truly personify the meaning of the words: "Courage and Integrity in Journalism". (reposted:

6.4.2023 4:56pm - (Replying to @FoxmanAbraham @jonathans_tobin and @AmmiHirsch) And thank you, Abe, for supporting Rabbi Hirsch.

6.4.2023 4:50pm - A timely tweet by Abraham Foxman, former head of @ADL . The so called "Israelism" film is a deceitful propaganda sham, sprinkled by (so called) Jewish actors. Please alert your campus Hillel Director, before he/she is lured into airing it on your campus for 'informed discussion". (reposted:

6.4.2023 4:24pm - (Replying to @Big_Worker and @People4Bernie) Thanks, but don't think I am as naive. I know "real world apartheid" when I see one and, more importantly, I know Israeli society inside out -- they are not lying to me -- 2 State + "end of conflict" solution is still the dream and principle of the vast majority of Israelis.

6.4.2023 4:17pm - (1/2) Two notes are worthy of attention. 1) As stated in my paper, Gibbard and Harper (1978) were the first to derive the consistency rule from Lewis's closest world semantics. (2) Halpern's completeness result guarantees that structural semantics does not (reposted:
6.4.2023 4:17pm - (2/2) produce additional rules, but consistency (and its compositional extension). It is thus comforting to know that the consistency rule is a necessary and sufficient derivative of a friendly formal semantics.

6.4.2023 3:56pm - (Replying to @academic_la) The bond is Zionophobia

6.4.2023 3:07pm - (Replying to @Dkarbowitz @AaronBandler and @deborahlipstadt) Ein Adam Sh'Ein Lo Shaah (every person as an hour of opportunity)

6.4.2023 2:19pm - (Replying to @rebels2008 @hsultan75 and @People4Bernie) The Israeli gov't is wise indeed. It has erected huge antennas tuned into Palestinian classrooms, mosques and media, eager to find an inkling of acceptance -- none thus far, save for surveys conducted for Western consumption.

6.4.2023 2:07pm - A most uplifting photo, to go with the uplifting spirit of the Israel Day Parade (reposted:

6.4.2023 1:52pm - Uplifting to see so many sane people in NYC. Where do they get the strength to stand up for what's right, true and noble, in the face of so much hostility and defamatory disinformation? It seems Zionism has done something right after all - it has given Jews a soul and a spine. (reposted:

6.4.2023 1:33pm - (Replying to @hsultan75 @rebels2008 and @People4Bernie) Beg to differ. When Palestinians say 2-states-solution they mean "A Palestinian state acting as a spring board for continuing the struggle to eliminate Israel." This interpretation should hold until we hear ONE Palestinian teacher, say to ONE class: "Israel is permanent kids".

6.4.2023 11:20am - (Replying to @DAGophile @DmitriGallow and @vlad_surdea) I believe the discussion about "reduction" should start with the question: reduction to what? Rather than with names of philosophers who tried reduction in the past. What kind of judgements we think could serve as primitives, replacing primitives s/a "listens to" (my favorite)

6.4.2023 11:08am - (Replying to @creaso1487 @ahadhaamoratsim and 2 others) Ireland for the Irish? By all means, as long as the Irish recognize 'London for the British', not if they claim London is a stolen land that needs to be "liberated" (by the Irish) tooth and nail.

6.4.2023 10:57am - I can whole heartily recommend at least one of the two books. Didn't read the other yet. (reposted:

6.4.2023 10:49am - Me too. Its amazing that a document of such stature could have emerged under such stormy and chaotic circumstances. (reposted:

6.4.2023 10:44am - (Replying to @TPrincato @Eradicator_NYC and 3 others) Sorry, no foul language

6.4.2023 10:34am - (Replying to @roberto_limongi) Looking forward

6.4.2023 10:30am - (Replying to @lepton939 @GPaulTurner and @People4Bernie) The only solution that I've found to be workable is in @EinatWilf 's The War of Return, precisely because Israelis too are not going away; a point that's not sufficiently emphasized in debates on that 1-state non-solution. The choice is between coexistence or mutual annihilation.

6.4.2023 10:16am - (Replying to @griergray3RD and @People4Bernie) I'd forgive Palestinians for this twist in their genetic making, if they accept that Jews are indigenous in their historical making.

6.4.2023 10:09am - (Replying to @natwitte) Westplaining is today the major obstacle to peace, more even than the settlements.

6.4.2023 3:42am - Based on your Tweet, I viewed to it again, and decided to RT it again because it answers questions that readers raised in the last week (especially the antagonistic ones). How Western Indulgence of the Palestinian Dream Has Obstructed the Path to Peace (reposted:

6.4.2023 3:29am - (Replying to @DAGophile and @vlad_surdea) Don't you think that what lot of earlier philosophers thought to be desirable is plainly impossible?

6.4.2023 2:32am - This may seem like other anti-Israel propaganda films we have seen in the past. It is not. It's coming to the Jewish Study Center on your campus. Its Jewish cast gives it the necessary cover to go under the radar, as it did at UCLA, (scheduled for June 7) @AaronBandler (reposted:

6.4.2023 1:52am - (Replying to @GPaulTurner @lepton939 and @People4Bernie) I dont know where you are getting your information, concluding with "without representatives of the Palestinian people." Please read "The War of Return" and let us know if you still believe in "without representatives of the Palestinian people."

6.4.2023 1:26am - You don't have to speak Hebrew to understand the excitement of the commentators, as Israel U-20 team beats Brazil 3-2 and player Turgeman "Plays Maradona on Argentinian soil" (reposted:

6.4.2023 12:00am - (Replying to @oldyishuvjew @Coffefe123 and 5 others) I am not replying to disingenuous questions, especially those that attempt to ridicule my account of the 1948 war, or from readers using foul language (eg "kick out" "apartheid"). Respectful readers truly curious about my childhood will be answered.

6.3.2023 11:48pm - (Replying to @vlad_surdea and @DAGophile) I cannot help you much on Woodward's counterfactuals because I find his account confused, and often circular. Some of my difficulties are expressed in this reply to his review of Causality

6.3.2023 11:21pm - For readers interested in a glimpse into @EinatWilf "The War of Return", here is a video summarizing the book and its message: For a Hague conference on the legalities surrounding Israel's re-establishment, see Zipperstein's (reposted:

6.3.2023 11:06pm - (Replying to @Eradicator_NYC @TPrincato and 2 others) I am not an expert in international law. What I know comes from other texts. But I would recommend @EinatWilf "The War of Return", especially regarding the opportunities missed to settle the problem, but ruined by Palestinians refusal to deal with a "non-existent entity."

6.3.2023 10:52pm - (Replying to @ChronicBabak @Coffefe123 and 5 others) You know everything, I do not feel comfortable talking to an all-knowing entity. No, the village was Jamusin, and the bombed factory between us was Meshi Zacks. Al Jazeera will probably discover that it was a weapon factory, but I have zero interest in conspiracy theories. Zero.

6.3.2023 9:43pm - (Replying to @TPrincato @Eradicator_NYC and 2 others) There is no "punishment for" here. Right of return is negotiated between the countries involved and, since Haifa was under British mandate, which no longer exists, immigration need to be negotiated between other sovereign countries involved. Ref. The War of Return

6.3.2023 9:36pm - (Replying to @Coffefe123 @oldyishuvjew and 5 others) At age 12, Egyptian war planes bombed both my village and the Arab village next door (there was a textile factory between us) and, the day after, the Arab kids, our playmates, did not show up to the playground.

6.3.2023 9:30pm - (Replying to @TPrincato @Eradicator_NYC and 2 others) Wars change the rules of immigration. Millions of Germans were expelled from territories that became part of Poland. These individuals and their descendants do not have a right to return and reclaim property, based on Polish laws and confirmed by international treaties.

6.3.2023 9:13pm - (Replying to @TPrincato @Eradicator_NYC and 2 others) Prioritizing immigration on the basis of security risks and special needs is not "racist". E.g. U.S. has programs to prioritize refugees from countries such as Syria and Iraq, based on conditions in those countries. Jewish history confers 'special need' status to diaspora Jews.

6.3.2023 8:55pm - (Replying to @rebels2008 and @People4Bernie) This is a hypothesis for which I've found no shred of evidence despite A long desperate search. Sifting though educational material, media content, Mosque sermons etc. led me to conclude that Westerner assumption that "there are plenty of Palestinians" is just plain wrong.

6.3.2023 5:22pm - (Replying to @AriDurresi and @larsweisbrod) I don't think God's friendship will help, unless the "interactions" are causal interaction, which amounts to causal knowledge. Statistical interactions will not do.

6.3.2023 5:19pm - (Replying to @Eve_Barlow) Is she really anti-Zionist? Or just a misguided virtue projector, lured by populist slogans?

6.3.2023 5:01pm - (Replying to @artistexyz @PetrSramekBench and @farzammolkara) Is this available for fun and profit?

6.3.2023 4:57pm - (Replying to @PHuenermund @causalinf and @VC31415) You mean some economists think DAGs, speak DAGs, identify with DAGs, then hide them in the drawer and publish strong ignorabilities? Not surprising, but what about their students? Don't they spy on them?

6.3.2023 2:16pm - Apropos! In my lecture yesterday people asked me what percentage of economists are using DAGs (vis a vis PO). I said 10-20 percent. What's your best estimate? @causalinf @VC31415 They also asked me why so few? But that's a subject of another Tweet. (reposted:

6.3.2023 2:04pm - Wait! All my friends in Israel (and I've many) are willing to commit to the doctrine: "Two states for two peoples, Equally legitimate and Equally indigenous." I know of no Palestinian (and I've read many) capable of saying it. Where does it leave the lost souls of @People4Bernie ? (reposted:

6.3.2023 1:46pm - (Replying to @LeoFergusonnyc) We do it both, by changing the political and material conditions in Palestine/Israel that make the shirt an object of derision, and with moral leadership, declaring the deriding "unacceptable".

6.3.2023 1:22pm - (Replying to @DAGophile and @larsweisbrod) But isn't Woodward's "Making things happen" relying of having a graph telling us who depends on whom which, in essence, amounts to reduction to counterfactuals. ??

6.3.2023 1:17pm - (Replying to @DAGophile @PHuenermund and @larsweisbrod) Why is the attempt to reduce causes to counterfactuals a failure? I see it as a great success.

6.3.2023 1:14pm - (Replying to @larsweisbrod) The emphasis is on "from scratch". This is still true; from passive data alone, lacking experiments or some causal assumptions, we cannot say anything about causation. Have you found an exception?

6.3.2023 1:04pm - (Replying to @jmwooldridge and @PHuenermund) In the language of Potential Outcomes, what was originally meant by "fixed effects", and what is meant now? I've never used it, what have I missed?

6.3.2023 1:01pm - The last supper? Or the one before the storm? So how come people have been warning me on the dangers of AI ? (reposted:

6.3.2023 12:39pm - (Replying to @LeoFergusonnyc) That's precisely why we coined the term Zionophobia, to separate anti-Zionism from antisemitism, to allow people to freely criticize any Govt they want, yet to label them racists when they deny the Jewish people's right to self determination.

6.3.2023 12:32pm - (Replying to @LeoFergusonnyc) If we do not turn a pro-Israel shirt into an object of respect we will become an object of attack when we take off all shirts.

6.3.2023 12:26pm - (Replying to @LeoFergusonnyc) A large part of that "solid ground" is manufactured by Fatima's professors. The belief that Zionism is premised on apartheid amounts to indicting the majority of living Jews for racism, which endangers the safety of all Jews.

6.3.2023 11:30am - (Replying to @LeoFergusonnyc) Watch not the words of Fatima, but the shining faces of her professors, gleaming behind her, as she desecrates the identity of her fellow students and ascend to academic stardom. (reposted: These professors are not 25 yrs old, they know where it leads. (reposted:

6.3.2023 9:35am - Note that I didn't ask @deborahlipstadt whether Fatima's speech was "antisemitic" or not; I asked if it was "acceptable". The term "antisemitic" has lost its punch many years ago and the Nexus "scholars" finished its emasculation. What remains is one issue: Acceptable or not. (reposted:

6.3.2023 9:05am - (Replying to @elderofziyonZ) Would the Biden Initiative define the 1936 Berlin Olympics as antisemitic? Many will argue NO, since it does not meet the criterion of "because of anti-Jewish hate".

6.3.2023 8:54am - (1/2) I am ready to support this historic initiative as soon as Ambassador @deborahlipstadt answers my letter If the Initiative cannot decide a public and well documented case like CUNY, what guidance does it provide to hundreds of other cases? Note also (reposted:
6.3.2023 8:54am - (2/2) also that the word "Zionism" does not appear explicitly in the entire document. (My faithful litmus test for the effectiveness of any attempt to deal with Zionophobic antisemitism.) Why hide it? Why obscure it in debatable interpretations? But let's hear about Fatima first.

6.3.2023 5:39am - (Replying to @tony_sebo @VoiceBoxNoisey and @JGreenblattADL) On the contrary, I support all groups who, like Zionists, are yearning for a homeland. Specifically, the Palestinians, whom Zionists have been approaching (since 1920) with: Two states for two peoples, Equally legitimate and Equally indigenous.

6.3.2023 5:28am - What's happening now at CUNY is a warning for everyone concerned with higher education. This video reveals a pattern of deception used by university administrators to perpetuate discrimination under the cover of "fighting antisemitism" @AdamL @AaronBandler (reposted:

6.3.2023 4:48am - (1/2) Dear Ambassador @deborahlipstadt , the paragraph you retweeted shows genuine intention to deal with a major problem faced by Jewish faculty and students. However, in the abstract, it has been interpreted differently by different people and organizations. To clarify its (reposted:
6.3.2023 4:48am - (2/2) language, can you apply it to the controversial CUNY graduation ceremony and help us determine whether Fatima's speech was acceptable or unacceptable? It is important to show that the Initiative we were all waiting for can make such determinations in concrete painful cases.

6.2.2023 10:05pm - (Replying to @ChronicBabak @Juaney_P and 3 others) Observe your technique. When facts do not fit your narrative you dismiss the sources as (1) Israeli Gov't conspiracy or (2) Pro-Israel advocates. Permit me to terminate this futile discussion, because I trust these 2 sources more that I trust anything I've seen elsewhere. Bye.

6.2.2023 9:46pm - (Replying to @ChronicBabak @Juaney_P and 3 others) I'm aware. It says so in my Hebrew version of 1937: Chapter 9 (titled Arviyey Eretz Israel) translated by Dr. Y. Yavin. But as I explained before, the point of my article was to see what Jabotinsky wanted his Hebrew readers to understand, not the Russians. This is an "error".

6.2.2023 3:54pm - (Replying to @ChronicBabak @Juaney_P and 3 others) Again, you are trying to pull Pallywood tricks on me by repeating a false allegation, repeat, repeat, repeat, as if that would make it true, repeat, repeat ..."your error", "your error" .... repeat, repeat.. Nice try.

6.2.2023 3:41pm - (Replying to @yudapearl @ChronicBabak and 4 others) Here is more information about the term "Pallywood", of which I was a direct victim - a picture of Muhammad Al Dura (the boy allegedly killed by Israeli bullets) was projected in the murder video of my late son Danny. Yes, Pallywood kills. (posted:

6.2.2023 1:52pm - (Replying to @ChronicBabak @Juaney_P and 3 others) You have explained the Palestinian perspective very clearly: co-existence is necessarily "colonization and apartheid" hence unacceptable, period. I've amplified Israel's 'agency' by highlighting her decision to control the WB in hope of impeding Palestinian drive to destroy her.

6.2.2023 1:32pm - (Replying to @joftius and @quantadan) For the life of me, I still can't grasp why anyone in the top photograph would be offended when warned that their march is supporting people totally committed to the destruction of the one ME country in which they are protected and respected. Please educate me.

6.2.2023 1:19pm - (Replying to @ChronicBabak @Juaney_P and 3 others) Pallywood is not "anti-Arab", it is a term coined to describe Palestinian techniques of spreading disinformation through fake news and media manipulation'_video One technique is to repeat a false allegation n times (eg, "your error") to make it believable. (posted:'_video)

6.2.2023 12:53pm - (Replying to @ChronicBabak @Juaney_P and 3 othersZ) Is it so hard to understand that Palestinian commitment to Israel's destruction (regardless of what she does) makes it impossible for Israel to leave the WB? At least we got you to admit that, for Palestinians, coexistence is "being colonized" hence unthinkable at all costs.

6.2.2023 5:06am - (Replying to @NinaTannenwald) Good point. One difference. Communism had a chance to be implemented and show its merits and faults. Zionism has not been given that chance. It was attacked at its bud and prevented from ever developing under normal conditions for us to judge its merits and faults.

6.2.2023 1:59am - (Replying to @ChronicBabak @Juaney_P and 3 others) I would put it a bit differently: The idea that Zionism must be crushed despite its moral and historical imperatives explains the status quo. Glad we are converging, albeit from different directions. Soon we would be neighbors.

6.2.2023 1:44am - (Replying to @BryonActon and @RepMikeLawler) True. Universities know that Qatari money will easily make up for any loss of Federal funds. Still, the Bill will have a symbolic value, and the legal fights over the issues involved may wake up parents and tax payers to look into the subversion of American education.

6.2.2023 1:35am - Today I'll have the honor of speaking at the workshop honoring my friend, colleague, and co-author Joe Halpern. See My topic: Axiomatizing Causal Reasoning My Conclusion: Friendly models (SCM) can be as expressive and as safe as unfriendly ones (eg., OP). (posted:

6.2.2023 1:23am - (Replying to @EmamiDrJ @VictorMuslin and 4 others) You are right. They act as if CAIR is the equivalent of @ADL , and they don't even comment on its presence at the table.

6.2.2023 12:59am - (Replying to @ChronicBabak @Juaney_P and 3 others) There isn't a single "error" in my article. (Please don't try Pallywood techniques on me). I have explained clearly where my teachers got their attitude toward Arabs, and the courage to preach relentlessly: "Today our enemies, tomorrow our neighbors."

6.2.2023 12:40am - (Replying to @ChronicBabak @Juaney_P and 3 others) What's your point? I quoted from the Hebrew book, as stated, and for a good reason. It was the Hebrew version that was meant to shape public opinion, educational molds and the attitude of my generation towards Arabs: "Today our enemies, tomorrow our neighbors". Viva the latter.

6.1.2023 11:38pm - (Replying to @ChronicBabak @Juaney_P and 3 others) I have the first edition of Jabotinsky's Medina Ivrit now in my hand, printed 1937, and I know some Hebrew. On page 69, I find the article Arviyey Eretz Israel, from which I quoted, to the best of my ability. Is that what you call "misrepresentation"? I would call it "FACT".

6.1.2023 11:03pm - ‘There Need to be Consequences’: Rep. Mike Lawler Introduces Bill Cutting Funding for Colleges that Host Antisemitic Events (posted:
6.1.2023 11:03pm - I support this bill, but it has be followed by an official apology to clarify what specifically was hateful in the speech, why, and who needs to be decriminalized.

6.1.2023 10:25pm - (Replying to @VoiceBoxNoisey and @JGreenblattADL) My long experience with dialogue has taught me that, unfortunately, one side considers "inclusivity" to be a disaster, if not a national disgrace.

6.1.2023 10:15pm - (Replying to @ChronicBabak @VoiceBoxNoisey and @JGreenblattADL) The status quo and its illiberal manifestations were created by attempts to crush a Jewish homeland, not by a quest for one.

6.1.2023 10:13pm - (Replying to @ChronicBabak @Juaney_P and 3 others) To learn about early Zionist and how they treated Arabs, read and don't tell me it's not representative, I've played with Arab kids till the age of 12. (posted:

6.1.2023 9:54pm - (Replying to @mapto @VoiceBoxNoisey and @JGreenblattADL) I hope you are not equating Fatima, CAIR and @RashidaTlaib with "moderate voices". I consider myself a "moderate voice" that is stifled by the extreme rhetoric of those three, who are labeling me "white supremacist". No, these kind of voices are not adding to the discussion.

6.1.2023 9:40pm - I am glad to bring this paper to the attention of our readers -- I am still digesting its new aspects and their potential merits. (reposted:

6.1.2023 9:17pm - (Replying to @anaghadoesmath and @vishnusreekr) I would add to these The Primer and how to teach from it

6.1.2023 9:09pm - (Replying to @VoiceBoxNoisey and @JGreenblattADL) Open dialogue only helps if you agree on the principles guiding the dialogue. I would not agree to a dialogue in which the legitimacy of Israel is a subject of discussion while that of Palestinian statehood is taken as a divine axiom. Symmetry yes, divine preferences no.

6.1.2023 8:46pm - (Replying to @1234_abcdef1234 @VoiceBoxNoisey and @JGreenblattADL) That's why we coined the term Zionophobia, to distinguish legitimate criticism of Israel's policies from denial of a Jewish homeland. Biden should seriously consider a national initiative to fight Zionophobia, rather than leave antisemitism at the mercy of vague definitions

6.1.2023 8:38pm - (Replying to @Juaney_P @JGreenblattADL and 2 others) It was Fatima who used the word "Zionists" and she equated it with "racism", which is a first degree offence, at least to many students present at that graduation ceremony.

6.1.2023 8:33pm - (Replying to @VoiceBoxNoisey and @JGreenblattADL) There is no such thing as "legitimate critiques of Zionism", since Zionism is the quest for a Jewish homeland, and not a Government policy.

6.1.2023 8:22pm - (Replying to @yiannis_entropy) Visual metaphors never intend to capture "the entirety" of groups it depicts, yet they are useful to convey ideas lost in ordinary conversations. For example, the idea that the entirety of Palestinians are committed to destroy a country where queers are respected and protected.

6.1.2023 7:27pm - This Rabbi should have become the head of all Major American Jewish Organizations, including @HillelIntl and @AJCGlobal (reposted:

6.1.2023 7:02pm - Anti-Zionism, whether expressed by David Duke or Roger Waters or Mohammed El Kurd or whomever, is all the same: raw, toxic Zionophobic antisemitism - which the Biden Initiative has left untouched. @deborahlipstadt @Daroff @Ostrov_A @academic_la @DavidHarrisNY @DavidSuissaJJ (reposted:

6.1.2023 6:29pm - (Replying to @yiannis_entropy) You sound frank and honest, so I'll answer you frankly and honestly. I still don't grasp what makes it offensive to metaphorically display naïve people singing the anthem of their executioners who dream to destroy the one ME country where they would be respected and protected.

6.1.2023 2:45pm - Moreover, CAIR's layers can now speak in the name of Biden's Initiative against antisemitism. After all, they are part of the Initiative, after being invited to contribute wisdom and insight. How did our sages put it: "Choose your partner wisely, or else ..." (reposted:

6.1.2023 12:48pm - We are anxiously waiting for a few Zionophobic readers to complain: This is just pink-washing, so that Israel can continue to "oppress the Palestinians and ... *insert juicy slogans here* ..." Tough, you can't fool the Zionophobes, they read your most secret intentions. (reposted:

6.1.2023 12:29pm - As predicted, CAIR's lawyers are picking up all the loopholes that Biden's Initiative has left open, and are using it to mute the Initiative and encourage future hate speeches, while claiming "free speech", Islamophobia and infinite love for Jews. "Show me your partners, and" (repost4ed:

6.1.2023 12:16pm - (Replying to @RepMikeLawler) On behalf of the Daniel Pearl Foundation I commend you on your determination to fight antisemitism on college campuses.

6.1.2023 5:32am - Beautifully put: Can't they at least pretend that they seek peace? (reposted:

6.1.2023 5:23am - Watch the Recording of a Discussion on the Zionist Legacy of Leon Pinsker with Einat Wilf, Aaron Schimmel, and Daniel Polisar via @mosaicmag (posted:

6.1.2023 1:33am - (Replying to @Polytope13) (1) There are few exceptions to every sentence. (2) I should have said "the only ME country" (3) I know their executioners views on Israel, and I think they know too. (4) I hope you join me in supporting the Jerusalem Pride Parade, today, starting soon.

6.1.2023 1:13am - Your latest diatribe on Israel, published in Deutsche Welle (“Opinion: Reassessing the approach to Israel”), continues the deceitful assault on the legitimacy of Jewish self-determination. (posted:
6.1.2023 1:13am - A comprehensive, well-documented and a must-read indictment of one of the most notorious Ex-Jew of our generation. Must read! @Ostrov_A @JGreenblattADL @CotlerWunsh @GeraldNGOM @EinatWilf @AIPAC @DavidHarrisNY

6.1.2023 12:31am - What a tragic song: "We would rather die in the diaspora". And I thought the Bund has expired in the Jewish utopia of Birobidzhan (though I can still hear this song in LA, HQ of the New Israel Fund.) (reposted:

6.1.2023 12:11am - (Replying to @Polytope13) What are they doing? They are singing the anthem of their executioners, who also dream to destroy the one country where they would be respected and protected. I find this to be suicidal and, listening carefully to the lyrics, calling for a paradoxical visual metaphor.

5.31.2023 8:40pm (Replying to @robsatloff and @ForeignAffairs) I was intrigued by the One-Staters logic. I goes: The Palestinians opposed the 2S idea from the day it was proposed, and they prevented it from happening tooth and nail, therefore, it wasn't meant to be. QED. 1S philosophers underestimate the Israeli objection to 1-state.

5.31.2023 6:20pm - (Replying to @aynumazi) I am keeping diligent record of all my tweets (for a future book) but, for the life of me, I still can't tell what you find to be "inappropriate and xenophobic". Unless you conflating Xenophobic with Zionophobic, God forbid.

5.31.2023 6:06pm - (Replying to @NotTriggerAtAll) Can't parse.

5.31.2023 5:54pm - (Replying to @aynumazi) I guess I did offend someone. Who? Veggies? Chickens? Hamas? or perhaps the Zionophobes in the picture above the chickens, blindly rushing to their demise?

5.31.2023 5:19pm - (2/2) In 2007, when she told President Bush about her experience, he asked her intently: "How long did it take you to acclimate in Israel?". When she answered: "Just a few months, we finally felt welcome", he instructed one of his staff to take record of her story in Israel. (reposted:

5.31.2023 4:51pm - I feel compelled to RT this reminder of the "Farhud Day"-- June 1st. My late wife was there when it happened, 6 yrs old, seeing the killing, looting and screaming. She carried these memories throughout her life, and sadly smile when someone would speak about Jews in Arab Lands. (reposted:

5.31.2023 4:13pm - (Replying to @RobertGSilvermn) Don't be harsh on them. I call them Ex-Jews, in honor of their parents and teachers who were probably decent human beings.

5.31.2023 4:03pm - I hope I don't offend my Veggie colleagues if I repost this poster - I just love chicken logic. (reposted:

5.31.2023 3:48pm - (Replying to @yasar_abughosh) The term Zionophobe is not imposed on any ethnic, religious or national group. It is a label to recognize those who take pride in denying Jews a homeland. Such people do exist, you know, and they need a verbal label of identity; I hope they take it as a badge of honor.

5.31.2023 8:27am - (Replying to @IsraelWarRoom and @jvpliveNY) I love it when Zionophobes invoke Islamophobia

5.31.2023 8:25am - Time for CUNY Chancellor to reverse the damage: (reposted:

5.31.2023 8:22am - ( Replying to @CMAriKagan47 @IsraelWarRoom and @CUNY) Not sufficient. See (reposted:

5.31.2023 8:20am - ( Replying to @IsraelWarRoom @CUNY and @HonestReporting) Not sufficient. See (reposted:

5.31.2023 8:15am - Lawmakers lash NYC college system over ‘antisemitic’ graduation speech via @timesofisrael (posted:
5.31.2023 8:15am - It's the wrong way to reverse the damage. A genuine apology should: (1) Identify the offender (2) identify the offended and (3) Acknowledge the value and inspirational qualities of the offended group. I am waiting to hear CUNY's officials praise Israel and Zionism. @SAFECUNY

5.31.2023 7:08am - (Replying to @IbrahimMaliyam and @TyroneS42704917) The idea of "ethnic cleansing" was invented by those who tried to crush Zionism, not by Zionists.

5.31.2023 6:08am - Zionophobes are not "pro-Palestinians". I don't know any racist group that does not pose as being "pro-some-noble-cause" for a cover; Zionophobes chose the "pro-Palestinian" cover, (reposted:

5.31.2023 5:45am - (Replying to @ExpatintheUK @JGreenblattADL and 3 others) The word Zionophobes was coined for exactly this purpose: to distinguish those who object to Israeli policy from those who deny Jewish right to self determination.

5.31.2023 5:40am - Glad @JGreenblattADL mentioned the CUNY scandal together with @rogerwaters hate-fest. Moreover, he courageously raised the growing threat of anti-Zionism, which the Biden's document glaringly avoids like toxic radioactive substance, perhaps to please CAIR and @RashidaTlaib . (reposted:

5.31.2023 4:58am - International Haters Solidarity Day: Zionophobes of all countries - unite. (reposted:

5.31.2023 4:35am - (Replying to @Daroff @GSDeutch and @J_Insider) Kudos to @Daroff for pointing out the incompatibility of CAIR with the fight against antisemitism. But the qualification "policies of anti-Zionism that are antisemitic" weakens our case - I can't think of aspects of anti-Zionism that are acceptable. Can anyone?

5.31.2023 4:21am - (Replying to @TyroneS42704917 and @bernardas) You're contradicting yourself, because a homeland is precisely a place where you live in peace, not be kicked off, not shot at, discriminated against, oppressed etc. etc. This is precisely what Palestinians deny their neighbor.

5.31.2023 4:08am - (Replying to @TyroneS42704917 @hypocoristicon1 and @IbrahimMaliyam) I'm adding this syllogism to the book of Zionophobic logic.

5.31.2023 2:42am - (Replying to @Bill_C54) Yes, this is what the logic of Zionophobes looks like.

5.30.2023 10:13pm - (Replying to @TovahVillalobos @PAULMAR88310557 and @ZounaidOsman) Not in my campus, @TovahVillalobos . Here, Zionists are the devil snakes, and Palestinians eternal angelic victims.

5.30.2023 10:09pm - (Replying to @PAULMAR88310557 and @ZounaidOsman) This is the Zionophobic narrative, fabricated on Al Jazeera. Please read the actual facts, as narrated by the actors themselves in 1920-1930's: (posted:

5.30.2023 9:59pm - (Replying to @janlewisartist and @Chamanit1) With a flavor of "unjustified fear of". This is precisely what Zionophobia conveys.

5.30.2023 6:05pm - The problem is that the report is focused more on fighting haters than on strengthening Jews. In doing so, it underplays the indispensable role that we ourselves must play in determining our fate. (posted:
5.30.2023 6:05pm - Adding to our discussion of Biden's initiative, I agree with @DavidSuissaJJ , with one twist: We have another source of pride, beside liturgical tradition, it is called: Israel. I do not believe another collective could have pulled a miracle of such magnitude against such odds.

5.30.2023 4:59pm - This week's report of CI articles has swollen to 9 pages: Several of the items are PhD Dissertations, taking us into the future. Readers concerned with explanation, should read Cao et al, on Semantics for Causing, Enabling, and Preventing Verbs Using SCM.

5.30.2023 3:29pm - (Replying to @LeiQiaosi) Being anti-Zionist means being against a Jewish homeland in any form or shape. If she wishes to speak against certain policies or actions, she should do so, but not against the core identity of other students. Americans did not demonize Muslims identity after 9/11.

5.30.2023 3:12pm - (Replying to @ZounaidOsman) You got the causal logic upside down. All the violence and suffering you are describing were caused, not by Zionism, but by attempts to crush Zionism in its bud. This is where the irrationality of Zionophobia looms.

5.30.2023 2:46pm - (Replying to @TyroneS42704917) I am a Zionist! Namely, I believe Jews have the right to a homeland. Am I a "problem" ?

5.30.2023 2:41pm - (Replying to @noah_greifer and @soboleffspaces) If you go outside causal inference, you have more reasons to look into the equivalence question: For what sets of variables, T and Z, would this equality hold: SUM_t P(y|x, t)P(t) =SUM_z P(y|x, z)P(z) See for complete answer.

5.30.2023 1:46pm - Uncharacteristic of European diplomats, @ukinsirael refrained from adding the moronic adage "We urge both sides to be nice to each other". (reposted:

5.30.2023 1:27pm - (Replying to @kyberknave and @ZounaidOsman) As I said, Zionophobes are truly phobic, more so than even Islamophobes.

5.30.2023 1:23pm - (Replying to @soboleffspaces and @noah_greifer) The last point need be emphasized, in other words, backdoor admissibility implies strong ignorability and, moreover, IPSW is asymptotiaclly equivalent to covariant adjustment, even under non-ignorability . See "Understanding Propensity Score":

5.30.2023 8:29am - (Replying to @DavidDeutschOxf and @tszzl) We are likewise praying that Israel's errors are in this category, and that "democratic institutions" will prevail over bad policies and bad leaders.

5.30.2023 8:03am - When I coined the word "Zionophobia", friends advised me to tone it down. After all, they said, "phobia" connotes an incurable sickly obsession, and you don't want to make it sound like that. I was right. (reposted:

5.30.2023 7:30am - I remember my parents' face as the news was announced on the radio. We all felt like a third Churban (destruction) is eminent, and only a miracle can save us from another genocide -- It did! (reposted:

5.30.2023 7:24am - As the holy book says: Its good to start the day on a bright note. (reposted:

5.30.2023 7:12am - (Replying to @JSherredAdams and @rogerwaters) Wrong

5.30.2023 1:16am - (Replying to @kamiyacho74 @Evolving_Ego and @rogerwaters) No way! Read before you speak. @rogerwaters was not condemned for criticizing Israel, but for incitement to genocide.

5.30.2023 12:13am - (1/2) Ears up, Ears up!! While we were busy with Biden's initiative, a memo was issued by Department of Education on "Discrimination against Jewish Students": Again, in vain you will search for the word Zionism; Govt agencies hire hordes of lawyers to help

5.30.2023 12:13am - (2/2) them avoid the number one word that captures most hostilities against Jewish students. It must be avoided at all costs, so that students can continue to be targeted, now legitimately, on suspicion of having Zionist thoughts. @CatherineLhamon @deborahlipstadt @Ostrov_A

5.29.2023 10:18pm - One is never tired of learning new things about great philosophers. Russell couldn't possibly have imagined that 80 yrs later, some of his own countrymen would be marching under a "From the River to the Sea" banner. (reposted: (updated 5.30.2023)

5.30.2023 12:13am - (1/2) Ears up, Ears up!! While we were busy with Biden's initiative, a memo was issued by Department of Education on "Discrimination against Jewish Students": Again, in vain you will search for the word Zionism; Govt agencies hire hordes of lawyers to help

5.30.2023 12:13am - (2/2) them avoid the number one word that captures most hostilities against Jewish students. It must be avoided at all costs, so that students can continue to be targeted, now legitimately, on suspicion of having Zionist thoughts. @CatherineLhamon @deborahlipstadt @Ostrov_A

5.29.2023 10:18pm - One is never tired of learning new things about great philosophers. Russell couldn't possibly have imagined that 80 yrs later, some of his own countrymen would be marching under a "From the River to the Sea" banner. (reposted:

5.29.2023 9:52pm - (1/2) Note that the sentence about exclusion is a description of what often happens, not of what we should label "antisemitic". The only targeting explicitly labeled AS is "Jews targeted because of their BELIEFs or IDENTITY." But nowhere does it say that, for some Jews, Zionism (reposted:
5.29.2023 9:52pm - is an integral part of their identity or beliefs. Nor does it say that denying Jews' right to self determination is antisemitic. In short, they left Zionism a legitimate punching bag for all evils in the world. Purposely? Negligently? Pressured? Perhaps @deborahlipstadt knows.

5.29.2023 6:54pm - The problem with Biden's Initiative is that it says nothing about such calls to exclude Zionist professors (like me) and students (like most of mine) from campus life. It seems that the 1000 brains behind the initiative worked hard (months, I'm told) to make sure it says nothing. (reposted:

5.29.2023 6:43pm - (Replying to @JSherredAdams and @rogerwaters) The only leaders I know who claim their neighbors have no right to exist are Vladimir Putin and Mahmoud Abbas. The US, as far as I recall, has been quite accommodating of its neighbors.

5.29.2023 4:47pm - (Replying to @SAFECUNY and @CNN) I find it hard to imagine that CNN would miss this opportunity to tell its viewers where higher education is heading @jaketapper @wolfblitzer @DanaBashCNN

5.29.2023 4:06pm - (Replying to @mrSatias @deborahlipstadt and 3 others) Mandela kept preaching non violence, and never advocated genocide against his adversaries. Listen to what the Palestinians advocate; today, not in 48; in Tel Aviv, not in Beit El; In the open, not in hiding. Tell us they are not to be classified as suspects of hostile intentions.

5.29.2023 3:55pm - This interview should have been aired on CNN, the news outlet that I normally watch when I am curious about higher education. @JackTapper @DanaBashCNN @wolfblitzer (reposted:

5.29.2023 3:02pm - I am surprise how the dozens of Jewish advisors on the Biden's document have not caught this loophole. Those who advocate the destruction of Israel, not for Jew hatred, but because they see Jews as a religion (hence undeserving of a state) would not be classified as antisemitic. (reposted:

5.29.2023 2:38pm - We have travelled a long way in the past century. In 1939, the Nazis did not have any flag to attack, tear, or burn, so they burned human beings. I hope @rogerwaters disciples continue to tear flags, knowing that Israel will protect human beings. (reposted:

5.29.2023 2:23pm - (Replying to @realEssYoo @StaceyEBurke and 2 others) I cannot in good faith accuse all Jewish organizations who dance and prance around the Biden's initiative of being driven by assimilation forces. I know some of them first hand -- they aren't. Securing a "seat at the table" is a more plausible explanation.

5.29.2023 2:07pm - But Biden's new initiative does not exonerate terrible schools from the charge of antisemitism, no matter how terrible you are. @SAFECUNY (reposted:

5.29.2023 2:00pm - (Replying to @gail_notestine @davidmdraiman and 2 others) The Bund is thriving in Birobidzhan and in Los Angeles. I didn't realize NJ has a chapter. (posted:

5.29.2023 1:48pm - (Replying to @kenkoonwong) I do not see any hindrance to applying the do-operator to continuous variables. Take the effect of dosage X=x on lifespan Y=y, P(y|do(x)) can be computed by the same surgery as in discrete models. Issues arise for "dosage X>x", but see

5.29.2023 1:08pm - Amazing video from the Smithsonian archives. I urge all readers to view it seriously, ask if they believed it actually happened in America, and compare it to what we have seen in Charlottesville, Virginia, CUNY's graduation, NY, and @rogerwaters rally in Germany. @SAFECUNY (reposted:

5.29.2023 11:50am - Seeing a handful of protesters penetrate @rogerwaters hate festival and voice a message of sanity, with Israeli flags, will cheer my heart for the rest of this week. It reminds me of the Jewish Partisan Song "We were here" (Anachnu Po). I wish I could be with them. (reposted:

5.29.2023 12:42pm - (Replying to @Evolving_Ego and @rogerwaters) Same Roger Waters. It's easy to be "anti-fascist" once Hitler gave fascists a bad name. What's hard is to know who are the real fascists these days. Waters chose Zionism (i.e., the quest for a Jewish homeland) to be his target, instead of Putin and Iran. Sad and stupid.

5.29.2023 12:09pm - (Replying to @realEssYoo @RobertGSilvermn and @MikeNew19512115) I am not sure it's only loyalty to the Democratic Party (I'm loyal too) or fear of losing a piece of the pie which this initiative is promising, both in visibility and in funding.

5.28.2023 9:09pm - (Replying to @g_ali_pdx) Can Zionophobic lips ever utter anything but profanities?

5.28.2023 9:00pm - (Replying to @FJnyc @RitchieTorres and 3 others) Thanks for reminding me of my 2022 tweet. Today, I am not as sure of the prediction I made last year, having seen the cheering faces of CUNY faculty behind Fatima, and the discriminatory behavior of CUNY's Chancellor against his own Zionist faculty and students.

5.28.2023 5:53pm - Thanks for joining the long line of awaiting. Unfortunately we haven't heard yet from any of the 1000 authors, perhaps they are still trying to decipher what 'Singling Out Israel Because of anti-Jewish Hate Is Antisemitic' means, and whether Roger Waters actually loves Jews. (reposted:

5.28.2023 5:02pm - (1/2) The key to resolving the conflict and is to understand why Westerners refuse to listen to what Palestinians are telling them, i.e., "Israel must go". @EinatWilf discusses it in "The War of Return":, but I find the question still perplexing: How can (reposted:
5.28.2023 5:02pm - (2/2) rational educated people, among them academic, commentators and reporters ignore or deny a phenomenon that is so ubiquitous, universal and critical. I wish I could interest 12 PhD students in Cultural Psychology to work on it.

5.28.2023 4:25pm - My ChatGPT warned me: Watch out! racists do not like looking in the mirror, especially anti-Zionists. (reposted:

5.28.2023 3:48pm - It is only when you ask the racists: Exactly what land is occupied? And they answer: "from the river to the sea" that you come to understand why it is still occupied, and who prolongs the occupation. (reposted:

5.28.2023 3:34pm - (Replying to @Bruced872 @deborahlipstadt and 3 others) A notable features of racists is their need to define what other groups are or aren't, without asking them - Racists Know Best (RKB). Please tell me what Zionism is, Mr. RKB

5.28.2023 2:51pm - I think everyone (including the 1,000 brains behind Biden's Initiative) understands what antisemitism is: Executing or advocating acts that harm the Jewish people as a collective. This includes of course stripping Jews of their homeland, i.e., anti-Zionism. Simple. (reposted:

5.28.2023 1:32pm - I wouldn't go as far as saying that Biden Strategy is a failure. It was given two litmus tests: (1) Roger Walters and (2) CUNY's graduation, and the Jewish world is waiting to see how it responds to the two tests, if at all. @deborahlipstadt @JoeBiden @AaronBandler (reposted:

5.28.2023 1:15pm - (Replying to @SteveSammut2 @deborahlipstadt and 3 others) Big difference between "ist" and "ism". Zionism is the belief in the right of Jews to a homeland. Zionists are people who work to protect and sustain that homeland. Which one does Waters want to "gently remove", and how? Is he for "ethnic cleansing" or "home demolition"?

5.28.2023 1:01pm - (Replying to @artillier @MichaelShurkin and 6 others) None mentioned a Palestinian who is not planning to dismantle Israel.

5.28.2023 12:52pm - (Replying to @g_ali_pdx @deborahlipstadt and 3 others) Roger Waters mentality has infested more than one victim.

5.28.2023 12:44pm - (Replying to @SteveSammut2 @deborahlipstadt and 3 others) It is only when you ask the protesters: Which land is occupied? And they answer: "from the river to the sea" that you understand why it is still occupied, and whose stain it is.

5.28.2023 9:07pm - (Replying to @artillier @SpaceGk and 6 others) I challenge all the paranoia experts, and the wise sages from "get back on earth" to name ONE Palestinian of stature who is not planning the end of Israel. Just one. Blindness kills. (posted:

5.28.2023 8:43am - Yet Biden's initiative, and the thousand brains behind it, have chosen to keep this racist ideology out of their founding document, so that the document would be praised by wider circles. @deborahlipstadt @AaronBandler @Ostrov_A (reposted:

5.28.2023 8:26am - Watch not the words of Fatima, but the shining faces of her professors, gleaming behind her, as she tramples on the identity of her fellow students and labels them "Racist". (reposted:

5.28.2023 8:14am - (Replying to @SpaceGk @MichaelShurkin and 5 others) I'll forgive @amnestyusa for its "double standard" -- Israel wants to be judged by higher standards -- but I can't forgive their failure to acknowledge, even once, the Palestinian national ethos of dismantling Israel.

5.28.2023 7:57am - (Replying to @willowt98239336 @deborahlipstadt and 3 others) "Roger Waters said so, so it must exist" said the gullible and misinformed.

5.28.2023 7:54am - It's a great honor to be on a Reading List next to Thomas Kuhn's The Structure of Scientific Revolutions, and to imagine that readers, like that lady in the photo, can enjoy the summer breeze through your words and thoughts. #Bookofwhy (reposted:

5.28.2023 7:30am - I love: "the distinction between pre-and post-1967 borders is a fiction sold to willing Western liberals, who do not want to acknowledge that cries to ?\200\234end the occupation?\200\235 really mean ending Israel..." No Palestinian thinks differently and no Western liberal is willing to listen. (reposted:

5.28.2023 7:20am - (Replying to @Francis_Watts @deborahlipstadt and 3 others) Thus spake our all-knowing Zionophobe .

5.27.2023 10:56pm - We are told that one thousand brains worked on the President's initiative for fighting antisemitism. Surely one of them can tell CUNY students how the initiative deals with their experience at graduation, where they were labeled "racists". One brain, please! @deborahlipstadt (reposted:

5.27.2023 6:03pm - (Replying to @Peter__Courtney @deborahlipstadt and 4 others) There is. Zionism is the belief in the right of Jews to a homeland. Zionists are people who believe that Zionism is the most noble enterprise of the 20th Century (echoing Nancy Pelosi). Big difference for Roger Waters, the "stain remover." Perhaps Biden's strategy can say more?

5.27.2023 3:41pm - Dear President Biden, Roger Waters urges his followers to "gently remove" me and my Zionist colleagues from the face of the earth. Can we count on some reaction from your new strategy on fighting antisemitism? @deborahlipstadt @Ostrov_A @ShelleyGspan @DouglasEmhoff @AaronBandler (reposted:

5.27.2023 1:49pm - This is an opportunity to show what the antisemitism strategy may do for Jewish students at the trenches: Would we hear from you? @deborahlipstadt @Ostrov_A @SAFECUNY @AmbassadorRice @ShelleyGspan @DouglasEmhoff (reposted:

5.27.2023 12:09pm - (1/2) I have zero feelings toward the hijabi speaker, given her Jihadi upbringing in Yemen. But I have zero respect for CUNY law professors cheering her while their Jewish and Israeli students are seeing their cherished symbols of identity smeared in public. I am also surprised at (reposted:
5.27.2023 12:09pm - the architects of Biden's initiative against antisemitism for not rushing to make this speech a showcase of what the initiative intends to render unacceptable on campuses receiving Federal Aid of any sort. @deborahlipstadt @Ostrov_A @JGreenblattADL @AaronBandler @DionJPierre

5.27.2023 8:45am - NYC?\200\231s public law school releases video of ?\200\230antisemitic?\200\231 commencement speech -speech/ via @timesofisrael (posted: ment-speech/)
5.27.2023 8:45am - (1/2) I am personally offended by CUNY professors cheering a racist commencement speaker who says: "May we rejoice... the fight against Capitalism, Racism, Imperialism and Zionism around the world (1:27 min into the video) President Biden new initiative on
5.27.2023 8:45am - antisemitism need be tested by its immediate actions on behalf of Jewish students at CUNY who were exposed to a speech that denigrates the core of their identity. I challenge the architects of that initiative to respond publicly and quickly @JGreenblattADL @Ostrov_A @SAFECUNY

5.27.2023 7:21am - (Replying to @RobBaumgartner4) I think it's too late now - eliminationism is the prevailing mentality across all tribes/families. Peace's only hope rests with Western actors imposing heavy costs on anti-coexistence rhetoric in education and media.

5.27.2023 6:13am - Well put, but this makes the Palestinian cause almost impossible, unless we shift paradigms and start with a picture of the ME as we would like to envision it 20 yrs from now: 2-States for 2 peoples, equally legitimate and equally indigenous. (reposted:

5.27.2023 5:59am - (Replying to @mircomusolesi) I am also in this CS paradigm which cares about the information content of the model, not its software/hardware implementation. The fact that NNs resemble neural circuitry does not make them a better tool for modeling Covid19 knowledge than, say, explainable Causal Graphs.

5.27.2023 5:37am - (Replying to @larsweisbrod and @sewenz) I have elaborated on this question in: "Radical Empiricism and Machine Learning Research".

5.27.2023 5:31am - Mahmoud Abbas scores symbolic victories at UN, Arab League (posted:

5.26.2023 11:00pm - No better way to consummate the Shavuot holiday than with the song that we used to sing in Kindergarten "Saleinu Al Ktefeinu" Dedicated to my K-teacher and to all readers for whom the song means as much. (posted:

5.26.2023 8:58pm - (Replying to @YasMohammedxx) Brilliant!

5.26.2023 8:41pm - (Replying to @blakeflayton) Brilliant! Post-logic logicians have formalized a new semantics for "meaningfulness", "value system" and "humanity". What remains to be seen is how "Israeli humanity" factors in the new notation.

5.26.2023 8:31pm - (Replying to @robsatloff) "Embracing a belief" doesn't make them antisemites, agree. But denying self-determination rights to those who do not wish to be thus embraced IS Zionophobic and, to simplify things, Antisemitism is a good first approximation.

5.26.2023 5:10pm - (Replying to @Jacob_Labendz) Flattening everything into AS hurts, agree! But let us do the Unflattening on a case-by-case basis. The word Zionophobia was coined for such unflattening. But, In the meantime, AS is a convenient surrogate for "socially unaccepted", such a word is badly needed.

5.26.2023 2:04pm - (Replying to @ComradeAdil_ @BanalSex_83 and 2 others) I read back our tweets and behold, it is you who is lost. My arguments are solid and consistent, yours are wabbling between facts and populist cliches. I propose the parties "equal legitimacy, and equal indigeneity", you propose the death of one, and call yourself "progressive".

5.26.2023 1:48pm - (Replying to @gregg_mashberg @crayandrizzy and @JTHanau) BDS cronies will use this sentence: "to fight a/s ?\200\234including efforts to delegitimize Israel,?\200\235 as a proof the Biden administration does not consider "delegitimizing Israel to be inherently A/S" hence it should be acceptable on campus.

5.26.2023 1:43pm - (Replying to @gregg_mashberg @crayandrizzy and @JTHanau) Agree; publicly announcing an urgency is a huge step forward. However, the points you bring up are merely expressions of this administration's sentiments, they cannot be used by my students to argue that a certain behavior is "antisemitic", namely, has no place in the university.

5.26.2023 12:51pm - (Replying to @ComradeAdil_ @BanalSex_83 and 2 others) Show me one "colonial project" by settlers speaking a language spoken in the land in ancient times, prior to that spoken by its current residents. Language, holidays, role-models, poems, etc. are not " this reason or that," these are the defining features of "homecoming".

5.26.2023 12:31pm - (Replying to @ComradeAdil_ @BanalSex_83 and 2 others) I think you meant to say that the attempt to strangulate Zionism has desecrated the land, and resulted in infinite sufferings to two nations, one of them still dreams of annihilating the other .

5.26.2023 12:24am - (Replying to @FoxmanAbraham) Abe, aside from telling Jews: "you will not stand alone", didn't you expect more teeth to the document itself? And isn't a tooth-missing document sending a warning signal to Jews, saying: "Others will decide how to diagnose and treat the disease - not you!"

5.26.2023 12:07am - (Replying to @Novalisu @Ostrov_A and @rogerwaters) ??

5.25.2023 11:23pm - (Replying to @ComradeAdil_ and @blakeflayton) The UN did not "offer", but rather "ratified" a moral imperative entailed by two legitimate national movements claiming ownerships of the same land, one by demographic presence, the other by historical connection - two claims based on different logics, yet carrying equal weight.

5.25.2023 11:10pm - (Replying to @blakeflayton) Some minor corrections. (1) Palestinians didn't just say no to THESE borders, but to ANY borders. (2) There were 1.2M jews in Europe and Arab countries, waiting to return to their historical homeland. (3) Palestinians didn't just "say no" but planned and attempted a genocide.

5.25.2023 9:25pm - The photo is missing two items: (1) A spokesperson for CAIR. (2) Twelve Jewish leaders chanting Halleluiah over an empty document. @AaronBandler @Ostrov_A @ZionessMovement @EndJewHatred @SAFECUNY (reposted:

5.25.2023 9:15pm - (Replying to @Jacob_Labendz) I don't buy it. For Israelis, Israel is a national, not a religious state. It is antisemitic to classify and treat Jews by any category that stands against their own choice, ignorance is not an excuse.

5.25.2023 8:49pm - Thanks so much for posting this beautiful interpretation of the Book of Ruth, which we are reading tonight (Shavuot holiday) and which highlights the centrality of Jewish peoplehood over Rabbinical orthodoxy. For summary, see (reposted:

5.25.2023 8:40pm - (Replying to @EinatWilf) I missed it - is a recording available?

5.25.2023 4:34pm - (1/2) I cannot get over my embarrassment each time I read this circular definition: 'Singling Out Israel Because of anti-Jewish Hate Is Antisemitic'. Dozens of Jewish organizations now read this sentence and go overboard praising the document for its wisdom and courage. Can't (reposted:
5.25.2023 4:34pm - they see its logical emptiness? It's like saying: "who is an antisemite? He who thinks like one" I understand the pragmatics of praising the administration for its good will and efforts but, man, I'd still be embarrassed. @JGreenblattADL @TedDeutch @AaronBandler

5.25.2023 4:15pm - I cannot get over my embarrassment each time I read this circular definition: 'Singling Out Israel Because of anti-Jewish Hate Is Antisemitic'. Dozens of Jewish organizations now read this sentence and go overboard praising the document for its wisdom and courage. Can't

5.25.2023 4:15pm - Sorry, @Ostrov_A , @rogerwaters does not qualify as anti-Semite according to the Biden new Document. Why? Because he never said he hated Jews, the one and only criterion in the document. Waters acts out of very noble motives, mostly progressive. (reposted:

5.25.2023 4:07pm - (Replying to @robsatloff) Well put. Same for those who advocate the destruction of Israel, not for malice (God forbid), but because they believe that Jews prefer to be stateless, pure benevolence; aren't they antisemitic?

5.25.2023 2:46am - (Replying to @dylanotes and @jdforward) If you come to my university and listen to what Jewish students are complaining about, you'd quickly realize that right-wing antisemitism is a minor nuisance practiced by laughable fringe groups, while serious hostilities come from left-wing anti-Zionist groups, mobilized by BDS.

5.25.2023 1:16pm - (Replying to @MarkGoldfeder) The document does not mention Zionism even once. It mentions Israel once and says nothing about denying Israel's right to exist. @AaronBandler

5.25.2023 1:08pm - Biden unveils antisemitism plan, says singling out Israel can be antisemitic posted: lan-says-singling-out-israel-can-be-antisemitic/00000188-5350-df79-a19d-f35c94600000)

5.24.2023 11:52pm - (Replying to @robsatloff and @TheIHRA)

5.24.2023 11:36pm - The Biden administration?\200\231s antisemitism strategy is expected to be released later this week. (posted: portedly-feature-ihra-nexus-definitions-of-antisemitism/)
5.24.2023 11:36pm - (1/2) What an empty debate! I too have reservations about some items on the IHRA definition of antisemitism, who doesn't? But I understand that this is not the time to show how scholarly clever I am, spoil a hard-earned wide consensus and give our enemies (reposted:
5.24.2023 11:36pm - an excuse to say: "See, even Jews can't agree on what antisemitism is". @AaronBandler @Daroff @JGreenblattADL @CotlerWunsh @POTUS @EinatWilf @DavidSuissaJJ @JNS_org @Algemeiner @DovWaxman

5.24.2023 11:49pm - (reposted:

5.24.2023 11:47pm - (reposted:

5.24.2023 11:46pm - Congratulations on overcoming the setbacks. (reposted:

5.24.2023 9:34pm - (Replying to @norabelrose) I have difficulty seeing the relation between the functional nature of NN and a causal model. The latter contains causal knowledge about the world, the latter does not. Is it merely a matter of "interpretation" to convert one to the other? Knowledge must come from somewhere.

5.24.2023 9:00pm - I feel for Palestinian "experts" who labeled Shavuot a "pagan holiday". You see, Jews have so many Biblical holidays that connect them to the land; Palestinian have none. Sad! (reposted:

5.24.2023 6:55pm - The 2023 Conference on Theoretical Aspects of Rationality and Knowledge (TARK 2023) will be held from Wednesday, June 28 until Friday, June 30, 2023, at the University of Oxford, UK. Accepted papers: see (posted:

5.24.2023 4:33pm - (Replying to @academic_la) These Ex-Jews have never stopped to read what their prayers say: Hu Yolichenu B'Kommemiyut L'Artzenu. (He will walk us in sovereignty to our land).

5.24.2023 4:27pm - Biden Makes Demands of Israel to Gain US Support for Saudi-Israel Peace: (1) Shelve judicial reform (2) Resume two-states negotiation. Who could ever imagine the Saudis to be the ones to help realize the Zionist dreams: (1) Democracy, (2) Co-existence

5.24.2023 4:10pm - The Daniel Pearl Foundation joins @daroff , @JGreenblattADL and many other community leaders in urging the @WhiteHouse to embrace @TheIHRA as the sole definition of #antisemitism in their strategy. Anything else permits anti-Zionist forms of antisemitism. s-with-a-sensible-definition#.ZG07DLWJkUk.twitter (reposted:

5.24.2023 3:54pm - (1/2) @KenRoth has a point! Feeling a debt toward Jews should not mean writing a blank check to the Israeli government. No, the check must be directly proportional to the level of existential threats that Israel finds itself at any given time, and should factor in both direct (reposted:
5.24.2023 3:54pm - (2/2) genocidal threats pronounced by Iran, Hamas, Hezbollah, the Palestinian leadership itself, and indirect genocidal threats implied by anti-coexistence Zionophobes, like @KenRoth @rogerwaters and their emboldened ilk.

5.24.2023 2:08pm - Dee also said he hopes to set up a meeting with Warner Brothers Discovery Chairman and CEO David Zaslav. (posted: -amanpour-cnn/)
5.24.2023 2:08pm - I join Rabbi Dee in wondering: Why does it take the threat of legal action to have CNN wake up and... simply report the truth about Israel?

5.24.2023 2:56am - (Replying to @_ZachFoster and @M_Lipshutz) How can we tell when a Zionophobe lost the thread of his arguments? When he starts using populist cliches like "apartheid" and "ethnic cleansing", thinking that the ugliness of these words would make up for broken thoughts.

5.24.2023 2:47am - Baby, Mama and Papa are doing well and, despite sleepless nights and countless sneers by jealous neighbors, papa wouldn't trade these moments for anything else. (reposted:

5.24.2023 2:23am - (Replying to @_ZachFoster) To make you more comfortable we have coined the word "Zionophobia," which hasn't been weaponize by the bad guys, and which you, as a jew, could use to condemn the ugliest discrimination against Jews -- denying them a homeland. Long live comfort.

5.24.2023 1:58am - (Replying to @JAsayuel and @AvivaKlompas) It's one thing to "sign a deal", it's another thing to honor it when your children swear that your neighbor has stolen your land.

5.24.2023 1:45am - Mr. President, you have called antisemitism “a stain on the soul of America.” Please do not let that happen on your watch. (posted:
5.24.2023 1:45am - A powerful Letter to President Biden concerning the necessity of using the International Holocaust Remembrance Alliance (IHRA) definition in fighting antisemitism.

5.24.2023 1:37am - Prof. Alan Dershowitz does not say how CNN can compensate victims of terrors for "trying to create a moral equivalence" between the victims and the terrorists. I wish there was some sort of acceptable apology, perhaps a commitment for a balanced coverage of the conflict. (reposted:

5.23.2023 10:06pm - Love the title: The internet and Hate Speech in a Post-Truth Society . Promises to cover vast territories, from Islamophobia to Asianphobia to Zionophobia. (reposted:

5.23.2023 9:36pm - To our readers in Latin America: Truly appreciating your Parliament's support of our fight against antisemitism and Zionophobia. @MarkGoldfeder (reposted:

5.23.2023 11:40am - (Replying to @ulfnielsen) For a moment I thought pharma aims to exclude data that comes from alternative universes.

5.23.2023 11:36am - (Replying to @ceobillionaire) Thanks for posting; thoughtful scenarios and convincing arguments.… (posted:

5.23.2023 9:48am - Here is weekly report of recently published CI papers: Several of them in a volume titled: real-world-evidence in Medial Product Development which got me a bit puzzled; is there a piece of evidence which is not "Real World"? (posted:

5.23.2023 8:08am - Orthodox Israel, who worships spirit over matter, should humbly learn from the spirit that propelled secular Israel toward these accomplishments, and cede claims for sole guardians of Jewish life. @EinatWilf (reposted:

5.23.2023 7:24am - (Replying to @DrJBhattacharya) I don't recognize "causal inference" in this fire storm. If I were to depict it, it would look like a boat going up the river, reaching a calm blue lake of clarity and serenity.

5.22.2023 10:12am - Once again, the American Anthropological Association will be voting on a (twice rejected) “Resolution to Boycott Israeli Academic Institutions”. I must share with you a website launched to clarify why it should be voted down decisively: (posted:

5.21.2023 1:35pm - (Replying to @AlfredoMorabia @DrRichBesser and 15 others) Multi-congratulations, Alfredo. How about a link to the index?

5.21.2023 1:19pm - (Replying to @kenkoonwong @anirbanb_007 and 2 others) Neural network is a model of the data, not of the world that generates the data. Its role in causal inference is to manage the estimation part of the process, going from finite samples to distributions; do-calculus takes us from distributions to causal questions.

5.21.2023 9:25am - 'BY ANY MEANS': Israel-haters march in Toronto, call to annihilate Jewish state, globalize terror | World Israel News (posted:
5.21.2023 9:25am - When I wrote that @RashidaTlaib Nakba Day is not about the 1948 Nakba, but about the Nakba associated with the destruction of Israel I knew what I was saying.

5.20.2023 6:04pm - (Replying to @aynumazi @jacobin and 3 others) What part of the model you have doubts about?

5.20.2023 6:02pm - (Replying to @aynumazi @jacobin and 3 others) I did not say "Oppressing" I said "harbor genocidal ambitions." Are you doubting the latter?

5.20.2023 4:40pm - (Replying to @terrible_archer @peteratmsr and @eliasbareinboim) The points you are raising are crucial for understanding what ChatGPT can do for us in the near future. I am still trying to formalize how much causal knowledge it can safely fake

5.20.2023 4:20pm - (Replying to @erikbiz) "Hatred" is a slippery concept. You can't get a Zionophobe to admit that he/she hates Israelis, for it's not socially acceptable to hate people. But you can easily get a Zionophobe to say that he/she is planning the end of Israel, a genocide that IS socially acceptable.

5.20.2023 1:54pm - Come to think about it, of all racist groups currently roaming our social media, Zionophobes are the only ones who actually harbor genocidal ambitions. (reposted:

5.20.2023 1:35pm - To all Portuguese speaking readers, including Brazilians, appreciating your support of the Israeli miracle, its moral imperatives and its contribution to world's prosperity. (reposted:

5.20.2023 6:50am - (Replying to @DavidHirsh) @_ZachFoster and other "Historians of Palestine" have not heard about Zionophobia; an uglier form of racism, of which they cannot claim innocence.

5.20.2023 5:50am - Speaking about "evidence" for causal hypotheses, in this episode of the @ezraklein podcast: we hear about statistical evidence (Rung-1), but not about the causal assumptions (Rung-2) that turn the data into a definitive evidence. Can it be repaired? (posted:

5.19.2023 11:50pm - (Replying to @soboleffspaces) Mediation is good example, and so is attribution, e.g., finding "causes of effects," or finding probabilities of harm and benefit . Non-compliance, however, is a Rung-2 problem.

5.19.2023 11:28pm - (1/2) So why am I so excited to hear that the oldest manuscript of the Hebrew Bible was auctioned yesterday (for $38,1M) and will soon be displayed in Tel Aviv Museum of the Jewish people: See for a closer look. The reason is probably (reposted:
5.19.2023 11:28pm - (2/2) that I recognize these very letters from first grade, and same very stories from second or third grade. I won't be surprised if Abu Masen proclaims it a "Stolen Palestinian property", since some of the authors lived in the Holy Land, which makes them Palestinian by birth.

5.19.2023 9:57pm - Many readers have expressed ambiguity regarding the distinction between Rung-2 and Rung-3 of the Ladder. I'm retweeting my answer with the hope clarifying this barrier. (reposted:

5.19.2023 4:10pm - Sharing @bariweiss conversation with Daniel Gordis, which I've found to be just the right thoughts with which to celebrate Israel's 75th anniversary. (posted:

5.19.2023 3:53pm - (Replying to @jmugan) If we have a causal model then, you are right, we can get an an answer to every counterfactual question. Now suppose you have only experimental studies, no model. Some counterfactual questions could still be answered, some could not. The former are Rung-2, the latter are Rung-3.

5.19.2023 12:49pm - Tlaib’s shameful weaponization of the Elie Wiesel Genocide Act against Israel (posted:
5.19.2023 12:49pm - Elisha Wiesel, son of Elie Wiesel, speaks against @RashidaTlaib 's attempt to hijack and invert a Genocidal Bill named after his father

5.19.2023 12:21pm - Made my day. Now I fee that: "We're back to this ancient city" (Chazarnu El Borot Hamayin ... Ba'Eer HaAtica" (reposted:

5.19.2023 1:23am - And what has Robert Malley learned about the Middle East in the past 20 years? (reposted:

5.19.2023 1:16am- (Replying to @LahavHarkov @JeffreyWoolf and @academic_la) Separately is not what Prophet Zakaria envisioned. He was smart enough to tell us that free and joyful Jerusalem means: "full of boys and girls playing in the streets."

5.19.2023 1:04am- (Replying to @SugyunA) This example (ie Simpson's Paradox) is indeed the best way I've found of convincing people that the barriers between levels of the Ladder are not artificial but fundamental.

5.18.2023 8:54pm - (Replying to @mcuban) I used to believe that the only reason a LLM cannot fake an LKM is the prohibitive complexity needed. ("You can't fake it unless you have it"). Now I am not too sure; there might be other barriers to faking, but I can't pin them down formally.

5.18.2023 6:09pm - I'd love to celebrate Jewish return to Jerusalem, but I don't like what I see on TV -- no women in the crowd, not even girls. This is not the Jerusalem prophet Zakaria envisioned: "And the Street of the city shall be full of boys and girls, playing in the streets." @academic_la (reposted:

5.18.2023 11:27am - (Replying to @Edmel2401 @johndebois and @HenMazzig) The barn's gate. Through the fox's lens, a chicken is an injustice, a neighbor is a crime, coexistence a humiliating defeat, and the barn's gate -- an apartheid.

5.17.2023 4:07pm - (Replying to @dita_gjuraj @daoudkuttab and @hahussain) I was there, in 1948, old enough to read newspapers. The establishment of the state in itself entailed an open arm of co-existence between two indigenous populations. What entailed expulsion & killing was an attempted genocide of one or the two indigenous populations.

5.17.2023 3:54pm - (Replying to @johndebois @AlGhanadreh_2 and @HenMazzig) The eyes of the beholders may see a settler in every homecoming neighbor. Some see a settler even one's own bed.

5.17.2023 3:47pm - Thanks for reposting these quotes. I would have reaffirmed my readiness to defend them ferociously in any forum and form, had they not become self-evident in our day and age, at least this is what I would like to believe. (reposted:

5.17.2023 3:55am - (Replying to @PWGTennant) Same as my advice to readers of this Twitter channel: Stick to the Scientific Paradigm: state what you want to know, what you already know, what data are available, and go from the latter to the former. This alone would get you ahead of 90% of Machine Learning people.

5.17.2023 12:23am - The Senate hearing we had today on how to regulate AI, with @GaryMarcus and Sam Altman, was even more illuminating: I'm at awe of the depth and seriousness of the Senators on the Committee. (reposted:

5.16.2023 11:41pm - (Replying to @YoniMichanie) Congratulations, Yoni!! Mazal Tov and lots of Nachat.

5.16.2023 11:39pm - They are still decisive today, in our war of spiteful survival. I wish Mahmoud Abbas would listen to Ben Gurion, and won't repeat the mistake that Azzam Pasha made in 1948. (reposted:

5.16.2023 11:34pm - Abbas speech provides an important glimpse at Palestinians mentality, just in case some readers have doubts, or wish to entertain the possibility that "not all Palestinians deny Jewish connection to the land" or "Not all Palestinians deny Israel's legitimacy". (reposted:

5.16.2023 11:14pm - In his lecture of 2 days ago, Yuval Harari makes more convincing predictions about the future of GAI and humanity. (reposted:

5.16.2023 10:23pm - @RashidaTlaib 's Nakba ceremony would not have been controversial had it aimed to commemorate the '48 tragedy. It became controversial by her declaring its purpose to be the reversal of the 1948 outcome, namely, inflicting a new Nakba on another people. You can't ignore the latter (reposted:

5.16.2023 10:11pm - Dangerous development. Neo-Nazis learned from the left that "Zionism" can be bullied with impunity (e.g., Berkeley Law); expect to see them copying slogans and becoming international champions of "From the River to the Sea". Hitler gave antisemitism a bad name. @EinatWilf (reposted:

5.16.2023 9:41pm - (Replying to @PavlosMsaouel @chubbyemu and 14 others) Very educational conversation, for me. I finally understood what the Big Data Paradox is. Never expected to be referenced with @XiaoMeng1 , with whom I disagree on almost everything, let alone individualized inference. Are you turning this discussion into a book?

5.16.2023 11:00am - Here is our weekly report of new articles on causal inference: Note the increased popularity of "counterfactuals" in these studies.

5.16.2023 2:26am - (Replying to @translucentlota and @RashidaTlaib) Ask the victim what they intend to do once the occupation is lifted and see if they deserve the angelic title "victims". Please ask, it's not a rhetorical question. I have asked, many many times. Ask @RashidaTlaib , Ask Palestinian intellectuals at your university. Ask BDS, etc.

5.16.2023 2:15am - (Replying to @translucentlota and @RashidaTlaib) Do I demand that all Israelis explicitly state Palestine’s “right to exist”? Absolutely. This has been my mantra since day-1: "Equally legitimate and equally indigenous". It has been echoed by all my Israeli friends (though I can think of some who won't), not by one Palestinian.

5.16.2023 12:10am - (Replying to @rogerra43393439) No, I don't know Meni Pe'er.

5.16.2023 12:07am - (Replying to @daoudkuttab and @LahavHarkov) A positive statement? Really? If you deny the Jewish people sovereignty in any borders, aren't you against Jews? You can't "Love Jews" and deny the Shoah, then deny their historical connection to the land. Positive statement would be "Equally legitimate, and equally indigenous"

5.15.2023 10:58pm - Palestinians ask us to take Nakba Day seriously, solemnly and respectfully. I tried, but @AnneBayefsky report, live from the UNGA circus, tells us that the Palestinians themselves do not respect the Day. Disappointment. (reposted:

5.15.2023 10:15pm - So be it. Nakba or no Nakba, May 15 is still Israel's Independence Day and, if the UN won't celebrate it - I will, on my own if needed. Wait, why on my own? Join me in singing Maluu Assameinu Bar, with the one and only, LA Shir Hebrew Choir: (reposted:

5.15.2023 9:22pm - (Replying to @johndebois and @HenMazzig) Don't put words into the mouths of people you don't know. No Jew has ever concocted a story to displace Palestinians. Read about the early Zionists and their pleadings for co-existence: (posted:

5.15.2023 9:01pm - (Replying to @levantophile and @hahussain) I believe Israelis are ready for such overtures, including symbolic acknowledgement of the Nakba, as long as the overtures are mutual. Here is a realistic arena where two-sided reckonings would be possible: (posted:

5.15.2023 8:27pm - Timely reading on both @NakbaDay and Israel's Independence Day. But asking Arab intellectuals to consider co-existence is asking the impossible, because, for them, it means an admission that the past 100 years of war was a dumb mistake -- a traumatic confession. (reposted:

5.15.2023 7:02pm - (Replying to @jeremy1174) You ignore one possibility. I have explored at length the question "why do they think the way they do?' and have found a satisfactory answer. Have I passed the test of "charity", or should I continue to ask?

5.15.2023 12:55pm - What I find obscene about the UNGA "special commemorative event" is that, after all, May 15 is also known as Independence Day to 8 million human beings. Doesn't it deserve some UNGA commemoration? (reposted:

5.15.2023 12:17pm - The UN is distorting the meaning of the Nakba @UnHerd (posted:
5.15.2023 12:17pm - On Nakba Day, we can understand the Palestinians wish to commemorate their suffering during their 1948 war on Israel. What we can't understand is their wish to repeat it.

5.15.2023 11:51am - (Replying to @Anatomi54244944) Unfortunately the two sides agrees on facts and causes -- they clash merely on interests, which a "value" issue. One side says "We, We, and We" the other says "Me, Me, and only Me"

5.15.2023 11:44am - Among the many respectful truth-seeking readers on this educational channel we occasionally meet an abusive deformity like the one below, a rare but revealing sample of Israel neighbor's mentality. (reposted:

5.15.2023 4:51am - The Corbyn we all knew has a point. The foiled Nakba that Palestinians attempted in 1948 did not end after all; they have not given up the dream of destroying Israel and are now proclaiming their intention openly and loudly. Today, we renew our calls for a threat-free Israel. (reposted:

5.14.2023 10:44pm - A friend shared this Opinion piece in the LA Times : "Can today's AI truly learn on its own - Not likely." I am posting it, not for it's analysis of "true learning" but for its many examples of "emergent properties". (posted:

5.14.2023 10:17pm - The IDF uniform adds to her grace and warmth; it reminds the viewers that there are real human being in Israel, creative, assertive, and demanding a day of normalcy. I wore this hat 1953-1956 but, no comparison, I was not as charming. (reposted:

5.14.2023 9:57pm - 75 years ago, Israel gained independence. How did the media cover it? (posted:
5.14.2023 9:57pm - Beautiful article by Forward, documenting how world media covered the creation of Israel, May 14, 1948. I am more interested, however, in how world media covers Israel's 75th birthday in 2023; having observed a weird pathology in 2008 -- @Ostrov_A (posted:

5.14.2023 9:11pm - (1/2) Today is May 14, 2023. On Friday evening, 75 years ago, we heard the radio broadcast: We have a State of our own, called Medinat Israel. There were few celebrations in the streets, much less than on Nov. 29, 1947, because the declaration ceremony was not announced in advance (reposted:
5.14.2023 9:11pm - (2/2) for fear of British interference, and the number of invitees was limited. Still, we knew that a new chapter had opened in the history of nations, and that we were fortunate to be part of it. The Egyptian war planes that appeared the next day did little to shake this feeling.

5.14.2023 8:39pm - ( Replying to @TPrincato @DeborahProudJew and 7 others) All nice and rosy, missing one tiny ingredient: Israeli stubborn wish to live.

5.14.2023 8:35pm - Each year, in celebration of Israel's birthday, newspaper editors feel a weird urge to invite guest writers to tell us why Israel should not exist. I haven't taken a tally this year, but I have done so in 2008, see I suspect the urge has not subsided (posted:

5.14.2023 7:58pm - Important to add: It took Truman, not days, not hours, but 11 minutes to recognize the new State of Israel. I would like to believe that President Biden would do the same, despite the morally deformed fringe of the Democratic party, @RashidaTlaib @BernieSanders @IlhanMN etc (reposted:

5.14.2023 7:34pm - I take seriously the idea of Israel becoming a Causal-AI superpower. While we in the US are debating how many trillions of neurons Deep Learning would need to emulate human-level intelligence, causal-ai will examine what it takes to climb rungs-2&3 of the Ladder, and climb it. (reposted:

5.14.2023 7:25pm - (Replying to @AsraNomani) Happy mother Day, and hug her for me.

5.14.2023 1:38pm - (Replying to @AleksanderMolak @ShalitUri and @ehudkar) Agree. Plus, Israel has three ingredients necessary for Causal-AI to flourish: (1) Intellectually curious researchers, (2) Startup oriented students, and (3) Significant geographical distance between researchers and their old US mentors.

5.14.2023 1:21pm - (Replying to @antogeo_) So, you fall for these posters, swallow the slogans behind them, and see no problem with the mindset of those who call them commonsense?

5.14.2023 12:50pm - When slogans of Jihad become posters in the West, we know we have problems with the mindset here at home. (reposted:

5.14.2023 1:10pm - (Replying to @kareem_carr) It is for this reason that we use Simpson's Paradox as the starting point in Primer I don't know of a better way to demonstrate the need for enriching statistics with causal vocabulary. It's also the reason why traditional statisticians shun it.

5.14.2023 1:15am - Maayan, you will be elated to hear that the #Bookofwhy has been translated into Hebrew, and is scheduled to be in print Sept. 2023 (Kinneret Publ.). Your beautiful illustrations will match the new sound of the text. The book, I hope, will turn Israel into a Causal-AI superpower. (reposted:

5.14.2023 7:37am - (Replying to @rogerra43393439) Please do. They started with "NOgdey Metsiut" which I changed to "Nogdey uvdot" - I'd welcome any improvement.

5.14.2023 5:01am - (Replying to @LynnLynzal and @Jerusalem_Post) They are usually inspired by the bible, Unicorn, unfortunately, has been criticized for having no meaningful lyrics.

5.14.2023 1:38am - (Replying to @kareem_carr) Glad you've found "Understanding Simpson's Paradox" to be instructive,, I have been using this paradox to separate between pre-revolution and post-revolution statisticians. The former insist that the paradox has nothing to do with causation.

5.14.2023 1:23am - (Replying to @DeborahProudJew @WRAL and 6 others) They look and sound like Islamic Jihad, same logic, same mentality.

5.13.2023 3:51pm - (Replying to @jonnyjkane Not blind, but tuned to intended violations first.

5.13.2023 1:53pm - (Replying to @EthanWilley7 @academic_la and @RepRashida) Those who accuse others of "ignoring" Nakbaic events of the past will do well spending one minute listening to the "Nakba of the future", planned and choreographed by @RepRashida , while ignored by her sympathizers in the West, including @BernieSanders and his Ex-Jewish followers. @timesofisrael

5.13.2023 1:37pm - Live Update: National security adviser confirms ceasefire, says Israel will ‘defend itself’ if attacks continue via (posted:
5.13.2023 1:37pm - I love the way Israel formalizes the new conditions for cease fire: "Quite will be met with quiet". In other words, a violation of "quite" is a crime, and will be dealt with as such, including the elimination of the perpetrators. When has this equation started to be questioned?

5.13.2023 8:42am - I would put it differently: If you shout "Free Palestine" and care not whether this shout calls for the erasure of her neighbor, your problem isn’t with "freedom", but with a basic human rights violation. (reposted:

5.12.2023 4:49pm - Scholars expound on Mount Ebal curse tablet with oldest Hebrew text via @JNS_org (posted:
5.12.2023 4:49pm - The Curses of Mt. Eival always scared me, as a kid -- we read them in school, at least once or twice a year. Now they found them in the oldest Hebrew text ever found in the Land of Israel. I'm no longer scared. I'm going to read them again, tonight!

5.12.2023 4:35pm - (Replying to @JOllgren) I believe I read it, and couldn't make up what is gained by thinking "exchangeability" instead of modern terms (eg ignorability, admissibility etc.) Is it more cognitive-friendly? Weaker assumption? Computationally more powerful? Testable? In short: why revive a pre-SCM term?

5.12.2023 3:54pm - It's the same trick they played in #USC, see They appointed a "Committee for Jewish life", without the participation of the 85 Zionist professors who complained and, notably, headed by VC of religious affairs. Religion is today's license for bigotry. (reposted:

5.12.2023 5:25am - A roundup of incidents, good and bad, happening on college campuses. (posted:
5.12.2023 5:25am - News from the 2nd front - college campuses. Lacking iron dome, we must fight bare hands. "We will prevail", a neighbor told us on May 15, 1948, as we were cuddling in a bomb shelter under Egyptian air bombardment. I havn't forgotten it, and now I am saying it to my students.

5.12.2023 1:51am - (Replying to @jeremy1174) I refuse to believe that you asking this question seriously.

5.12.2023 1:08am - Same story again, when you complain about antisemitism they nominate an Imam to head a "task force" on religion and after-life. I told you to complain about Zionophobia. (reposted:

5.12.2023 1:00am - I am willing to bet this clip was produced by Al-Jazeera, or Islamic Jihad, not @amnestyusa (reposted:

5.12.2023 12:47am - A shameful apartheid wall. (reposted:

5.12.2023 12:10am - Reminds me of Judgment Day; Oh, how I wish to see the reception in Heaven: Oh, when the Saints go marching in, Oh, when the Saints go marching in Oh Lord, I want to be in that number, When the Saints go marching in (reposted:

5.11.2023 5:19pm - Not convincing. All they have to do is read Al-Jazeera and learn how Israel is committing cardinal crimes even under deadly rocket fire. (reposted:

5.11.2023 2:05pm - (Replying to @MichaelALewis10 @KevinZollman and 11 others) But do you agree with the basic premise that researchers should think Trials (rung 2) as opposed to think natual counterfactuals (rung 3)?

5.11.2023 12:40pm - CNN anchors do not do it out of malice. On the contrary, they believe Israeli lives are a "controversial" issue, hence a good journalist must not take sides. (reposted:

5.11.2023 1:38am - (Replying to @john_c_hawkins and @amanpour) This is my concern, yes.

5.11.2023 1:37am - Inspired by Hacking writings, I feel an urge to tell historians/philosophers that someone of Hacking's caliber should write the glorious history of causal thinking. The #Bookofwhy has several good chapters, but it's written by a Whiggish historian -- time for a fresh look. (reposted:

5.11.2023 1:25am - (Replying to @anzarmlateef and @BernieSanders) That's precisely what Israelis believe you want, -- a Palestinian state instead of Israel - and what they vowed to resist tooth and nail, despite the euphamism.

5.11.2023 1:18am - (Replying to @JamesSteeleII @rlmcelreath and 4 others) The secret to understanding moderation, DAGs or no DAGs, is to ask yourself "who cares", namely, what kind of research questions would we answer by quantifying "moderation." The rest is just math: express the question mathematically and you're done, this is "moderation".

5.10.2023 11:51pm - (Replying to @bullfrog35 and @BernieSanders) They can't get it to their head that, just like Palestinians, Jews are a people, a nation, not just religion, and will not return to Dhimmitude no matter what nice words one uses to describe it.

5.10.2023 10:37pm - What's wrong with the Nakba event @RepRashida is celebrating is that she tries to erase two related Nakbas: (1) The one attempted and foiled in 1948 and (2) The one she and her compatriots are openly espousing to execute against Israel if, God forbid, given the chance. (reposted:

5.10.2023 10:13pm - In memoriam-- Ian Hacking (1936-2023) He is credited with bringing a historical approach to the philosophy of science. #probability #causaltwitter #Bookofwhy (reposted:

5.10.2023 9:48pm - (Replying to @soboleffspaces) Sorry to hear of his passing. A brilliant historian.

5.10.2023 9:41pm - With a few tiny fixes it would apply to Palestinians as well. (reposted:

5.10.2023 9:28pm - These groups and individuals would swear in their mothers' honor that they do not espouse antisemitism, and Nancy Pelosi would vouch for them, with @BernieSanders approval. What disqualifies these groups is their inability to utter the word "co-existence", ie, Zionophobia. (reposted;

5.10.2023 8:25pm - Here is my choreography of an event for mutual recognition of past tragedies and past mis-deeds: Dedicated to Congresswoman @Rashida 's Nakba event. A permission is hereby given to Al-Jazeera to republish this article in full. (reposted:

5.10.2023 7:59pm - (Replying to @erikbiz) Agree on both counts.

5.10.2023 7:39pm - (Replying to @slowhoneybee) I bet Al-Jazeera covered it A to Z. (reposted:

5.10.2023 7:19pm - Congratulations to @vardi for being elected Foreign Members of the Royal Society of London You missed the coronation with three days. (reposted:

5.10.2023 7:19pm - Israel 'catastrophe' event hosted by Rashida Tlaib canceled after outrage (posted:

5.10.2023 5:04pm - Just called my sister in Tel Aviv. She can't get down to the bomb shelter in time (hip fracture) so she's given up, decided to stay in her 5th floor apt. and hope for the best. Rockets can be merciful, even when people don't. (reposted:

5.10.2023 4:53pm - ‘Hateful Material’: EU Demands Freeze of Palestinian Education Aid Over Antisemitic Textbooks Funny, the European Parliament is asking Palestinians to change the DNA of their national identity. Hallucinatory. Next they'll ask them to say "co-existence". Impossible. (posted:

5.10.2023 3:21pm - New papers on Causal Inference posted in the past TWO weeks: Can't keep up with the breakthroughs

5.10.2023 2:00pm - OOPPS, this is the right link to the slides, but several slides did not properly transfer. Please forgive. ps. it was 2010, not 2008. (reposted:

5.10.2023 12:54pm - (Replying to @raittioe @smueller and @_MiguelHernan) Is there any question about incomplete adherence that is not answerable from this analysis:

5.10.2023 11:56am - Readers asked if I can share the slides of my 2008 lecture at Georgia Tech: Yes, I've found them: I love the lecture so much, that I'm going to use same slide next week at Rand. (posted:

5.10.2023 11:28am - (Replying to @raittioe @smueller and @_MiguelHernan) Works of @_MiguelHernan are always worth looking at, but I don't see an analysis of harm and benefit in the ones you posted. Can you pinpoint?

5.10.2023 11:22am - Just finished listening to Yoav Shoham seminar on "Understanding understanding": . He does not subscribe to our formal & performative definition of "understanding" (capability to climb all 3 rungs of the ladder), but he didn't offer an alternative either. (posted:

5.10.2023 3:01am - Listening to Jeff Hinton on @Amanpour show he seems disappointed with DL. But he attributes DL deficiency in understanding to insufficient neurons, compared to humans, not to the absence of world models. Are we heading for another DL spring? (posted:

5.10.2023 2:02am - (Replying to @RCkadt) No, we share the same culture of caring and minimizing suffering of all people. We differ in how we count the effect of Never Again: I include the people whose lives were saved due to the Never Again determination; I don't think you do.

5.9.2023 8:14pm - A wise decision by @SpeakerMcCarthy , but it deprive @RashidaTlaib from reaffirming in public what pro-coexistence folks have been claiming for years: Palestinians cannot depict a future that does not entail the demise of their neighbors. @JGreenblattADL @DavidSuissaJJ @Ostrov_A (reposted:

5.9.2023 8:00pm - I think we should thank Berkeley Law School for this monumental achievement. Their decision to bar Zionist speakers made students think if they really wish to have a "Berkeley" sign on their diploma. (reposted:

5.9.2023 6:36pm - (Replying to @academic_la) This is only half the answer, @IslamEssa_Gaza . I wouldn't blame you if your government is terrorist. I would blame you if your heart and mind is with your government and, if it isn't, for not telling us how they differ.

5.9.2023 6:07pm - My heart goes to the thousands, perhaps millions who will be sleeping tonight at a 30 seconds distance from a bomb shelter. And may the spirit of the Lag-BaOmer Holiday strengthen the hands of all those who defend them. (reposted:

5.9.2023 5:52pm - (Replying to @richard_landes and @Mivasair) The Rabbi has a point; why should we listen to what the BDS says its mission is, when it is so ugly and racist, why don't we just imagine what they should say and life would be so much nicer, ecumenical and out-of-boxish.

5.9.2023 5:26pm - (Replying to @HeynLamb and @Ostrov_A) Those who count bodies and ignore intentions are bound to increase the former, regardless how careful the count.

5.9.2023 10:57am - The Globes article will not be complete without mentioning Israel's first computer, The Golem, built in 1953 by my UCLA colleague and mentor, Jerry Estrin. See (reposted:

5.9.2023 10:46am - (Replying to @AxelMerk) Oh, you are right, I also have this book on my shelf and, you would be surprised, it makes me feel schizophrenic: Was this author me? Where did he get the time? How blind can an author be? etc.

5.9.2023 10:40am - (Replying to @EpiEllie @LuisaBorrell and 2 others) To make me and other readers "see it", please point to it in the context of Theorem 2 and Corollary 1 of, where the identifiability of Natural effects is demonstrated.

5.9.2023 5:57am - (Replying to @EpiEllie @LuisaBorrell and 2 others) I don't see a 5th assumption in the identifiability proof of the Natural Effects here: But I am always open to new discoveries.

5.9.2023 4:02am - (Replying to @LuanaGoriss and @JAsayuel) What is even more interesting is that I didn't know about the existence of "The Jewish Book of Why". My loss, I'm sure.

5.8.2023 9:30pm - The rain in Israel is the only "Colonized Rain" known to mankind. It's known for producing good neighbors. (reposted:

5.8.2023 2:43pm - I am in favor of a Mahattan Project aims to develop "controllable AGI". I see two hindrance: (1) Big corps who would rather sieze control of the uncontrollable, (2) Deep Learning community, who would rather not re-start doing AGI from basic scientific principles. (reposted:

5.8.2023 2:04pm - (Replying to @chuluutiinam) The Hebrew translation read: Notes from the House of Death. I am not there yet. My trenches are merely background for optimistic poems.

5.8.2023 1:57pm - The absence of Jewish representation is endemic to all so called "Equity, Diversity and Inclusion" projects in academia. Glad this Orwellian scandal hits the news. (reposted:

5.8.2023 10:06am - (Replying to @whysomanykaykay) This is called the "Mediation Fallacy" which is mentioned 6 times in #Bookofwhy. The classical example is Berkeley's sex-discrimination .

5.7.2023 11:56pm - Holy Mackerel! A reader sent me a video of a talk I gave 15 years ago and, by gosh, it is so much better than the talks I give today, that I can't resist sharing it: It doesn't cover transportability, true, but the conceptual framework is crystal clear. (posted:

5.7.2023 8:59pm - (1/2) "Be loud and proud in our Jewish and Zionist identity!" This is what I tell my audience too. However, I also tell them to think twice before complaining about antisemitism. There are better weapons around, see and, judging by the many Zionophobic (reposted:
5.7.2023 8:59pm - (2/2) reactions to Noah's video, you can see why these weapons are more effective. No one would get the nerve to accuse us of "conflating" anti-Zionism with antisemitism; we don't. We deal with the former as an uglier form of bigotry, more dangerous than Islamophobia.

5.7.2023 7:17pm - (Replying to @joshuanbarbozam) Please stay the course, we need every rational philosopher we can find to fight the heathens.

5.7.2023 7:05pm - If I were to write another book, I'd title it: "Poems from the Trenches". Anyone knows an enlightened, risk-taking publisher? (reposted:

5.7.2023 6:42pm - Readers familiar with my critics of economists' tribalism vis a vis Causal Inference,, would be illuminated by this discussion with J. Heckman about tribalism within economics itself, and how it affects what economists do and write:

5.7.2023 4:37pm - Beautiful! It's the only passport in the world written in the language of Abraham, Moses and the Maccabees. I had to wait a year to renew mine; now I feel connected. Congratulations! (reposted:

5.7.2023 5:17pm - (Replying to @lostinio) I've just heard President Biden taking pride for the upbringing he has had, his Irish heritage, etc. Humans need to communicate their uniqueness, and group identity is the quickest method available to us. A passport is labled group identity which has gone through some scrutiny.

5.7.2023 5:01pm - Pain is intersectional, except Israeli pain. Grandmothers are intersectional but, in @RashidaTlaib 's mind, there aren't any in Israel. There aren't even people there praying for a day of normalcy. Rep. Tlaib tries to bring this bigotry to the floor of the US Congress. (reposted:

5.7.2023 3:33pm - Alerted by @ArnoRosenfeld , I hereby modify my observation to read: Hebrew is the only "colonizer language" known to mankind which was created in the "colonized" land prior to the "colonizers" arrival. Even Roger Waters and @RashidaTlaib should agree to this version. DM them. (reposted:

5.7.2023 11:16am - (Replying to @ArnoRosenfeld) We need to ask our archeologist colleagues if any evidence exists that Afrikaans was spoken in South Afrika before Europeans arrived.

5.7.2023 3:08am - Addendum-1. Chanukah is the only "colonizer holiday" known to mankind which commemorates a victory that took place in the "colonized" land itself, not outside it. Can you share Addenda - 2, 3, 4... (reposted:

5.6.2023 2:54pm - Hebrew is the only "colonizer language" known to mankind which was created in the "colonized" land itself, not outside it. (reposted:

5.6.2023 10:03am - Reminder, BHL today at the Landmark, 7 pm. (reposted:

5.6.2023 9:53am - Glad to see the logic of Evidence Based Medicine (EBM) put on trial. (reposted:

5.6.2023 9:32am - The first paragraph of "Why I am only a half Bayesian" summarizes these qualitative lessons fairly well. (reposted:

5.6.2023 1:45am - (Replying to @nasim_rahaman) Are we using a different Chat? Mine is And I can't get him to change from "correlation=0"

5.6.2023 1:30am - "In summary," says ChatGPT4, "the correlation between the two coins, given that the bell rang, is zero." It seems that we can make it think like a statistician, an economist, an epidemiologist, etc. and see how those folks approach problems. It's called "psycho epistemology." (reposted:

5.6.2023 1:18am - What a smart Chatbot! I changed "causal" to "statistical", "Pearl" to "Fisher" and look what I got: "the outcome of one coin does not affect the outcome of the other coin. Therefore, the probability of each coin being a HEAD is 0.5, regardless of the outcome of the other coin."

5.6.2023 12:18am - Can you share the prompt that made GPT4 arrive at the correct answer, as opposed to the answer I got: "This means that the correlation coefficient between the two coins is zero, both before and after the bell rings".

5.5.2023 9:37pm - (Replying to @AngeloDalli) Yes. If it employs a "causal discovery" program, like IC, the DAG discovered in our case is unique (assuming no unmeasured confounders). See Causality chapter 2.

5.5.2023 9:30pm - (Replying to @AngeloDalli) This seems like the right conclusion.

5.5.2023 9:25pm - (Replying to @ShawhinT) What makes a system "complex" vs. "simple"?

5.5.2023 9:38am - (Replying to @NonInferior) Thanks for clarifying. Kahneman's "what you see is all there is" is a name given to a fallacy of human thinking, not to a principle of rationality one should adopt. Now it coheres with the principle of "ropes behind the data" in #Bookofwhy

5.5.2023 9:26am - Thanking @ADL for speaking out against the anti-Zionist character of Quds Day celebrations. One comment: Calls for the destruction of Israel need not "devolve into antisemitism" to be condemned as racist and genocidal; they are despicable on their own (de)merit. @JGreenblattADL (reposted:

5.5.2023 1:36am - The full story of how the murderers of Rina, Maia and Lucy Lee were sent to their rightful place. I'm sure @RashidaTlaib is rejoicing today the victory of divine justice, but I haven't checked on @IlhanMN . (reposted:

5.5.2023 1:05am - (1/3) Readers trying to teach GPT4 the secrets of CI should note that it has difficulty understanding a simple collider: Q. Consider two coins and a bell that rings whenever any of them comes up HEAD. What is the correlation between the two coins, given that the bell rang? Ans.
5.5.2023 1:05am - (2/3) As we assume that the coins are fair and independent, the correlation between them before and after the bell rings is zero. Interestingly, the explanation given is causal: "The outcome of one coin does not affect the outcome of the other coin, so there is no relationship or
5.5.2023 1:05am - (3/3) or association between them. This means that the correlation coefficient between the two coins is zero, both before and after the bell rings". This is what we'd expect from a 7 yr old: Explanations must invoke causal intuition, but the connection to correlation is erronous.

5.4.2023 7:28pm - (Replying to @ThomasE13297605 @YairWallach and @pythiccoder) So, the Azzam Pasha "extermination" quote of Nov 1947 does exist after all. But does it really matter, considering that even if they only meant to subdue the Yishuv back into dhimmi-hood, and the Yishuv was determined to resist it bunker to bunker, the result would be the same.

5.4.2023 5:17pm - (Replying to @ddayan) I had the same thought in mind. @YairWallach knows that the moment he speaks facts, SOAS will fire him. I once spoke there (2007?) - a snake pit carved in deceit.

5.4.2023 1:58pm - Indeed. Azzam Pasha (Sec. Gen. Arab League) never said: "pushing to the sea". All he said: "a war of extermination and momentous massacre which will be spoken of like the Mongolian massacre and the Crusades". I was 11, old enough to tell "massacre" from "pushing". Comfort me. (reposted:

5.4.2023 1:35pm - What a great way to start the day -- it makes you feel everything is possible! It is! (reposted:

5.4.2023 10:23am - (Replying to @DavidDeutschOxf) I had a similar experience with chatgpt4. I asked if such data would justify arguing that inoculation is more harmful than beneficial, and it answered that in the 19th century medical practices were not as advanced as they are today.

5.4.2023 6:08am - It sounds like some journalists truly wish to believe that Israel is thirsty for Journalists blood, and would use any shady evidence to affirm their wish. To me, that does not seem appropriate on World Press Freedom Day. (reposted:

5.4.2023 5:42am - Putin's father was confident too. (reposted:

5.4.2023 5:36am - We were told in kindergarten that retribution is an ugly non-constructive act. We were also told that justice is to be pursued. This would generally make one confused -- not today! Not when you look at the faces of Rina, Maia and Lucy. (reposted:

5.4.2023 12:41am - A very profound observation, despite its darkness. I interpret it to mean: If God forbid Israel goes down, world Jewry will never be the same as it was - having touched the fountain of life, albeit briefly, will forever change how Jews would view their place in the cosmos. Huge (reposted:

5.3.2023 11:36pm - (Replying to @academic_la) "There is no other hand" said Fiddler on the Roof. A guy that intends to boycott Israel is like a guy applying for restaurant license and refusing service to blacks.

5.3.2023 11:13pm - If you weren't able to attend Yossi Klein-Halevi Lecture, “Israel at 75: A Crisis of Identity” presented by Yossi Klein Halevi, a recording of the event can be viewed here: Don't be frightened by the pessimistic sound of the title. (posted:

5.3.2023 10:27pm - Scourge or Success: How the Media Commemorated Israel’s 75th Birthday (posted:
5.3.2023 10:27pm - Many of our readers will be interested in this panoramic view of how the world/media/bloggers commemorated Israel's Independence Day -- from celebrations to bemoaning's

5.3.2023 6:41pm - Congratulations to @vgcerf on this month long recognition for laying down the foundations for the internet. When Vint received his PhD (1972) I was new to UCLA, but I recall 2-3 memorable discussions, perhaps because they were lively (a euphemism for stormy). (reposted:

5.3.2023 6:18pm - (Replying to @vardi) You were too nice to this Zionophobe guy, Michael Kardoush, who pretends he never heard of Azzam Pasha, and that the Palestinians had no lobbyists at the UN in 1947. He'll never forgive you for suggesting "mutual recognition" -- a death threat for every Palestinian patriot.

5.3.2023 5:30pm - (Replying to @leilhol and @Ostrov_A) Another exhibit for the International Museum of Zionophobic Mentality, this time an absurdity believing that Israel wishes to ruin its reputation by killing journalists. Zionophobes come in all shapes and forms. @CotlerWunsh

5.3.2023 2:41pm - The overriding message of his new film, “Slava Ukraini,” is that the civilized world cannot afford anything less than “total victory” for Ukraine, lest it puts us at risk of another world war. (posted:
5.3.2023 2:41pm - Please join me, and my friend and mentor BHL, in a special screening of “Slava Ukraini” and Q&A with Bernard-Henri Lévy on Sat. May 6 at 7PM at the Landmark in Westwood. See (posted:

5.3.2023 1:55pm - (Replying to @nachik @intelligence2 and 3 others) I did follow this thread, and commented that I would never debate under such title. The one under which I would enjoy debating Mehdi Hasan would be: Is anti-Zionism a form of racism? But he blocked me.

5.3.2023 1:50pm - (Replying to @MichalSabra) People like this not only "exist" on this earth, but they are essential core Zionophobic mentality. I call them "the 'what about ...' vultures.". e.g., "what about Rachel Corie", "what about Dir Yassin". etc. As if one evil justifies another.

5.3.2023 1:19pm - A perfect example of bad counterfactual reasoning: In retrospect, David had an enormous advantage over Goliath. In retrospect, the winner has a provable advantage over the loser. In retrospect, the winner knew he had an advantage, else he wouldn't have dared etc,, @YairWallach (reposted:

5.3.2023 1:04pm - Another exhibit for the International Museum of Zionophobic Mentality. (reposted:

5.3.2023 11:56am - (Replying to @DavidDeutschOxf) This is precisely how our high school teacher introduced us to logic (in 1950) and how he tried to show us that logic can correct causal misconceptions. Unfortunately, he did not have the tools. @kncukier

5.3.2023 10:55am - (Replying to @pmagrass) What we can make of this is that Chat has not learned predicate logic. This deficiency is understandable, and possibly reparable, since predicate logic is 130 yrs old. I'm still playing with propositional causal logic, which is only 30 yrs old, and which Chat hasn't learned.

5.3.2023 4:32am - (Replying to @pythiccoder) I remember Fawzi al-Qawuqji name quite vividly. But I prefer to focus on Azzam Pasha. The former was just a gang leader, the latter a strategic leader of five genocidal armies that attacked us on May 15 1948. Especially I like his apology (to Abba Eban) "we must exterminate you"

5.3.2023 4:18am - A fascinating conversation with @AsraNomani on @InfinityMessage , concerning radical Islam and its unholy alliance with many dubious forces in our society. I've learned things I should have known but didn't stop to ask. Thank you @AsraNomani for bringing them to the open. (reposted:

5.3.2023 4:09am - Thank you, Ambassador @NaorGilon for mentioning our late son, Daniel, on this #WorldPressFreedomDay. His writings, his words, and the way he viewed humanity have become inspirational to Journalists everywhere. We hope to see them become imperative norms of the profession. @AsraNomani (reposted:

5.2.2023 11:19pm - Multi-congratulations to @DaphneKoller on her election to the National Academy of Sciences. Daphne is co-author of "Probabilistic Graphical Models" (2009) and co-founder of Coursera. As we say in Hebrew: Chizzki V'Nitchazek (In your strength lies ours)

5.2.2023 10:53pm - With a couple of refinement the guy would qualify for Al Jazeera anchor. (reposted:

5.2.2023 10:49pm - (Replying to @NelsonDaleSmith @oldbooksguy and 3 others) Honored to be nominated.

5.2.2023 10:41pm - (Replying to @IamHassiLassi) Under an appearance of "objectivity" the video omits the giant elephant in the room: Palestinian rejectionism of any form of Israeli sovereignty, in any borders whatsoever. Rejectionism is the core of Palestinian identity and has been THE major obstacle to peace; it still is.

5.2.2023 8:50pm - Azzam Pasha wasn't a fringe militant; he was the Sec. Gen. of the Arab League, the head diplomat arguing for the Palestinian cause. I was 11, old enough to read newspapers, and to witness Pasha's "momentous massacre" attempted when Egyptian war planes bombarded us on May 15, '48. (reposted:

5.2.2023 8:36pm - One of evolution's greatest achievement has been the development of an eye with a retina, to distinguish the safe from the dangerous. If this is "apartheid", @Amnesty is Pinocchio. (reposted:

5.2.2023 8:11pm - This "debunking" fits Palestinian history re-writing. On October 11, 1947, 5 months before Dir Yassin, Azzam Pasha (Sec. Gen. Arab League) declared out loud: "a war of extermination and momentous massacre which will be spoken of like the Mongolian massacre and the Crusades." (reposted:

5.2.2023 7:51pm - Speaking of AI contributions to humanity, here is one: Separate the safe from the dangerous. (reposted:

5.2.2023 7:42pm - Congratulations to LA Rabbi @RabbiWolpe on becoming an ADL Fellow!! @JGreenblattADL (reposted:

5.2.2023 2:35pm - (Replying to @IamHassiLassi) Israelis, because they were born there (mentally, of course) and have vivid memory of that (historical) childhood, and have proven their ability to resurrect that childhood through miraculous nation building, and, importantly, offered coexistence to their neighbors.

5.2.2023 1:50pm - (Replying to @SarvagyaGupta) #Bookofwhy treats another version of the game.

5.2.2023 1:40pm - (Replying to @soboleffspaces) what is MSAS?

5.2.2023 11:44am - Very cool and thoughtful. There is clearly great potential to using LLM in CI tasks, possibly leading to a new hierarchy and a new ladder, where the input is not "data", nor "experiments", but "texts about data and experiments". (reposted:

5.2.2023 11:18am - I was once interviewed for Al-Jazeera (Doha, 2006); the reporter could not bring himself to say "two peoples". For him, there was only ONE, and "coexistence" meant "War on Palestine" . (reposted:

5.2.2023 10:53am - We must humbly admit that homecoming after 2,000 years is rather unprecedented in the history of nations. We must strongly insist, though, that yearning for 2,000 years is equally unprecedented, and perhaps doubly relevant to what counts as "indigenous". (reposted:

5.2.2023 10:24am - If you valued my thoughts you would chant with me: "Equally Legitimate and Equally Indigenous" and, if Palestinians valued my thoughts, they would start with the latter. (reposted:

5.2.2023 8:02am - Readers who like this introduction may also be interested in the addendum: which neutralizes some of the critics we heard from trialists and clinicians. (reposted:

5.2.2023 7:50am - (Replying to @GeraldNGOM and @NYT) We have chatGPT to compose automatic anti-Israel "reports", no need to ruin the reputation of "journalists" like @satariano @paulmozur & @PatrickKingsley . Poor souls.

5.2.2023 7:32am - (Replying to @kyberknave @m4xim1l1an and @NYT) If a slope is "apartheid" because it can turn "slippery" the earth would be flat.

5.2.2023 7:29am - (Replying to @m4xim1l1an and @NYT) I wouldn't be surprised at all if it is used to track #settlerterrorists and, if not, it certainly should. Glad you asked: "are ALL Palestinians terrorists?" You made my point: We need fast algorithms to let the innocent majority pass through, while @NYT screams "apartheid"!.

5.2.2023 7:50am - (Replying to @GeraldNGOM and @NYT)

5.2.2023 12:48am - (Replying to @tkaiser_science and @smueller) I don't see why the former would present special problems, or a different kind of analysis.

5.1.2023 11:54pm - How I wish I was informed someday that an AI algorithm I've developed in the past is successfully used by Israel to track terrorists and prevent the killing of innocent civilians. If the @NYT counts terrorists surveillance as "apartheid," may God protect their reporters. (reposted:

5.1.2023 9:56pm - Congratulations to @BrumerDelilah on her runner-up-ship for the 2023 "Journalist of the Year" title. (reposted:

5.1.2023 9:46pm - The moral deformity of @amnestyintl is well documented. What's still missing is a credible analysis of how deformities of this magnitude can metastasize organizations with such noble charter. (

5.1.2023 6:48pm - In honor of the #JewishAmericanHeritageMonth that starts today, I am retweeting my speech at the graduation ceremony of UCLA Jewish students, in which I tried to define "that extra color we call 'Jewish'". @noatishby @StandWithUs @DavidSuissaJJ @Ostrov_A (reposted:

5.1.2023 10:43am - Nostalgia. Today, May 1st, International Workers Day, members of my generation remember how they marched through town with red flags, marching bands, singing the International Communist Anthem: Rise up cursed folks.. Hebrew: Kum Hitnaera Am Chelecha ... (posted:

5.1.2023 6:18am - US House passes pro-Israel measure 401-19, but growing number of Dems vote against via @timesofisrael (posted:
5.1.2023 6:18am - J-Street is rudely insensitive; its protest of the resolution's omission of 2-state-solution ignores the feelings of most Palestinians, for whom a 2-state-solution is a death-threat.

5.1.2023 5:50am - A nugget of wisdom is better than a mountain of hype. A thoughtful interview on LLMs by Stuart Russell: (posted:

5.1.2023 5:20am - The high priests of Critical Race Theory (CRT) promised us (on this Twitter account) that this ugly things would not happen in the classroom -- they do! (reposted:

5.1.2023 5:10am - (Replying to @lackboys3 @RachelMiniK and 17 others) With a stellar list like this, I'll be working twice as hard to qualify! #Israel75

5.1.2023 4:56am - "“Today, we celebrate 75 years of vibrant democracy in the heart of the Middle East, Seventy-five years of dynamism, ingenuity, and groundbreaking innovations. You have literally made the desert bloom.” (reposted:

5.1.2023 4:33am - A new 'under the radar' assault. (reposted:

5.1.2023 4:30am - Behold, behold! Not only Pinocchio and Humpty Dumpty, George Orwell too was a Muslim. (reposted:

5.1.2023 4:08am - "This should not be happening" is not the right response to the @guardian happening. The right is response is to ask whether the @guardian is willing to ask "WHY it is happening", as did David Hirsch: (reposted:

5.1.2023 4:03am - (Replying to @DavidHirsh and @guardian) Is the @guardian willing to ask why or not? This is the question.

5.1.2023 3:12am - GPT4 has similar trepidations: Me: Can you give an example of a question which you, lacking true understanding, find it hard to answer while an agent possessing true understanding would be able to answer? GPT4: Certainly, Question: What is the meaning of life? (reposted:

4.30.2023 2:51pm - The one good thing about the new guardian of #hrw: She won't be able to say: "As a Jew, ..." and blame her hatred for Israel on abusive upbringing. (reposted:

4.30.2023 11:14pm - (Replying to @joshualightman @WNewcorn and @KenRoth) The word "sociopath" sounds like an attempt to defame the guy, but I am trying to understand scientifically how such mentality can evolve in an ordinary human brain.

4.30.2023 2:59pm - Replying to @dylanarmbruste3 The definition sounds harmless, but I would be hesitant quoting pre-revolution authors.

4.30.2023 2:39pm - (Replying to @peterhimmelman) My mouth trumpet is flexible, see here: It can give you a C# rooster if needed. Ready too. (posted:

4.30.2023 2:31pm - We are informed of an interesting seminar at Stanford: The title: "Understanding understanding" would probably trigger sparks of curiosity among readers who subscribe to our formal definition of "understanding" (capability to climb all 3 rungs of ladder). (posted:

4.30.2023 1:41pm - Oh President Lula, How the world has changed? In 1947, the Arab presented us a very compelling reason why they can't accept a Palestinian state: "Not even the size of a postage stamp" they said, with zeal and moral righteousness. Has the North changed in their moral compass? (reposted:

4.30.2023 1:11pm - (Replying to @Its_Spriggan) DM me at your risk.

4.30.2023 1:0pm - What is Epidemiology? Two recent JAMA papers deal with this question: (1) Sheldrick's "Randomized Trials vs Real-world Evidence, How Can Both Inform Decision-making? and (2) Lipsky and Greenland's "Causal Directed Acyclic Graphs" (reposted:

4.30.2023 12:46pm - Quoting directly from the UN Resolution 181, of Nov. 29, 1947: "Independent Arab and Jewish States , ...shall come into existence in Palestine ... not later than 1 October 1948." President Lula was 3 years old then, and I was 11, whose memory do we trust? (reposted:

4.30.2023 12:33pm - (Replying to @RabbiWolpe) No, No; the world needs an occasional reminder that The Guardian has "editorial standards."

4.30.2023 10:12am - (Replying to @jazchaz and @POTUS) Is his remark on Danny in this video? I couldn't find it.

4.30.2023 9:37am - (Replying to @MarcioMinicz @_MiguelHernan and @AleksanderMolak) I would put it a bit differently: The knowledge needed to answer most research questions is not in the data. The word "pattern" already connotes features of the data, the word "knowledge", though shunned by researchers outside AI, refers to a conserved commodity.

4.30.2023 9:20am - (1/2) What's unique about the Moment Magazine Symposium is posing the question "What does it mean to be Pro-Israel?" to people who work day and night towards the elimination of Israel, eg, @PeterBeinart , @CecilieSurasky , Aziz Abu Sarah, George Bisharat and others. Their answers (reposted:
4.30.2023 9:20am - (2/2) reveal both the tragically misguided structure of Palestinian mentality, and the distorted mindset of Zionophobic Ex-Jews. Jewish education should be re-structured to counter these two deformities. @NadineEpsteinDC , @EinatWilf , @academic_la , @JGreenblattADL , @Ostrov_A

4.29.2023 8:20pm - (Replying to @Anatomi54244944 @tdietterich and 2 others) The nuggets are causal assumptions together with what can be derived from them. Most interesting Inferences in causal-ai are polynomial time.

4.29.2023 4:21pm - I like @YejinChoinka 's "Dark Matter" metaphor of commonsense, but I think the metaphor of a new dimension (as in Plato's cave) is a better one. The relationship between data and understanding is clearer, and we have a laboratory - causal AI - to test ideas on that relationship. (reposted:

4.29.2023 12:47pm - Glad Moment Magazine chose to highlight my answer to the question: "What does it means to be pro-Israel today?" Here is my answer in full: (posted:
4.29.2023 12:47pm - I should add to it a paraphrased quote from Herzl: "Israel is a homecoming to the Jewish fold before it becomes a homecoming to the Jewish land." (Herzl used "Zionism", not "Israel", in the First Zionist Congress", see (posted:

4.29.2023 11:26am - A colleague of mine is about to testify before a policy-minded Government agency, and is asking for examples where observational studies were used to answer causal questions connected with unemployment and pension plans. Any pointers?

4.29.2023 6:46am - (posted:
4.29.2023 6:46am - Another inspirational reading with which to consummate our celebration of Israel's 75th birthday. @EinatWilf @Ostrov_A @CotlerWunsh @mishtal @DavidSuissaJJ @noatishby

4.29.2023 4:32am - (Replying to @depalan) Not sure I can agree that fighting words like "racism" should remain in their original, narrowly defined semantics of skin-color based "race". Words have a tendency to broaden, so as to embrace wider spectrum of despicable bigotry.

4.28.2023 10:55pm - When I see a cover of a book that I read, I jump with glee to meet other smart readers, but when I discover it's my own, I humbly say: "That's a fairly good book." This one really is. (reposted:

4.28.2023 10:41pm - Inspiring Shabbat reading for the weekend celebrating Israel's 75th birthday. (reposted:

4.28.2023 9:28pm - @CotlerWunsh , Happy birthday to your father, @IrwinCotler , May he continue his relentless fight for decency, reason and human rights till 120. @TheRWCHR

4.27.2023 10:38pm - (Replying to @bentriderdeb) Sounds like Ivan the Terrible.

4.26.2023 11:52pm - I didn't think much about this interview but, listening to it again, I love it -- I love being interviewed by someone who says: "Wait a minute, this is against everything I have been teaching..." How many professors can say that? (reposted:

4.26.2023 1:51pm - (Replying to @DrBobGoldberg) Never give a Zionophobe the benefit of sickness like anti-semitism. Zionophobia is not a sickness, but a well calculated strategy aimed to eliminate a people's homeland. They know what they are doing.

4.26.2023 10:42am - A word of caution, the book is indeed ONLY for those seeking common sense approaches to reasoning. (reposted:

4.26.2023 1:19am - Keep track of this list of 19 names. It would come handy when you get a request to support the re-election campaign of any of these shameless law makers. (reposted:

4.25.2023 7:30pm - Dedicated to Yaron London @tavlool , in friendship and admiration. I hope he still approves of my translation, despite the changing nature of The Chase. (reposted:

4.25.2023 7:19pm - I don't know a better way of ending Israel Independence Day than singing MIRDAF with Chava Alberstein and reading a translation of "The Chase" (approved by Yaron London), below: @EinatWilf @academic_la @SharonNazarian @Ostrov_A @CotlerWunsh @newzionists (reposted:

4.25.2023 6:03pm - The current crisis presents a historic opportunity to refashion the Jewish state in a way that embraces the original aims of the founding: protecting both democratic rights and freedoms and also the unique Jewish character of the state. (posted:
4.25.2023 6:03pm - A thoughtful soul-searching of Israel's history, being, crisis and bright future. "The love in this country surpasses that of any nation on earth."

4.25.2023 4:41pm - Moreover! The gurus of the statistical establishment do not help the inversion, because they belong to the 1/10th who do not know why. They are in fact contributing to the same myth in Machine Learning: "If our method is wrong - we should be penalized eventually and notice it." (reposted:

4.25.2023 3:37pm - (1/2) I've just watched the "torches lighting" ceremony from Mt. Herzl, in Jerusalem and, call me naïve, or call me a poet, I can't understand why the miracle called Israel is not celebrated in every household around the planet; show me a serious educator who wouldn't see it as (reposted:
4.25.2023 3:37pm - (2/2) THE most inspiring role model of what society can achieve, against all odds, when it acts together towards one noble cause. Show me. 4.25.2023 3:28pm - I salute UK Jewry for starting to treat British Zionophobes, not as antisemites, but for what they really are: Ugly Zionophobes. (reposted:

4.25.2023 3:15pm - I rarely respond to tweets with "I absolutely and resolutely agree!" This is one such tweets. Why? Because the number of Statisticians who say: "Why not use Bayes updating?" still exceeds the number of those who know why by a 10:1 ratio. It should be reversed. (reposted:

4.25.2023 9:32am - Variables "just exist", but their values are observed, when we care to look, and events (changes in variables value) can even "occur".

4.25.2023 9:25am - A new week opes, and a new list of CI papers is with us: So many of which deserve attention!! My philosophical instincts led me to: "Reflections on the asymmetry of causation", by J Ismael, which connects thermodynamic gradient with modern accounts of CI

4.25.2023 8:22am - Sorry for the typo. It should read, of course: "to rebel against their stat professors", not "revel". The past two days I met three more statisticians asking: "But why not use Bayes updating on causal models?" Here is why: (reposted:

4.24.2023 12:04pm - (Replying to @DavidHirsh) It's all a result of indicting Corbyn on charge of antisemitism instead of the more serious change of Zionophobia.

4.24.2023 11:30am - As part of Israel Independence Day celebrations at UCLA, Yossi Klein Halevi will lecture on "Israel at 75: A Crisis of Identity". See: @DovWaxman An earlier video, summarizing Yossi's topics of interest can be watched here (posted:

4.23.2023 11:06pm - One of the songs that will be part of the memorial sing-along tomorrow is "Shema Israel - When the heart cries only god hears" Here is the original performance of that song by Uziah Zadok (13 ys) (reposted:

4.23.2023 10:35pm - Tomorrow, as part of the Memorial Day Ceremony Commemorating Israel’s Fallen Soldiers and Victims of Terror, I will light a candle in memory of our son, Daniel, and i'll sing along with other bereaved families, in honor of their fallen loved ones: (posted:

4.23.2023 10:19pm - (Replying to @revprez) Ruth understood that G-d is merely a symbol of a people's oneness; IOW, when you identify yourself as a member of a collective, you automatically adopt whatever symbol that collective uses to represent its oneness. Today, that religion has lost its appeal - the symbol is Israel.

4.23.2023 7:11pm - (Replying to @ChelseaParlett @5_utr and @nickchk) Honored to be on this "had to pick one" list, but I would not characterize #Bookofwhy as a "stat book"; it is in fact an "anti-stat book" in that it calls to stat students to revel against their professors who think causation and write down regressions: (posted:

4.23.2023 - 6:57pm - The way I put it is similar to the way Ruth put it: "Your people are my people, and your G-d is my G-d". Meaning: Peoplehood first, Religion second. Even G-d understood it when He introduced Himself as: "who took you out of Egypt", namely, "If you forget history, you forget Me!" (reposted:

4.23.2023 5:01am - @DavidSuissaJJ is one of the few leaders who understand that Jewish students will never gain respect by merely displaying victimhood; "Zionophobes not Welcome" is their demand for equity and dignity. "Zionophobia and Islamophobia Out of My Campus" is a viable alternative.

4.22.2023 7:37pm - Replying to @newzionists I have hoped you won't stop at complaining about "abhorrent display of intolerance" but continue to demand "Zionophobes Out of Campus!" Bullies respond to push-backs, not to whining.

4.22.2023 9:23am - When I was a student we invited mostly "controversial speakers". When have Jewish students become herd followers to the point of shunning debates and bowing to common denominators? Is that the reputation you wish Yale students to have upon graduation? @MiriamElman @CotlerWunsh (reposted:

4.21.2023 4:16pm - One more rooter here for @BrumerDelilah - Our National Finalist for the 2023 National High School Journalist of the Year. Finger crossed! (posted:

4.21.2023 4:12pm - @KenRoth legacy in Jewish history cannot be un-tarnished with a book. (reposted:

4.21.2023 4:08pm - Hillarious! How different is it from: "But it's the Hamas who are launching rockets at Israel." (reposted:

4.21.2023 3:55pm - US Federal Court Rejects Challenge to Texas Anti-BDS Law (posted:
4.21.2023 3:55pm - To win a law suit against BDS is not hard, but to hear an Attorney General clearly explain why, is unique: "[anti-boycott laws are] increasingly necessary as the radical left becomes increasingly hostile and antagonistic toward Israel"

4.21.2023 1:28pm - As a witness to the hijacking of the Center for Jewish Study at my university ( @UCLA ), I highly recommend Jarrod Tanny's in-depth discussion of the bankruptcy of Jewish Studies as a field and its complicity with rising antisemitism in academia. @JGreenblattADL @DovWaxman (reposted:

4.21.2023 1:02pm - Words can win wars. The appearance of "Zionophobes Not Welcome" posters on US campuses signals the beginning of the Zionist students pushback and the end of Zionophobic racism in academia. (reposted:

4.21.2023 7:27am - (1/2) A colleague asked yesterday for my opinion of James L. Gelvin's book "The Israel-Palestine Conflict; One Hundred Years of War," Cambridge University Press, 2005. It so happened that I wrote a review of this book (unpublished), posted here: It's worth
4.21.2023 7:27am - (2/2) reading primarily for understanding the literary techniques used by many so-called "historians" who pretend to present a balanced account of a topic they feel strongly about.

4.21.2023 5:47am - (Replying to @AlexanderNaume2) But I hope you agree that, in order to strike a balance between personal and societal goals, we need a model for estimating the likelihood and each goal is satisfied by any given action. This is what causal inference is about.

4.20.2023 11:49am - (Replying to @soboleffspaces) Jon Williamson puzzles me. For a minute I thought: perhaps he is truly unaware of SCM, but then I recalled a 2009 paper I published in a volume he co-edited: ; he couldn't possibly escape solutions to problems he poses as "wishful's" . Puzzle.

4.20.2023 4:51am - Sorry, their theme is "Evidential Pluralism", not "pluralistic evidence" as I wrote. The pluralism involved consists of data + mechanism which, essentially, stands for data + causal model, (i.e., SCM) but lacking a formal representation of the latter. (reposted:

4.20.2023 4:29am - Hurray! Our paper on Personalized Decision Making is now available on line: Enjoy the final corrections we made, including the addendum: (posted:

4.20.2023 4:14am - (Replying to @rodakker @_bcalexan and @a_sarentorbic) LLMs are not AI. AI is the attempt to rectify those distortions.

4.20.2023 4:09am - Those who mock their neighbors' miracle of independence raise doubt of whether they deserve one. (reposted:

4.20.2023 3:45am - (Replying to @ShMMor and @Martin_Kramer) The cliché is even more grotesque: Palestinians have resisted 2S tooth and nail and have rendered it impossible so regrettably I have to conclude Israel can't continue to exist. Socratic logic par excellence.

4.20.2023 3:25am - (Replying to @CotlerWunsh) He made his point; he should now retract and join the ceremony, to honor Am Israel.

4.20.2023 3:23am - (Replying to @ThePearlPost and @NorthwesternU) Congratulations! Delila, Stay the course!

4.20.2023 1:26am - (Replying to @JewDefense @ABC and @kugrieshaber) So, please join my calling and urge every Holocaust Museum to dedicate a pavilion to the Warsaw Ghetto Uprising and, organically, to the creation of Israel, 1917-1948, which inspired the uprising and symbolizes Jewish revival. An anti-Holocaust-Survivor.

4.20.2023 1:10am - I've scanned the book and it seems to me that the authors are not aware of the causal revolution. As a result, their theme "pluralistic evidence" remains an ungrounded "theme", lacking footings in models, methods, or algorithms. Disappointment. (reposted:

4.20.2023 12:05am - Evidently, the link provided is broken. Here is a direct link to the article: @Aizenberg55 (reposted:

4.19.2023 11:45pm - I've written a whole article about this misconception: See if you can use some of it in your disputation with @PeterBeinart and other Zionophobes. (reposted:

4.19.2023 5:26pm - (Replying to @rodakker and @a_sarentorbic) Science progresses by refutation, not by confirmation. That we cannot infer a causal graph from observations by any method whatsoever is Theorem #1 in causal inference. Is it a weakness of SCM? No, it's a statement about reality. Like the impossibility of squaring a circle.

4.19.2023 9:15am - Are you serious? Testability is one of its crown achievement.

4.19.2023 6:49am - (Replying to @Queemdius and @a_sarentorbic) You bring up an important point, the history of statistics cannot be told without a causal perspective, and the same can be said about AI. If this is true, then the history of statistics cannot be told by a statistician - I hope Stephen Stigler agrees!

4.19.2023 6:30am - Palestinian 15-year-old arrested for East Jerusalem shooting that injured 2 Israelis via @timesofisrael (posted:
4.19.2023 6:30am - A video of the actual shooting is provided in this article, and I don't know a decent person, Democrat or Republican, who was not waiting anxiously for this arrest.

4.19.2023 6:07am - (Replying to @amyalkon and @havivrettiggur) Likewise, in 1882, following the Odessa pogroms, a fellow by the name of Leon Pinsker published a pamphlet titled "Auto-Emancipation -Warning to His Fellow People, from a Russian Jew" in which he urged the Jewish people to strive for independence, thus launching Zionism.

4.19.2023 5:49am - A paper that endeavors to formulate the assumptions behind synthetic control models in the language structural causal models should be saluted, let alone one that relaxes those assumptions. (reposted:

4.19.2023 5:34am - The story of two instructors who used #Bookofwhy in high school is told here: It's doable, because causation settled in our minds before college algebra. (reposted:

4.18.2023 10:15pm - Dear @ChancellorCUNY . My CUNY colleagues are complaining about Zionist professors systemically discriminated against and excluded by @CUNY and the @PSC_CUNY (see @SAFECUNY) and you are conveniently speaking of antisemitism -- seeking license for inaction. Can you spell Zionism? (reposted:

4.18.2023 9:46pm - Anyone who hasn't heard this siren sound on Holocaust Remembrance Day, in Israel, or hasn't seen these cars stopping in the middle of the highway while drivers stand, heads down, in memory of the fallen, will not understand why Never Again has ceased to be a slogan; it's a fact. (reposted:

4.18.2023 7:33pm - Extremely valuable, factual and insightful article on the historical context of the Holocaust, as a culmination of a century long systematic riddance of European Jewry. The early Zionists of 1880 foresaw it clearly. Another exhibit for Holocaust Museums to consider. (reposted:

4.18.2023 6:12pm - The problem is not that this outspoken lady is allowed to speak as if she is not a racist. The problem is that the world refuses to take seriously what she says and recognize that this is precisely what 99.99% of Palestinians say, since 1929. (reposted:

4.18.2023 12:13pm - Extremely interesting paper on causal models in multi-agent environment: I am retweeting it to readers of this channel to mull over, while I do the same. (reposted:

4.18.2023 8:13am - Here is our weekly list of new publications in Causal Inference Enjoy, explore, and share breakthroughs we cannot live without.

4.18.2023 7:54am - It is for this reason that we should urge every Holocaust Museum to dedicate a pavilion to the Warsaw Ghetto Uprising and, organically, to the creation of Israel, 1917-1948, which inspired the uprising and symbolizes today's meaning of Never Again. An anti-Holocaust Survivor. (reposted:

4.18.2023 7:44am - (Replying to @RacksSound and @hshu22) I haven't seen a Zionophobe able to look in the mirror, once exposed.

4.18.2023 7:11am - Thanks for the mention and, for my records, I am copying the other educational channels you mentioned: @MelMitchell1 , @alepiad , @rasbt , @kareem_carr , @DataChaz , @TivadarDanka , @svpino .

4.18.2023 6:54am - (Replying to @Half_full1) Many many thoughts. The one prevailing says that slogans such as "oppressing" and "their land" are cover-ups for active current ambitions to complete the genocidal aims of 1948. They would otherwise be phrased in terms of more humane aims, say "co-existence" and "shared land".

4.18.2023 4:57am - (Replying to @RacksSound and @hshu22) This is what Zionophobes normally do when someone exposes their racism.

4.18.2023 3:16am - (Replying to @hshu22 and @RacksSound) I wouldn't even invoke the notion of "blame". Unless Zionophobes have invented a new physics, it is a physical impossibility for anyone to realize "never again" without opposing genocidal attacks. It's physics, not blame.

4.18.2023 2:40am - Which makes everyone wonder: Do they really need a provocation to do it, if they could? (reposted:

4.17.2023 10:42am - When I speak at schools, kids often ask me if I am a Holocaust Survivor. No, I say, I am an anti-Holocaust-Survivor. I am part of a generation who vowed to end all Holocausts, then was attacked by 5 armies, on May 15 1948, and miraculously prevailed, to pursue that vow.
4.17.2023 10:42am - Today is #HolocaustRemembranceDay. We remember 6 million Jews - including my grandparents, uncles, aunts, cousins - murdered by Nazi Germans and their collaborators. We vowed never to forget, and our vow is echoed, ringing and amplified in Israel's resilience.

4.17.2023 8:49pm - The number of Arab-speaking Jews who wish to be called "Arab-Jews" is less than the number of Russian-speaking Ukrainians who wish to be called "Russians". I know it, my wife was one. Only Zionophobes label people against their chosen identity. Is @GarySpeddding one? (reposted:

4.17.2023 8:19pm - My limited knowledge of cell cycle models prevents me from fully appreciating this PNAS paper, but its successful use of graphical models to discover causal links from cell development data makes me confident in seeing more CI applications in genetics. (reposted:

4.17.2023 3:34pm - (1/3) AlJazeera editors really think that people in the West are so stupid as to fall for the idea that Israel gains anything by preventing Muslims from praying, or by inciting violence in front of reporters cameras. Trouble is, some Westerners do fall for it.
4.17.2023 3:34pm - (2/3) Aljazeera: "Dome of the Rock saw tens of thousands of Palestinians praying despite occupation's restrictions, attempts to stop them from reaching al-Aqsa on Qadr Night." If true, Israel must be leading such a failing occupation. Quote Tweet📷الجزيرة فلسطين📷 @AJA_Palestine ·
4.17.2023 3:34pm - (3/3) Quote Tweet📷الجزيرة فلسطين📷 @AJA_Palestine ·3hقبة الصخرة تتزين بعشرات الآلاف من المصلين الفلسطينيين برغم قيود الاحتلال ومحاولاته عرقلة وصولهم إلى الأقصى لإحياء ليلة القدر #الجزيرة_في_رمضان #صورة📷24306

4.17.2023 10:42am - When I speak at schools, kids often ask me if I am a Holocaust Survivor. No, I say, I am an anti-Holocaust-Survivor. I am part of a generation who vowed to end all Holocausts, then was attacked by 5 armies, on May 15 1948, and miraculously prevailed, to pursue that vow.
4.17.2023 10:42am - Today is #HolocaustRemembranceDay. We remember 6 million Jews - including my grandparents, uncles, aunts, cousins - murdered by Nazi Germans and their collaborators. We vowed never to forget, and our vow is echoed, ringing and amplified in Israel's resilience.

4.17.2023 12:15am - (Replying to @HuelleAlexander) Now we have two kids shouting: Where is the Emperor's new Model and its Foundation? Do I hear three kids?

4.17.2023 12:10am - (Replying to @GeraldNGOM) This is obvious to every Palestinian and to everyone who has had a chance to talk to a Palestinian. But it's not mentioned at all in the Foreign Affairs article, written for Americans, who are too rational to grasp Middle East irrationalities.

4.16.2023 11:38pm - (1/2) It must be me. I'm reading and reading, page after page, and I don't understand a word of what they are saying. Moreover, "they" count to 108 authors. I feel like Rip van Winkle. Where have I been the past 20 years? Before the FM revolution? When a "Model" meant a model. (reposted:
4.16.2023 11:38pm - (2/2) Or are we watching the Emperor New Cloths? with no child to scream: "Where are the cloths"? Where is the model? Where is the foundation?

4.16.2023 8:42pm - I find this 60 Minutes program fascinating, especially listening how Google's researchers describe the capabilities and limitations of LLMs which we have been discussing here on this channel, eg. adaptability, black-box-ness, hallucinations. A different yet fascinating language. (reposted:

4.16.2023 6:55pm - We are informed that everything under the sun, including AI and LLM, is "Foundation Models" Poor me, still haven't been able to figure out what FM is. Am I the only one left behind? (posted:

4.16.2023 6:00pm - When we ask: "can we generate automatically" we need to specify "from what?". From passive observation? from playful manipulations? From actions and assumptions?

4.16.2023 4:24pm - (Replying to @AlexJohnLondon and @conitzer) I would only change "(unless you already know that)" to " (unless you already know something about it)". what you knew + data = much more than what you knew

4.16.2023 3:50pm - Very insightful observation. Causal reasoning won't scale up, while Machine learning won't solve toy problems. Conclusion = ? Should we trust the latter or boost the former? (reposted:

4.16.2023 11:52am - I wish I was allowed two votes. (reposted:

4.16.2023 11:47am - (Replying to @conitzer and @AlexJohnLondon) You are pointing to one problem: Pharma corruption of data. I am pointing to another: Misunderstanding of causal inference and what it says about EBM.

4.16.2023 11:27am - I often wonder why "Probabilistic Reasoning" was accepted with greater enthusiasm than "Causality" (3:2 in Google citations). My answer: Probability is less threatening to traditional education, allowing all the mistakes to continue unabated. Causality rebels against them. (reposted:

4.16.2023 2:01am - The problem with this "one state" article is not in the facts but with the logic. It goes: "The promise of a two-state solution made sense before the Palestinians proved that they will forever resist it; now that they did, it's no longer viable." Lovely logic: Takes 2 to tango. (reposted:

4.16.2023 1:27am - (Replying to @AndrewHires) There is no consensus on the level of danger, agree, but I believe there is consensus on possible dangers of depleting our resources on technologies that (a priori) exclude understanding, at the expense of those that aim towards understanding. @GaryMarcus

4.15.2023 2:48pm - I assume They are referring to a 1987 paper of mine "Do we need second order probabilities",, to which I devote a chapter in the book "Probabilistic Inference..." 1988. (reposted:

4.15.2023 1:43am - Israel's President meets Holocaust survivors. "I am the girl that came out of the forest, who could not read or write till age 12, now meeting the President of my State. Tell me, Is there a victory greater that this?" (reposted:

4.14.2023 4:01pm - Mighty congratulations to Dr. Ang Li, who has accepted a faculty position at FSU (Florida State University, Tallahassee, FL). We welcome FSU as a new center of CI research, specializing in theory and applications of personalized decision making. See

4.14.2023 3:42pm - Grateful to @smueller for clarifying the relationships between the PoC bounds obtained from multiple experiments vis-a-vis those obtained from observational studies with a known graph structure. Both will revolutionize personalized decision making: (reposted:

4.14.2023 9:25am - (Replying to @LuisaBorrell) I think the emphasis here is on the words "rightfully confused." It's been 30 years now that I am trying to understand what g-formula-g-computation is; I've given up.

4.13.2023 9:28pm - For North African Jews, it’s a night when we turn over our homes from Passover to Mimouna, a joyous family and community-oriented cultural celebration that’s all about blessings, smiles and sweets. (posted:
4.13.2023 9:28pm - A belated Happy Mimouna Day to all my Ex-Moroccan friends who celebrated it.

4.13.2023 3:25pm - Important announcement from the Journal of Causal Inference (JCI): We are looking for an energetic new talent to fill the position of Editor in Chief: If you are qualified and enthusiastic for undertaking this exciting job, please apply before April 30.

4.13.2023 3:13am - (1/2) @fabianbgng 's paper describes clearly why causal mechanisms need be taken into account in defining fairness criteria, let alone estimating degrees of fairness. A word of caution, however, on the use of "matching", which is not a causal notion. #Bookofwhy shows how misleading (reposted:
4.13.2023 3:13am - (2/2) matching can be (page 274) when applied blindly. In general, "fairness" is governed by counterfactual logic, i.e., Rung-3 of the Ladder of Causation. Direct and indirect effects must be defined at Rung-3, using counterfactual notation.

4.13.2023 2:20am - (Replying to @DavidSKrueger @lastpositivist and 2 others) Causality does not "solve" fairness, but it is necessary for solving problems associate with fairness, simply because "fairness" is a causal notion; it cannot be defined nor discussed in any other language.

4.13.2023 12:03am - It's still not the root cause. (reposted:

4.12.2023 8:21pm - (Replying to @suzannTee and @erichorvitz) We had a few intimate sessions.

4.12.2023 8:04pm - I've seen people in deep prayer attacked by less moral armies. I'm sure you have seen some too. I leave it to your judgment. (reposted:

4.12.2023 7:51pm - You can't fight antisemitism without addressing Zionophobia, squarely and explicitly. Priorities have shifted! (reposted

4.12.2023 7:37pm - Whether or not IDF is "the most moral army in the world" is told by the faces of these worshipers, none of whom shows any fear whatsoever when asked and forced to move back a few steps. They know they are perfectly safe, in the hands of this moral yet dehumanized army. (reposted:

4.12.2023 7:13pm - Going by the title alone, "Council on Academic Freedom" may be either an extreme right-wing organization or a radical left-wing movement. However, the authors: Steven Pinker and Bertha Madras render it safe for ordinary Sapiens. (reposted:

4.12.2023 6:31pm - Do two men walk together Unless they have planned? (Amos, 3.3) (reposted:

4.12.2023 4:05pm - (Replying to @noahdgoodman @akatzzzzz and 2 others) I think when you work out the theory the difference between probabilistic and causal dependencies will surface automatically. Recall, conditional independencies are not axiomatizable, whereas counterfactuals ARE.

4.12.2023 1:48pm - You made my day! Now I only wish that one of the authors of LLM would read #Bookofwhy. I just received a proof of the Hebrew translation, in case they need an truly approachable version. (reposted:

4.12.2023 1:42pm - (Replying to @noahdgoodman @akatzzzzz and 2 others) The question you ask in this work goes beyond the causal/probability dichotomy. It deals with going from local judgments to global inference. In what sense do conditional probabilities chain? I think they don't. Can you elaborate?

4.12.2023 1:36pm - (Replying to @ThomSeaton @EinatWilf and @Adi_Schwartz) There are 120 professors on my "strong Zionist" list, who are not too happy about the "Bundist Occupation" of Jewish institutions at UCLA.

4.12.2023 1:30pm - (Replying to @dggoldst and @erichorvitz) I bet you never forgave Ward Edwards for that. He was a straight shooter, and strongly opinionated. It was through him that I met the Tversky-Kahneman team.

4.12.2023 1:24pm - (Replying to @erichorvitz) Did you figure out why gpt4 ordained you and Jack as the authors of the QMR paper? Do you understand why we are spending precious time asking why GPT4 did this or that?

4.12.2023 12:58pm - (Replying to @erichorvitz) The guy is changing his mind. Look what he tells me: 📷 The term "Bayesian network" was first published in the paper "Probabilistic Reasoning in Intelligent Systems: Networks of Plausible Inference" by Judea Pearl in 1988. Doubly wrong: (1) It's a book, (2) 3 yrs too late.

4.12.2023 5:12am - (Replying to @peteratmsr) How do you ask "via Bing"?

4.12.2023 3:32am - (Replying to @EKochba) It takes counterfactuals to stay strong and young at any age, but when you have only 34 yrs left, it comes natural.

4.12.2023 1:11am - (Replying to @eleroymason) Hard core Bayesian statisticians claim there is nothing Bayesian about Bayesian Networks. Why, because they do not assign priors to parameters. Go argue with statisticians.

4.12.2023 1:02am - (Replying to @BillBainbridge5) I don't know much about Control Loops, but why shouldn't LLM access Google Scholar, validate its sources, double check its answers, and save me many routine key strokes.

4.12.2023 12:51am - (1/2) Our slightly belated List of new papers in Causal Inference Having spent many sessions on personalized decisions, our interest would naturally be drawn to the paper on Bounding Probabilities of Causation, which I am still
4.12.2023 12:51am - (2/2) digesting. But do not let me bias you, the list contains many titles that I would love to unpack, given the time and bandwidth needed. Hoping for your help in unpacking.

4.11.2023 11:49pm - (Replying to @taravanand @AJAveritt and @eliasbareinboim) Intriguing !! What is the simplest C-DAG you can draw to demonstrate the problem, its solution and, most importantly, the input information required of the user.??

4.11.2023 11:41pm - (Replying to @entropypromoter) I'm not trying to use it as a replacement to Google scholar. I am trying to uncover its chain of associations, and what type of text makes the chain stronger or weaker.

4.11.2023 11:34pm - I remember my parents reading the newspaper in the kitchen, trying to hide it from me, 11 yrs old, inquiring "who did it?" (reposted:

4.12.2023 11:28pm - One of my most beloved authors. Sad to know he will not surprise us again with a new color of beautiful Israel. (reposted:

4.12.2023 11:23pm - Have you ever tried the word "Zionophobe" to neutralize racists of this breed? It has worked miracles for me. (reposted:

4.11.2023 11:05pm - (Replying to @AvivaKlompas and @FranceskAlbs) You are asking @FranceskAlbs to answer a moral question, as if she is capable of understanding such questions, as if she herself has not emboldened those killings by bestowing automatic impunity to the killers.

4.11.2023 10:25pm - (1/3) My question was: "When was the term "Bayesian network" first published? DPT-4 answered: The term "Bayesian network" was first introduced in the paper "Probabilistic diagnosis using a reformulation of the INTERNIST-1/QMR knowledge base" by John S. Breese and Eric J. Horvitz, (reposted:
4.11.2023 10:25pm - (2/3) published in 1988. The paper discussed the use of Bayesian networks as a representation for capturing the probabilistic relationships between variables in the context of medical diagnosis." This answer has 3 errors: (1) The INTERNIST paper was NOT authored by Breese and Horv.3
4.11.2023 10:25pm - (3/3) (2) The words "Bayesian Networks" are not mentioned at all in the INTERNIST paper. (3) Bayesian Networks were introduced 3 yrs earlier,1985, then 1986, in a highly cited paper (3,200 citations). I assume some paper by Breese was in DPT-4 training set, and mentioned all terms.

4.11.2023 9:58pm - Lyrics: @FranceskAlbs Music: @KenRoth Choreography: Supreme Ayatollah (reposted:

4.11.2023 9:24pm - (1/2) (Replying to @neuro_data) The answer was generic: While I cannot pinpoint the exact related works that led to the incorrect association, it is possible that the prominence of the 1988 paper by Breese and Horvitz in the context of Bayesian networks played a role. This paper, which applied Bayesian
4.11.2023 9:24pm - (2/2) (Replying to @yudapearl and @neuro_data) networks to medical diagnosis, might have been more frequently mentioned or discussed in the training data, leading me to associate it with the first appearance of the term." And btw, the 1988 paper mentioned was NOT authored by Breese and Horvitz.

4.11.2023 4:24pm - (Replying to @Philosofrenzy and @kevinmcld) Of course it can. You and I are perfect proof that it can.

4.11.2023 4:03pm - I took it tongue in cheek, knowing that it meant something else. I am still exploring what it really means, so I can properly interpret what next to expect of it once it says "thank you". Any idea?

4.11.2023 3:47pm - (Replying to @kevinmcld and @Philosofrenzy) Pearl once wrote that "thank you" is the most common counterfactual expression in the English language.

4.11.2023 3:39pm - (Replying to @soboleffspaces) @eliasbareinboim has written a couple of chapters expanding on Page 24 in Causality. Note that do-expressions (not functions) is what we get from reinforcement learning (RL).

4.11.2023 12:57pm - Totally agree. It's the "apparent confidence" with which GPT-4 cites wrong papers and/or authors that blew my mind. It shouldn't be hard to insert an "I think" in front of fabricated facts. (reposted:

4.11.2023 12:46pm - Agree. It is more natural to define DAGs as an abstraction of SCM (with unknown functions), though a stand-alone treatment of DAGs is also possible (as a collection of experimental data), as is done in Causality page 24. (reposted:

4.11.2023 12:20pm - An opportunity for readers to hear, in English and in-person, the cry of 8 million Israelis under two vicious attacks - terror and condemnation. (reposted:

4.11.2023 6:42am - (Replying to @mcxfrank) I was set out to test if GPT-4 can pass our definition of "understanding" and ended up surprised that it doesn't double check its answers, even when it is available in its data base. (reposted:

4.11.2023 6:30am - They claim millions of things but their key, conflict prolonging claim is: Tel Aviv is a stolen land. (reposted:

4.11.2023 6:15am - Opinion | What can Israelis and Palestinians learn from Northern Ireland’s 25 years of peace? (posted:
4.11.2023 6:15am - Many pundits enjoy drawing analogies between North Ireland and the Israeli-Palestinian conflict. Most fail to note one essential difference: The IRA never claimed London was on stolen land -- the Palestinians do.

4.11.2023 3:04am - (1/4) I had a few intimate sessions with GPT-4, which I intend to summarize systematically on my blog. In a nut shell, after GPT failed to find where Bayesian Networks where first published it confessed: I gave you a paper that does not exist, and doesnt mention "Bayesian Networks"
4.11.2023 3:04am - (2/4) because the title was cited heavily by some prominent author in my training text, and because the topic (diagnosis) seemed related. Me: Was it so hard to double check? You do have those papers in your data set? GPT-4: I guess I should have, but I wasn't programed to do so.
4.11.2023 3:04am - (3/4) It confirmed that the way it selects an answer is through a chain of associations and similarities leading from the question to the answer, and that its method relies in transitivity of associations (i.e., A-B and B-C implies A-C). It acknowledged possible violations of
4.11.2023 3:04am - (4/4) transitivity, but failed to give an example of such violation. It thanked me profusely for producing such example (two-coins and a bell), and we both agreed that the reason intransitivity seems unassailable in this example is that it invokes causal, not statistical relations.

4.11.2023 12:59am - (Replying to @ThomSeaton @EinatWilf and @Adi_Schwartz) I've read "War of Return" and I advice all readers to read every word of it, if they wish to understand why the conflict has been going on for 100 years with no resolution in sight.

4.11.2023 12:55am - (Replying to @JohanDH2O @noah_greifer and @RWJE_BA) It may seem like another case of tomahto/tomeito until you try to articulate what the condition is for the validity the g-formula. Try it without the back-door and see how far you can go.

4.10.2023 8:58pm - Retracking, sorry. The murdering ring is being chased in Nablus, not Jericho. Is there a person in the civilized world who is not praying for their quick capture? (reposted:

4.10.2023 6:30pm - Just watched PM Netanyahu on Israeli TV: (1) Yoav Gallant is reinstated as defense minister. (2) All Israel's problems, terror, discontent, etc. are blamed on the opposition and the previous government. Barring the latter, he could have reclaimed his Mr. Smart legacy.

4.10.2023 6:13pm - Whenever a University hires an Ex-CAIR executive, it is doomed to lose some of its most creative faculty to identity politics. CAIR is a hotbed of hatred and intolerance. (reposted:

4.10.2023 1:06pm - (1/2) Readers have asked many times: How can you be sure that, deep inside, Palestinians are not disposed to some sort of peace with Israel? This new post summarizes the key evidence: Denial of any Jewish connection to the land. Nothing could be more telling
4.10.2023 1:06pm - (2/2) We, soldiers of co-existence, have been spending all our energies fighting obstacles such as (1) Palestinian terrorism (2) Israeli settlements, and more, turning blind eye to the one obstacle that dwarfs them all: Palestinian denial of their neighbor. Time to switch lenses.

4.10.2023 12:25pm - I welcome introductory courses in causal inference -- the world needs it badly. I havn't read the content but, seeing the abstract, I can tell that it is not one of those courses titled "causal inference" that ignore the problems of transparency, identifiability and testability. (reposted:

4.10.2023 12:12pm - Request for readers who are proposing expansions to the ladder: Please provide a typical question answerable at layer i that is not answerable by layers j < i . Plus, if possible, how the needed new information is encoded within (or outside of) SCMs. (reposted:

4.10.2023 3:06am - (Replying to @soboleffspaces) Nice graph. I remember how Imbens tried once to convinced me that he can select the right covariates by thinking ignorability -- the envy of us, mortals.

4.10.2023 2:57am - What move diplomats to make statements and say nothing. My gardener can do it at 1/10 the salary. (reposted:

4.10.2023 2:53am - Algerian Radio is adapting a Palestinian mentality. (reposted:

4.10.2023 2:46am - I assume they are searching for the murderer of the Dee sisters. ?. (reposted:

4.10.2023 2:33am - The setup in this paper is the same as in @soboleffspaces work on hip-fractures,, where the proper ignorability condition was derived from the model, rather than assumed a priori (bad habit of PO thinking). @smueller , @EliBenMichael , (reposted:

4.10.2023 2:30am - (Replying to @haaretzcom) So why does @haaretz display the idiotic poster of the first group?

4.10.2023 1:41am - I join @lauder_ronald in applauding the decision by Columbia University to build an academic center in Tel Aviv, thus assuring students and faculty that BDS mentality and Zionophobic racism have no place on US campuses. President Bollinger should make it explicit. (reposted:

4.10.2023 1:15am - (Replying to @rogerra43393439) Professor Fuchs is a long-time pollster. I heard him on Israeli TV explaining how they overcame biases due to Holiday vacation etc. Evidently they group voters by the party they last voted got, then reweigh the results knowing the group size from the last election.

4.9.2023 10:47pm - The reason is that "bound-narrowing" occurs only when observational data are confounded. It is only then that additional information about individual behavior can be extracted beyond RCT data. We show it in (reposted:

4.9.2023 10:09pm - This paper assume strong conditional ignorability in an observational study which, essentially, gives us both experimental and unconfounded observational data. The bounds on P(benefit) and P(harm) are same as in Causality Eq. (9.9), and do not improve by including the latter. (reposted:

4.9.2023 8:18pm - Significant upset in Israel polls: Netanyahu Crashing, Gantz Far Ahead, Lapid Second. Just 20 percent of respondents in the Channel 13 poll say that Netanyahu is doing well as a prime minister, compared to the 71 percent who say that he is doing badly. Opposition gets 64 mandates

4.9.2023 7:42pm - I love it! If we are destined to be controlled by someone, its better be Jews; they at least have a sense of humor ! (reposted:

4.9.2023 1:50pm - "Ani Maamin" ad sung at the funeral (reposted:

4.9.2023 1:45pm - Today I sing "Ani Maamin" (I believe), in the funeral of Rina and Maya Dee (15 and 20), who were murdered on their way home with their mother. (The mother is still fighting for her life.) It means "good will prevail over evil" -- as my grandparents sang in Auschwitz -- it will! (reposted:

4.9.2023 12:53pm - (Replying to @AviBachsh) Are you serious? And I thought she was changing her mind due to my unique power of persuasion. I feel cheated.

4.9.2023 12:43pm - (Replying to @DKedmey) Oh, NO! Robert Malley again? How many times should @USEnvoyIran be proven wrong, before being replaced by someone who understands the Middle East? @omriceren

4.9.2023 10:52am - 'Elimination of Israel' on the horizon, IRGC commander says | World Israel News (posted:
4.9.2023 10:52am - I am mulling and mulling over this announcement by Iranian officials. What could the "collapse and elimination of Israel" be like in the mindset of an IRGC commander? Same as past rhetoric of Ayatollah's, or something totally new: A coordinated 4-front attack? Can't stop mulling!

4.9.2023 9:45am - A truly illuminating time-line of AI history, from Turing (1950) to GPT-4 (2023). Grateful to @mishadavinci for posting this gem of history and for squeezing a seat for me on Mt. Olympus. (reposted:

4.9.2023 1:56am - No! I am telling myself. You can't let this picture of 6 brothers and sisters, victims of the latest attacks on Israel remain on your retina without sharing it with others who perhaps haven't seen their smiles, their hopes, their humanity, and their mourning families. (reposted:

4.9.2023 1:42am - @kenroth has a point. The purpose of the IHRA definition of antisemitism is, indeed, to distinguish Zionophobic racism from legitimate criticism of Israeli policies, to expose the genocidal agenda of the former, and to limit the outreach of its top ideologues, @kenroth being one. (reposted:

4.9.2023 1:01am - Sharing an informal interview I did last week with Adel Daoud's, to Social Science audience. I believe it contains some new anecdotes, absent from my formal lectures, which you may find relevant to the history and logic of causal inference and AI. Thanks! (reposted:

4.8.2023 11:50pm - To all my Christian followers on this educational channel: Happy Easter! The day of Jesus resurrection! Well matching the Jewish Passover - which commemorates our people resurrection. (reposted:

4.8.2023 8:05pm - Sick of seeing Al Jazeera cesspool overflowing social media, I am tuning to @IDF for real-time information on what's happening in Israel. It's amazing how refreshing dry facts can feel when fake News hide the day. (reposted:

4.8.2023 7:25pm - Britain 2023! (reposted:

4.8.2023 5:07pm - Hats off to @FleurHassanN for the courage to be interviewed by Al Jazeera on the latest attacks in Israel. Last they interviewed me (in Doha) they vowed never to do it again. They now quote the UN who quoted Al Jazeera on Israel's wrong doing. Evidence-Based Deceit -- EBD. (reposted:

4.8.2023 4:40pm - We can argue till dawn whether Israel has done its share for 2SS, but one thing is beyond dispute: Israel teaches its children that 2SS is an ideal long-term goal, Palestinians teach their children that 2SS is unacceptable. I have personally experienced the former. (reposted:

4.8.2023 4:00pm - (Replying to @robshookphoto and @hahussain) The state that Israel want to be is no different from Canada or Swizerland, Multi-ethnic but still proud of its history and distinct from its neighbors.

4.8.2023 2:10pm - (Replying to @_gamalan) Sure. GPT-4 started quoting me (albeit sloppily) when trying to convince me that it knows about "confounding", but it didn't realize that Soldier-2 deconfounds the effect of Soldier-1 in the firing-squad. It needed help in getting it, then argued forcibly for it.

4.8.2023 1:47pm - (Replying to @Burning12196087 @L4tkeMuncher and 2 others) Sorry, but when I read what my Palestinian academic colleagues write about Israel's future, this is exactly what the Palestinians are planning to do, given the chance.

4.8.2023 1:32pm - @FranceskAlbs 's logic is missing one tiny ingredient: The so called "people it oppresses" are also those who call for Israel's demise once liberated. @FranceskAlbs is yet to endorse a 2SS under: "Equally legitimate and equally indigenous". We are waiting. (reposted:

4.8.2023 1:16pm - (Replying to @academic_la) The guy is as dangerous as he is nut.

4.8.2023 1:04pm - (1/2) (Replying to @MandE_NEWS) I've asked: "We started with a definition of non-confoundedness in the form of an equation about conditional independence of potential outcomes. Yet when I asked you to check non-confoundedness on a specific problem (the firing squad) you gave me arguments based on diagrams,
4.8.2023 1:07pm - (2/2) (Replying to @yudapearl and @MandE_NEWS) path blocking, colliders, etc. etc. Why did you switch to this language instead of remaining in the potential-outcome framework and convincing me that choosing W = S2 satisfies the independence condition required by the defining equation? " That's when GPT-4 confessed.

4.8.2023 12:49pm - Especially surprised by the ability of a Rung-1 extrapolator (GTP-4) to take text containing instructions on how to climb from Rung-1 to Rungs 2-3, and sometimes (not always) follow those instructions when given text with Rungs 2-3 information. (reposted:

4.8.2023 11:39am - (Replying to @robshookphoto and @hahussain) Trouble is, Westerners expect to hear this from Hamas, but when told that this is also the agenda of EVERY Palestinian in the West Bank, they refuse to accept. "It can't be" they say, "surely, some just want freedom and independence." Asked to name ONE, and they can't.

4.8.2023 5:24am - (1/4) Today I scored my first victory trying to convert GPT-4 into a commonsense Reasoner. To my question: Isn't possible that the two frameworks [potential outcome and DAGs] are equally formal, and saying that one is "a more formal and direct way to assess confoundedness" is
4.8.2023 5:24am - (2/4) merely repeating a myth propagated by careless scientists? GPT-4 answered: You are correct that it is not necessarily the case that a more intuitive framework is less formal or direct than a less intuitive one. In the context of causal inference, the potential outcomes
4.8.2023 5:24am - (3/4) framework and the graphical causal model framework are both formal approaches to understanding and estimating causal effects. My previous statement about one being "more formal and direct" was not accurate, and I apologize for the confusion." A confession I couldn't get from
4.8.2023 5:24am - (4/4) A confession I couldn't get from any humanoid trained in the potential outcome culture. Victory!

4.8.2023 1:48am - I dream of a new Rashida. One who has the courage to tell Al-Jazeera: "Come on, guys, we all know that Israel has no interest in wreaking violence at Mosques - it makes her look very bad on TV. So why don't you come up with more credible stories for what's happening in Al-Aqsa?" (reposted:

4.8.2023 1:18am - (Replying to @ansarikashif7 and @FranceskAlbs) A new causal model for current wave of terror: It is all fabricated by PM Netanyahu. The notion of terror, especially against innocent civilians and Italian tourists is so foreign to Palestinian mind that it had to be fabricated by PM Netanyahu. Readers RT-ing it = believing it.

4.7.2023 3:14pm - (Replying to @AsraNomani @YasMohammedxx and 8 others) I would label Al-Jazeera "deceit-affiliated" network, you don't need a state to be committed to lies anti-Western propaganda.

4.7.2023 10:31am - Guess who is heading this flock of blood- smelling vultures -- of course, it's @FranceskAlbs ! As soon as the number of rockets launched at Israel exceeds 20, she knows exactly who to blame: (reposted:

4.7.2023 2:48am - For a split second I thought @PatrickKingsley of the @NYT is at it again, trying to convince us that Israel has something to gain from violence at the Al-Aksa Mosque. Luckily, this time the picture does not serve his "storming" agenda. (reposted:

4.7.2023 2:13am - Wishing @Cotler a broad and curious audience on her important Northeast speaking tour. (reposted:

4.7.2023 2:07am - Just as Israel's enemies began to celebrate a crack they thought to have found in Israel-US relations, President Biden surprises them with an ironclad rebuff: Stay away, vultures! Go find yourself another prey! (reposted:

4.7.2023 1:43am - (Replying to @stephensenn @Raamana_ and @f2harrell) No comment.

4.7.2023 1:40am - (Replying to @stephensenn @Raamana_ and @f2harrell) This necessary assumption, that "relative bioavailability is transportable from the trial to target population", is what SCM undertakes to enable/verify/test, based on more ascertainable assumptions, like those encoded in the model. We are closer to clarifying the two approaches.

4.6.2023 9:48pm - (1/2) Evidently, one of those horrified by the idiotic implications of @PatrickKingsley is Brad Lander @bradlander , who rush-rushed to tweet: "Horrific. As we prepare for Passover, Jews around the world have the responsibility to make clear that this is not what we mean (reposted:
4.6.2023 9:48pm - by “next year in Jerusalem.” Really? Lander's voting base must be so uneducated and gullible that he finds it necessary to apologize for world Jewry, as if Israel takes pleasure in preventing Muslims from praying. Lander, btw, is on Sarsour's band wagon: (posted:

4.6.2023 9:18pm - Tbh, I too didn't expect such a coherent and thoughtful answers from GPT-4. How did you prompt it? Note however that abilities 6 and 8 have already been captured in do-calculus, and other layers too (eg social intelligence) are arguably within reach. Thanks for sharing. (reposted:

4.6.2023 9:04pm - The closest I get to the "Humean vs. Kantian" debate is in my article on "radical empiricism and Machine Learning Research" I try to avoid invoking names of famous philosophers, simply because they did not have computers on which to base theories -- we do. (reposted:

4.6.2023 3:21pm - The most idiotic words are those implying that Israel has any interest in preventing Muslims from praying at the Al-Aqsa Mosque. Words coming from major media outlets, including @nytimes correspondents in Jerusalem. @PatrickKingsley (reposted:

4.6.2023 2:57pm - (Replying to @DKedmey) This is indeed the question I am trying to answer by playing with GPT4. Whatever the answer, it shouldn't require a "major" AI breakthrough, since counterfactuals have already been algorithmicized.

4.6.2023 11:17am - IDF says Hamas behind rocket barrage from Lebanon; Israel expected to retaliate via @timesofisrael (posted:
4.6.2023 11:17am - Israel is under attack!!! My mind goes to where my heart pounds.

4.6.2023 11:03am - Another empty tweet from a Norwegian politician. Can they ever tweet content? (reposted:

4.6.2023 10:49am - (Replying to @dggoldst and @MattDeakos) I would like to believe that many statistics instructors do rebel seeing how stat textbooks describe Simpson's paradox, but choke the rebellion inside rather than echo Novick's "This is ridiculous!" The stat literature is glaringly deficient of rebels for causal intuition.

4.6.2023 10:10am - (Replying to @stephensenn @Raamana_ and @f2harrell) I am not representing any argument, just sitting quietly waiting to hear what "the purpose" of RCT is, hoping it is somehow related to the clinical questions asked.

4.6.2023 10:02am - (Replying to @ArziARStudio) Is this my latest photo?

4.6.2023 10:00am - (Replying to @f2harrell and @Raamana_) I would gladly switch to another idea, once we adapt another definition of what the purpose of RCT IS, rather than what it is NOT. Hopefully, the new purpose would be related to the clinical research question asked.

4.5.2023 1:15pm - (Replying to @DKedmey @DavidSuissaJJ and 2 others) We use the @StandWithUs Haggadah, it has the picture of Herzl there, and a time-line summary of 3,000 yrs of Jewish history.

4.5.2023 4:52am - (Replying to @blakeflayton) Beautiful answer - symmetry can only go so far. Nakba's victims seek to destroy Israel. Jewish expulsion victims do not seek to destroy Iraq.

4.5.2023 1:55am - (Replying to @taha_crypto1) Songs do not prove that someone ancestors were living somewhere (no one tries to prove that), they prove a vivid, authentic and unfakable historical attachment of today's claimants themselves. Try it on China, it won't work.

4.5.2023 1:21am - (1/2) Two special features will be added to our Seder tomorrow night: (1) A prayer for the State of Israel (especially "bestow Your light and truth upon its leaders, ministers, and advisors") and (2) A moment of silence in gratitude to the IDF soldiers who are freezing cold (reposted:
4.5.2023 1:21am - on their watch, so that their families can celebrate the Passover holiday the way they did when they were toddlers. Join me (Rabbi Yehuda), if you feel likewise; you don't have to be an ordained Rabbi to make the holiday purposeful. @DavidSuissaJJ @bandlersbanter @RabbiWolpe

4.5.2023 12:42am - (Replying to @taha_crypto1 @FranceskAlbs and 4 others) I gave you a test: Please point to one historical event prior to 1900 that you are celebrating in songs or a holiday. It's just amazing!

4.5.2023 12:14am - (Replying to @daoudkuttab) May he live to splash in the Mediterranean together with my grandchildren, when peace descends on us. Amen! Inshallah!

4.5.2023 12:04am - Is it merely ignorance for anti-Zionists to dis-believe that tomorrow night Jews from Prague, Moscow, Geneva, Berlin, New York, London, Warsaw, Athens, Baghdad and Rabat will be narrating the Exodus story as if they themselves came out of Egypt, the house of bondage? Ignorance? (reposted:

4.4.2023 11:43pm - (Replying to @taha_crypto1 @FranceskAlbs and 4 others Not only is it believable, it is testable as well. I gave you a test: Tell us a story of a Mamluk hero, sing an Ottoman song, celebrate a Canaanite holiday. Go to Biblical times and every Jewish child will pass the test half asleep. Tomorrow night we'll pass it blind-folded.

4.4.2023 6:53pm - (Replying to @SayedMhmdAli and @bentriderdeb) I'll walk with you the Straight Path from here to Mt. Olympus, but do me a favor and drop that "Hell Fire" you were using on my friends who love Israel.

4.4.2023 6:22pm - Ingenious. With a few minor fixes, I would be willing to sign this letter. (reposted:

4.4.2023 5:37pm - (Replying to @hartificialman) what was missing?

4.4.2023 5:26pm - (1/2) (Replying to @MattDeakos) Why? I thought every thinking organism would rebel hearing of a drug that is good for me good for women and is bad for a person. Hear what Novick says: "The apparent answer is, that when we know that the gender of the patient is male or when we know that it is female we
4.4.2023 5:29pm - (2/2) (Replying to @yudapearl and @MattDeakos) do not use the treatment, but if the gender is unknown we should use the treatment! Obviously that conclusion is ridiculous. (Novick 1983, p. 45). If a statistician calls a conclusion "ridiculous" wouldn't anyone else?

4.4.2023 5:05pm - (Replying to @abiylfoyp) Nice, the key argument is #3, which a normal person would use without the fancy words of "risk benefit analysis". But, as GPT-4 admits, it's not too good in logic as it is in collecting texts and rumors. Both are needed.

4.4.2023 3:54pm - (Replying to @JoshuaParkhurst and @Ostrov_A) He didn't even say "maybe". All he said was that some foreign diplomats are concerned about such possibility, which is exactly what Israel's Govt and media have been saying for years, i.e., "some outside sources say." Time to accuse and time to heal!

4.4.2023 3:38pm - (Replying to @academic_la) I just ordered 3. One for Passover, One because it's really good and one for BDS's sake. How lucky we are that one key stroke can make us feel like we have made the world a better place.

4.4.2023 4:34am - Thanks for bringing the #BalfourProject to my attention. And, just in case you missed the ubiquitous significance of the Balfour Declaration, here is my reminder: Please read it at the Conference. (reposted:

4.4.2023 1:06am - (Replying to @PrzemekPWr @EinatWilf and 3 others) Glad you asked. I had the same click. Social Intelligence and Political Intelligence should be two AI branches to emerge from causal inference.

4.4.2023 12:46am - Fascinating new idea: Israel's democracy has its roots in the semi-autonomous political life of Jewish communities in the diaspora -- and that's the secret of its resilience. Illuminating!! @EinatWilf @Ostrov_A @DKedmey @academic_la (reposted:

4.3.2023 10:57pm - (Replying to @drstoyanpavlov and @TeddysMom8) I find it very useful as a representative of "a man in the street" (which I very rarely meet). I am disappointed ocassionally when it doesn't argue like "a man in the street." It's still very useful for me to know what shallow internet conversations have been like.

4.3.2023 10:49pm - Monday brings us another bunch of causal inference publications My old fascination with philosophy was reignited by the paper of Vasse, defending Causal Decision Theory. I was surprised though that the debate with Evidential Decision Theory is still on.

4.3.2023 9:28pm - (Replying to @TeddysMom8) That's why I wished GPT-4 would argue: "Risk should be considered, fine, but what about all those who were saved by the vaccine? And all those who would die if we ban the vaccine, even temporarily. " (as we do in #Bookofwhy page 44)

4.3.2023 7:51pm - (1/2) Moving GPT-4 from Simpson to Smallpox. Me: If more people died from smallpox inoculations than from smallpox itself, what's wrong with the argument that inoculation should be banned until it is made safe. GPT-4: You raise a valid concern, and it's important to consider the
4.3.2023 7:51pm - (2/2) risks and benefits of any medical intervention, including vaccination. If it were true that more people died from smallpox inoculations than from smallpox itself, the argument to halt inoculations until a safer method was developed could be considered reasonable." Slippery.

4.3.2023 7:21pm - (Replying to @julioalexo) It is essential to separate the two because, invariably, university administrators take antisemitism to mean "a license for inaction". They designate a "task force" with four philosophers to decide what it is, and that's where it gets buried. See USC

4.3.2023 7:02pm - (Replying to @RobotLaunch) I believe 1Billion is very conservative figure, counting educational budgets, Gothic Cathedrals and industrial training.

4.3.2023 5:34pm - ‘Historic Victory’: US Gov’t Rules That University of Vermont Failed to Address Campus Antisemitism (posted:
4.3.2023 5:34pm - First time we see The US Department of Education Office of Civil Rights (OCR) ruling on "antisemitism and anti-Zionist" harassment and discrimination in a US University. The victory is in seeing "anti-Zionism" spelled out explicitly, not buried indefinitely under "antisemitism."

4.3.2023 2:54pm - (Replying to @yudapearl and @Drombolombo) Why the weighted average argument works in causal setting and not in Simpson's Paradox: See The Sure-Thing Principle:

4.3.2023 2:47pm - (Replying to @Drombolombo) I meant grade in school, but it could also be pay grade. The answer is "definitely NO" because drug that is good for me good for women and bad for person does not exist. Sure Thing Principle . A theorem in do-calculus.

4.3.2023 11:52am - (Replying to @dimgrr) Nice, except I've never heard of the "Principle of consistency", especially when in stochastic settings, where by "X causes Y" we mean "X raises the probability of Y". Worth asking for reference to "consistency".

4.3.2023 4:49am - Why the weighted average argument works in causal setting and not in Simpson's Paradox: See The Sure-Thing Principle: (reposted:

4.3.2023 4:42am - (Replying to @mcdonalds_tim) The answer is correct, but the reasoning is not, because the rule of total probability applies as well to Simpson's paradox, where the issue is association, not causation and, Lo and Behold, there we do have sign reversal. The impossibility is a theorem in do-calculus.

4.3.2023 2:39am - I am curious about your conclusion: "Adding more controls need not bring you closer to identification! " So, how does one know if the controls available are sufficient, redundant, or bad? (reposted:

4.3.2023 1:59am - (Replying to @KarlPettersso10) I am not sure about the meaning of "over a life time". Can you prove it from do-calculus?

4.3.2023 12:48am - (Replying to @davidboxenhorn) We should not conflate AI with GPT-4. The latter is indeed very useful for reflecting the current wisdom of the crowd which is often not reflected in the literature of in what your politically correct colleagues tell you.

4.3.2023 12:43am - (Replying to @bytebiscuit) Every stat instructor can parrot this slogan. I'm trying to test causal understanding.

4.3.2023 12:35am - Unfortunately, it didn't get it. Me: Shouldn't the same argument works in Simpson's Paradox? GPT-4: If smoking is consistently associated with a grade decrease within every age group, then it should be associated with a grade decrease overall, as you correctly pointed out. Wrong (reposted:

4.3.2023 12:26am - (1/2) Luckily, it's teachable: Me: But if smoking causes a grade decrease within every age group, shouldn't it also cause an average grade decrease overall? Isn't the latter just a weighted average of the former, which makes it impossible to reverse sign? GPT-4: "You are correct. (reposted:
4.3.2023 12:26am - (2/2) If smoking consistently causes a grade decrease within every age group, it should also lead to an average grade decrease overall. I apologize for the confusion in my previous response." Correct! But I'm still not sure it got it; same heuristics straps people in Simpson Par.

4.3.2023 12:14am - (Replying to @rootsaandwings) No. By "causal" we mean the Probability of grade increase if you decide to smoke is higher than if you decide not to smoke.

4.2.2023 11:46pm - (1/4) Got my first session with GPT-4, amazing! Though it failed its first causal understanding test. Me: "Is it possible that smoking causes grade increase on the average and, simultaneously, smoking causes grade decrease in every age group?" GPT-4: It is theoretically possible
4.2.2023 11:46pm - (2/4) for smoking to cause an average grade increase overall while causing a grade decrease within every age group. However, this scenario would likely involve some form of Simpson's paradox, where the overall relationship between two variables reverses when accounting for
4.2.2023 11:46pm - (3/4) a third variable (in this case, age). ....etc. It failed because (1) I asked a question the answer to which is "definitely NO, no exceptions", and (2) It read somewhere about Simpson's Paradox and a "third variable," overlooking that the question was explicitly causal,
4.2.2023 11:46pm - (4/4) not associational (" Is it possible that smoking CAUSES ..."); statistics texts do not normally ask causal questions. In summary, GPT-4 understanding is at the level of a well-read statistics instructor -- Amazing in its breadth! But still Rung-1.

4.2.2023 9:04pm - (Replying to @noatishby) Sad to see you go, Noa, but the war is getting nastier, and your comrades in the trenches need you. You do not need a title to be part of the War Cabinet -- your seat will be waiting for you.

4.2.2023 8:47pm - The ocean of Israeli flags that have dotted the landscape in 2023 speaks to a deep, visceral patriotism that cuts across much of Israeli society. (posted:
4.2.2023 8:47pm - A re-discovered hard truth: Israelis love their country. "Now that they’ve seen who they voted for, many of the coalition’s voters have buyer’s remorse". @DavidSuissaJJ

4.2.2023 5:29pm - (Replying to @forde_nathaniel @stephensenn and 3 others) Enders's analysis is still stuck in Rubin's marshland. It warns readers: "These assumptions ... are inherently unverifiable. Your analysis will stand or fall depending on how plausible each assumption is..", yet makes them as incomprehensible as possible using stat notation.

4.1.2023 4:06am - (Replying to @ralfduschef) I can't buy this dictum: "The social building of trust is beyond engineering," perhaps because I remember how people argued (before 1947) that "playing chess is beyond engineering".

4.2.2023 5:16pm - (Replying to @setspace and @ralfduschef) This is precisely why we need to understand humans using new tools, new languages and new logic.

4.2.2023 5:13pm - Refreshing to see missing data analysis elevated from the statistical marshland in which it was initially hatched, towards causal modeling, where it obtains transparency, identification and testability. (reposted:

4.2.2023 4:57pm - (Replying to @f2harrell @elmir1omerovic and 3 others) It's not so much the difference between Bayesian and frequentist approaches, as it is between statistical and causal thinking. The latter identifies when the former makes sense, and adds to it transparency plus testability.

4.2.2023 2:10pm - (Replying to @f2harrell @matloff and 4 others) Beg to doubly differ. Missing Data is a causal, not statistical problem. Causal analysis is indispensable for showing when statistical heuristics (e.g., multiple imputations) work and when they don't. See examples:

4.2.2023 11:45am - Statisticians' wars over p-values and Bayes' subjectivity are merely a prelude to the gigantic wars still awaiting them when the number of samples increases and at issue is whether statistics is the language of science. See "Why I am only a half-Bayesian" (reposted:

4.2.2023 12:55am - (Replying to @3scorciav @AlexTensor and @berkeley_ai) NSF and top-funding agencies do fund some small-scale research in Rungs-2&3. However, large-scale multi-universities proposals are reviewed by Rung-1 experts and, naturally, are funded to Rung-1 investigators. Exceptions may exist, but not to my knowledge.

4.1.2023 3:29am - (Replying to @Alberto_Begue_ and @HannesMalmberg1) I should get it. Thanks.

4.1.2023 2:40am - (Replying to @HannesMalmberg1) Beautiful! What Chatbot is it?

4.1.2023 1:23am - Call me sentimental, @jazchaz , but don't ask me why tears start sliding down my cheeks when I hear this song, and why it makes me so absolutely sure Israel will survive this crisis. Here is the original, sung by Gali Atari in 1986:,vid:uyv5WCzR7z0 (reposted:

3.31.2023 6:50pm - That's another reason why I like Twitter. (reposted:

3.31.2023 6:39pm - I'd strongly support the idea of a Manhattan Project of intense research to make machines more trustworthy and interpretable (regardless of, or in parallel with a moratorium.) The premature super-investment in non-interpretable technologies is the core of our problems. (reposted:

3.31.2023 11:31am - (Replying to @fhuszar) Counterfactuals help us reason the world in which we live. I find it hard to imagine anyone interested in the process by which GPT-4 chooses its words; it ain't our world.

3.31.2023 2:09am - (1/2) I am on the waiting-list for GPT-4 access. Once I get it, my first task would be to teach it counterfactual reasoning on the firing-squad story (GPT-3 had hard time grasping that I have zero interest in whether shooting rifles is legal in California), and then move to an (reposted:
3.31.2023 2:09am - (2/2) isomorphic story (e.g., sprinkler and rain) to see if GPT-4 got the logic, namely, if it forms a causal map of the story and utilizes the wisdom of the map, without being bugged down with the idiosyncratic details of sprinklers and rains.

3.31.2023 1:35am - (Replying to @CotlerWunsh) Hear, Hear, in its entirety, and add to it the Prayer: "Barech et Medinat Israel, Resheet Tsimat Geulatenu... and send your light and truth to her leaders and experts and bless them with your good advice." Amen!

3.31.2023 1:20am - (Replying to @fhuszar) The theorem says nothing about the causal graph; it simply states what cannot be done without some certain assumptions. Are you saying we have zero interest in the effects of interventions?

3.31.2023 12:38am - (Replying to @CotlerWunsh) Agree on the potentials that this crisis holds for Israel: The backbone of Israel's society has discovered its true identity; secular, Zionist, principled and, most importantly, respectful of broad consensus. I join Harari's call to add Megilat HaAtsmaut to the Pessach Haggadah.

3.31.2023 12:01am - I strongly condemn this vicious, defamatory and baseless post by @WounterDeVriendt against @GeraldNGOM . I hope decent members of the Belgian Parliament will examine the facts and join me in this condemnation. (reposted:

3.30.2023 11:35pm - Prof. Yuval Noah Harari: Judicial coup in Israel ‘will destroy academic freedom’ (posted:
3.30.2023 11:35pm - This is an English summary of Harari's lecture (in Hebrew) which I've posted here a few days ago.

3.30.2023 9:57pm - First good news from CUNY (The City University of New York). Congratulations @SAFECUNY ! (reposted:

3.30.2023 9:45pm - (Replying to @agarret7) Apologizing - My acronym-knowledge is limply.

3.30.2023 2:49pm - (Replying to @ATaher_ @julioalexo and @_ZachFoster) The Mufti had a stronger grip on the Palestinian people in 1930's than any "president" I know. Are you disavowing his teachings vis-a-vis the need to crush Zionism, and to prevent European Jews from escaping Europe?

3.30.2023 2:43pm - (Replying to @skassabr) No. I do not know of any Palestinian capable of uttering the words: "Equally legitimate, and equally indigenous" when the objects of discourse are the Jewish and Palestinian PEOPLES (forget "religions"). And I have challenged them on at least 100 occasions; they CAN'T !!!

3.30.2023 12:53pm - (Replying to @KillianForde) When was the last time you told your son that his real home is Bergen, from which you were expelled by force, that you never signed off your property rights to anyone, and that he should expect to return there soon, to sing again the poems written there, which he knows by heart.

3.30.2023 11:42am - (Replying to @julioalexo @ATaher_ and @_ZachFoster) I can understand why Hajj Amin al-Hasayni befriended Hitler and mistook Zionists to be crusaders. What I cannot understand is why Palestinians today, especially their intellectuals would perpetuate the Mufti's mistakes to prolong and amplify their people sufferings.

3.30.2023 11:25am - Thank for posting. I wasn't aware of the fact that the Peel's Commission (1937) stated explicitly that the Arabs charges had no basis. The Arabs refused, at that critical moment of history, to consider European Jews as part of the deal, thus sealing their fate in the ovens. (reposted:

3.30.2023 5:04am - (Replying to @FJnyc) According to Rashid Khalidi, Palestinian national identity was forged in the 1920s, but it was too late "to crush the Zionist movement". It's the only national identity in the world forged on "crushing" rather than collective memories.

3.30.2023 3:40am - (1/2) (Replying to @MSkywalker @NewwAmericanna and @_ZachFoster) The idea that "We never really left that land" is not a Zionist invention, but a thread governing governs 80 generations of Jewish thoughts. My grandfather said 3 times a day "The merciful one will soon walk us, in sovereignty, back to our country." So, please, do not treat
3.30.2023 3:49am - (2/2) (Replying to @yudapearl @MSkywalker and 2 others) this yearning as legends of the past; treat them as forces operating in the present. Once you do, we can begin to define "indigeneity" with the Palestinians and, hopefully, converge on "equally indigenous".

3.30.2023 3:25am - (1/3) GPT-4 should know that my statements about DL never rising above "curve fitting" is not an opinion, nor a "common concern" but a mathematical theorem. You cannot answer causal questions from passive data alone, unless you enrich the data with: (1) causal assumptions or (reposted:
3.30.2023 3:25am - (2/3) (2) interventional experiments. GPT-4 essentially agrees with this statement by admitting that its conclusions rest on causal assumptions extracted from the training text. It is still quite impressive for GPT-4 to be able to treat stories about experiments as experiments,
3.30.2023 3:25am - (3/3) and stories about causal assumptions as assumptions. The conversion, from stories to substance, would be very useful as a user- interface in many causal inference tasks. We need to make sure, though, that it can handle the substance properly, assuming a proper conversion.

3.30.2023 2:41am - (Replying to @MSkywalker @NewwAmericanna and @_ZachFoster) You make it sounds like an afterthought; "Oh, legends say we had something then, so let's claim something now." Wrong! We never really left. Certainly not in spirit. And it's the spirit that defines indigenous people, not a place of residence. Equally indigenous, or course.

3.29.2023 10:37pm - (1/3) Sounds like our "historian" has no history. * Show me one case of colonialists moving into a country they perceived to be the birthplace of their history • One case of settlers speaking a language spoken in the old land before the one spoken by its contemporary residents (reposted:
3.29.2023 10:37pm - (2/3) *One case of settlers whose holidays commemorated historical events in the land to which they moved, not in the lands from which they came. • One case of settlers who did not name towns "New XXX", like New York or New Wales, but by names those towns enjoyed in ancient times
3.29.2023 10:37pm - (3/3) (Israeli town are not named "New Warsaw", "New Berlin," or "New Baghdad") • One case of settlers who narrated their homecoming journey for eighty generations in poetry, prose, lore, and daily prayers. More on this topic, here:

3.29.2023 2:51pm - When a Palestinian intellectual says "It is amazing ..." you can expect a shallow regurgitation of some populist slogan, combined with a profound emasculation of Palestinian agency. @dianabuttu @Ostrov_A @EinatWilf (reposted:

3.29.2023 2:30pm - Note also: The name was "Palestine Liberation Organization", not "Palestinian" -- an essential distinction, when it comes to aims and agenda. (reposted:

3.29.2023 2:13pm - Can some pro-Palestinian activist explain to their "historians", intellectuals and spokespersons that denying Jewish roots in the holy land does more to prolong the occupation than all terror attacks put together? They don't seem to get it! (reposted:

3.29.2023 1:49pm - (Replying to @desai_pratik) I have the same feeling but I can't put my finger on any scholarly work on this theory; my knowledge on language acquisition is shallow.

3.29.2023 12:39pm - (Replying to @Anatomi54244944 and @PrzemekPWr) Zionists covenant with G-d was: We will support your plan about the end of days, if you support ours about tomorrow -- he signed it.

3.29.2023 12:28pm - (Replying to @hearnimator) And there should always be a discussion of Zionophobia in the context of Islamophobia; religion has not monopoly on people's identity.

3.29.2023 4:24am - (Replying to @blakeflayton) I witnessed an earlier peace rally -- the "Land for Peace" euphoria of August 1967, before any settlement was in sight. It was rebuffed by Khartoum's "Three No's" which proved the occupation inevitable. The emasculation of Palestinians agency is fundamental to their history.

3.28.2023 10:58pm - (Replying to @chaimlevinson and @bezalelsm) Amnon Shapira was my High School mate at Ironi-Datti (Zeitlin). We were both in Bnei-Akiva when it was a Torah V'Avoda movement, before it was highjacked by messianic ghosts. Shalom Amnon, Shalom Chaver, Yehi Zichro Baruch.

3.28.2023 7:43pm - (Replying to @AleksanderMolak and @yunliwoo) Unfortunately many stat-ML authors are now adding the key word "causal inference" to their papers, though they have nothing to do with causal inference. I am using my own formula, it's not perfect, but captures the low lying fruits.

3.28.2023 4:42am - Like other Mondays before this, we are welcoming 5 pages of new publications in the many aspects of causal inference: Please search for and share breakthroughs.

3.28.2023 3:06am - If you are stuck on RCT and looking to action something useful from #Bookofwhy, consider generalizing RCT findings. Can you do it without do-calculus? I doubt it. Now watch this:, or this:, or this:

3.27.2023 11:34am - If I were not a disbeliever, I would swear that it was God who listened to my prayer and implanted a seed of responsibility into Bibi's head, to halt the Judiciary overhaul and allow time for reaching wider consensus. But given who I am, I'm waiting for him to reinstate Galant. (reposted:

3.27.2023 6:03am - McGill Tribune Rejects Pro-Israel Column Submitted By Jewish Student Saying Zionism Doesn’t “Align With Our Values” via @honestrepcanada (posted:

3.27.2023 4:01am - My anxiety over the events in Israel has risen to the point where all my secular convictions have given way to a prayer: Oh Father in Heavens, Bless the State of Israel, Shield it with your love, and spread over it the canopy of your peace. Please join me: (posted:

3.27.2023 3:39am - Ben Gurion departures halted as Histadrut declares ‘historic’ strike over overhaul via @timesofisrael (posted:
3.27.2023 3:39am - It's a major development: Israel's largest Union (Histadrut) joins the "stop the madness" strike to halt Government plans.

3.26.2023 6:23pm - Nice test. I was suggesting generalizing from story to story, but time is a much better filter against bluffing. (reposted:

3.26.2023 6:17pm - The photos are distasteful, but ignoring half of the country is a disaster.

3.26.2023 4:43pm - I can't brush it off so easy. I've defined "understanding" as the ability to answer questions on all three levels of the Ladder. Sticking with this definition, GPT-4 has not failed yet, especially if it can generalize from story to story. I haven't tested it on this dimension.

3.26.2023 4:37pm - Very nice! From now on, if anyone has a question about #Bookofwhy, don't ask me, ask GPT-4.

3.26.2023 3:07pm - Will Bibi lead Israel to suicide? Tamut Nafshi? Cry, the beloved country! Za'aki Erets Ahuva! Livshi Bigdei Tifartech, Ami! (reposted:

3.26.2023 12:54pm - For this reason, GPT's understanding should be tested on questions whose answer is "NO, it is not possible under any circumstances". Such questions are the core of causal logic. e.g., can a drug be good for men, good for women and bad for people? (reposted:

3.26.2023 12:46pm - Netanyahu fires defense minister Gallant for calling to pause judicial overhaul via @timesofisrael (posted:
3.26.2023 12:46pm - It's now Avi Dichter that can save Israel, by refusing. Israel's saviors change by the day, the country is waiting for one with clout and courage. It's time for Kohelet's staff to come forth and say: We Erred!

3.26.2023 12:36pm - Fairly amazing. GPT seems to have read #Bookofwhy. The answer is wrong though. The question did not state that K and U are initially uncorrelated. If they are correlated then it is feasible that they would turn uncorrelated after conditioning on X.

3.26.2023 12:27pm - Good suggestion, but #Bookofwhy was written for people, i.e., a species with inborn crave to represent its understanding of the world in a road-map called "causal model". Unless CHAT is endowed with such craving, I doubt he could understand #Bookofwhy.

3.26.2023 6:24am - Asking a "Palestine Historian": "Why do you lie?" is like asking a Holywood producer: "When did you shoot your last movie?"

3.26.2023 6:16am - It sure is an expression of hate, even racism, but why give the racist an instrument to deflect his racism, and to cover it with populist slogans that exonerate him from antisemitism. Why not pinpoint precisely, and undeniably, the kind of racist he is -- an ugly Zionophobe.

3.26.2023 5:50am - "Jewish study major" - Another glimpse at the sophisticated educational level reached by the "Jewish Studies Association". @StudiesZionist (reposted:

3.26.2023 5:28am - I knew he would take that route, and that is why I was careful NEVER to call @KenRoth an antisemite, always a Zionophobe -- a stronger indictment which he cannot deflect nor deny. (reposted:

3.26.2023 5:18am - That's the time to say: We erred!

3.26.2023 5:16am - Could not get the question.

3.26.2023 4:43am - Agree. All eyes are now looking at Likkud ministers to take a stand against the reform, but perhaps more effective would be for the staff of Kohelet to come up and say: "We erred! We underestimated this or that.. We erred." They will be forgiven, and commended for their courage.

3.26.2023 1:03am - Stupid and extremist calls come from both sides, what does not come is a clear analysis like this one: (posted:

3.25.2023 11:51pm - I'm stuck on the questions and the input information.

3.25.2023 11:49pm - I heard them too. But as King Solomon the wise said: There is time for everything under the heavens. A time to give birth, and a time to die; a time to plant, and a time to uproot the plant. And humble me says: There is also a "manner" for everything under the heavens. Manner!

3.25.2023 11:43pm - Palestinians have created a new logic for cause-and-effect, called "triggered by Zionists". It has already given rise to a great number of scientific innovations, startups, and research universities. Watch @AhmikAser for additional progress. (reposted:

3.25.2023 11:16pm - Israel refuses to surrender, 300,000 protesters in Tel Aviv alone. Cry, the beloved country! Za'aki Erets Ahuva! Livshi Bigdei Tifartech, Ami! (reposted:

3.25.2023 6:26pm - Israeli defense minister, Yoav Gallant, calls for halt to judicial overhaul. The ball is now in the hands of other sane ministers in the government; will they rise to the vital call of sanity? Will they listen to the desperate call of their country? Next week will tell. (reposted:

3.25.2023 8:54am - Same goes for 'Jewish Voice for "peace"', 'New Israel Fund', perhaps also schizophrenic J-Street, regardless of whether they're overtly or covertly pro-BDS.

3.25.2023 1:59am - For our Hebrew speaking readers, Prof. Yuval N. Harari (author of Sapiens) is here giving the best analysis I've found of the Judicial crisis in Israel: He mentions a pending decision by the academic senate to close the Hebrew University in Jerusalem. (posted:

3.24.2023 5:41pm - Jerusalem security forces on alert as tens of thousands at 1st Friday Ramadan prayers via @timesofisrael (posted:
3.24.2023 5:41pm - If you are praying (as I do) for the Ramadan Holiday to be celebrated in Jerusalem without violence, the following is an encouraging report.

3.24.2023 4:37pm - The lullaby "Palestinians want peace" has long been replaced by "From the river...". They now get offended by the former. "Peace with whom?" they ask, "with an entity that does not exist?" Don't offend them, please.

3.24.2023 3:55pm - When @jeremycorbyn calls for justice, equality and peace we should all freeze in attention; has he been notified about "truth"? (reposted:

3.24.2023 3:49pm - 40 years of Zionism brings us to 1936, when I was born, and we were 400,000 Jews in Eretz Israel. (By records of the Mandate Authorities) (reposted:

3.24.2023 3:37pm - When a reader asked: "Do Palestinians love Allah more than they hate Israel?" other jumped: "How dare you offend a whole culture?". I think a culture born on negation needs to be understood that way. (reposted:

3.24.2023 3:25pm - (Replying to @DavidHirsh) I don't think Zionophobes care about France; they'd only dig it if asked: Does Islam have some God-given right to exist? Wow!! They'd jump, "How dare you compare religion to a country?" The next thing to ask is "Who said religions have monopoly over existence?". Now they get it.

3.24.2023 3:08pm - (Replying to @jiaxx213umn) I cannot tell from your paper what questions can be answered by your do-DAGs that cannot be answered by ordinary DAGs. Also, if there are such questions, does a do-DAG convey more information from users?; if so, what is this extra information?

3.23.2023 11:12pm - (1/2) Do you have a GPS in car? Note that, in addition to those sweet "turn right, turn left" instructions, you also get a road map. Why? To enable you to pick a detour, in case you hit a road block or a construction site. So, next time you hit a perturbation over what your LLM (reposted:
3.23.2023 11:12pm - (2/2) has been trained on, ask her if she can give you a detour. If she has a world model, she would have no problem doing so; otherwise, she will try to fake one, and it will show immediately. @DavidDeutschOxf @GaryMarcus @ylecun

3.23.2023 10:06pm - (Replying to @GeraldNGOM @yoavgallant and @netanyahu) I was hoping; we were all hoping. The rest is history (or, rebuke of history).

3.23.2023 10:03pm - (Replying to @DavidDeutschOxf and @michael_nielsen) Is there a test to distinguish between the two?

3.23.2023 1:14pm - I havn't read this essay before - thanks. It nicely compiles the arguments why AI needs causation, but I do not find its arguments against AGI to be compelling; they actually border on the metaphysical -- beyond me. I aim to build intelligent machines, not to speculate. (reposted:

3.23.2023 9:53pm - (Replying to @AlexanderNaume2) Fine, but this does not imply sufficiency. We still need to differential results of actions from results of observations, and more.

3.23.2023 9:19pm - This demonstration takes place a few feet from the house where I was born. But in my days we didn't have such demonstrations, we played together Haredim, Chofshiim and Masortiim. It was before the Dark Ages, then the city renamed Herzl Steet Rabbi Schach St. --dark onto darkness. (reposted:

3.23.2023 8:43pm - Peoples with no historical records are yearning to go back to the Ottoman Empire, arguably the most repressive, corrupt and dis-functioning regime in human history, to comfort their need for victimhood and to negate their neighbor's history. (reposted:

3.23.2023 8:22pm - The opening plenary set the tone by urging attendees to think big, to dream big, to see far down the road. (posted:
3.23.2023 8:22pm - it's been a gloomy day in Israel. Netanyahu missed an opportunity to heal the nation. What elevates me to hopes and action are the words of @DavidSuissaJJ (rephrased): "Think long-term, create future memories for generations to come." Amen!

3.23.2023 8:05pm - (Replying to @AngeloDalli @soboleffspaces and 3 others) I can resonate to (1), (4) and possibly (5), because I've tried them on the logic of causation and proven them valid. I'm hesitating about 2 & 3, because I haven't seen them formalized yet, awaiting a toy example with 4-5 variables. Note: 1 requires modeling assumptions, not 4.

3.23.2023 7:16pm - (1/2) Commenting on: "DeepMind successfully demonstrated “a recurrent network with model-free reinforcement learning to solve ..[causal problems] (Dasgupta et al., 2019). RL uses actions, to (reposted:
3.23.2023 7:16pm - (1/2) make up for missing models. Still, model-free RL cannot fully climb to Rung-2 of the ladder, as shown by @eliasbareinboim, not to mention climbing to Rung-3.

3.23.2023 6:58pm - (Replying to @Dude62261752) I don't know any of my students, past and present, who believes in either: (1)Model-free deep learning, or (2) "AGI is impossible".

3.23.2023 10:54am - What happened to CUNY? - A fairly incriminating report!!! (reposted:

3.22.2023 10:02pm - (Replying to @prem_k @soboleffspaces and @GaryMarcus) Thanks for posting this article. It mentions causation 5 times, but I am not sure I would agree with the authors on how they phrase it.

3.22.2023 8:54pm - Refreshing to see Students Supporting Israel carry the torch of sanity to the heart of the "2nd Annual Palestinian Mentality Week", which started today at UC Berkeley! (They call it "apartheid week", but the idea is the same: From the River to the Sea, no one else but We & We) (reposted:

3.22.2023 6:45pm - (Replying to @soboleffspaces) I wouldn't call it a 'framework' until I see at least: (1) A new question answered by the integration. (2) A recognized CI researcher in the References, (3) An equation describing what one hopes to get out of the integration.

3.22.2023 6:25pm - (Replying to @soboleffspaces @stephensenn and 2 others) I agree that the integration problem should better be phrased as "what questions can be answered by {RCT+RWD} combinations, that cannot be answered by each component alone?" The ability to answer questions entails logical deduction, and rules out hand-waving.

3.22.2023 6:16pm - (Replying to @soboleffspaces @rtaira18 and @stephensenn) The authors are not known as CI contributors, so one would expect them to be unaware of the Ladder of Causation and of what quest

3.22.2023 5:13pm - (1/2) The reason I asked "Can they do it?" is that I do not believe it can be done without understanding the principles of integrating RCT and RWD data, and those principles will get buried in noise unless we can demonstrate them on toy examples. Moreover, though everyone agrees (reposted:
3.22.2023 5:13pm - (2/2) that it's silly to ignore RWD about the "outcomes of medical products in the population," it's hard to point to a simple example where such integration is guaranteed to yield something that each study alone cannot yield in isolation. Perhaps readers can help in the search?

3.22.2023 1:48pm - (Replying to @soboleffspaces @dylanarmbruste3 and 2 others) Some philosophers (I among them) would take it as evidence that humans carry DAGs in their heads. But I am sure others would object.

3.22.2023 4:52am - (Replying to @_J_Pe) Would you still resist the temptation if Putin's mentality was driving force behind Russian national identity?

3.22.2023 4:21am - As readers know, "attribution science" deals with the "probability of necessity" (PN), a counterfactual term that is analyzed here and Note that, in climate context, we do not have experimental data, but monotonicity holds. (reposted:

3.22.2023 3:33am - (Replying to @_J_Pe) The wisdom you wish to be distilled from history teaches us that not all "national identities" are equally benign, some can even be dangerous, especially those who are bent on denying national identity to their neighbors.

3.22.2023 2:18am - (Replying to @Dude62261752) As I read this paper, it assumes a (triangular) structure of IV type, and further imposes shape restrictions (eg monotonicity) on some of the functions. In our "Personalized Decision" we make no structural or functional assumptions, hence the bounds.

3.21.2023 6:31pm - (1/4) "Proven by practice" may be a good guideline for repetitive routines for which we have a valid criterion of success (e.g., using seatbelts). But when we seek justification for a totally new practice (Fisher's RCT), we can't rely on practice alone. Also, we may wish to ask
3.21.2023 6:31pm - (2/4) whether the success observed can be improved, or whether it is not due to some temporary condition, etc. We then seek conditions under which success is guaranteed. "Guarantee" is a logical notion, demanding deduction. A good example is the practice of "taking the average" of
3.21.2023 6:31pm - (3/4) several observations to improve precision of measurements. It was "proven by practice" by many astronomers, until Thomas Simpson (1747) asked: Can we prove it? He did! And by so doing he articulated the conditions under which improvement is guaranteed. Now let's go back to
3.21.2023 6:31pm - (5/5) RCT and ask "what guarantees us that going through the procedure specified we would get a better answer to our question: "causal effect =?" Causal Inference answers this question, all others fumble around it, some do not even agree on what our question is. The proof is easy!

3.21.2023 2:44pm - Forgot to tag @stephensenn @soboleffspaces ; an extremely important question touching on the entire enterprise called "causal inference" = reduce research questions to a deductive exercises, showing that the desired conclusion logically follows from the available premises. (reposted:

3.21.2023 2:32pm - (Replying to @_StephenOlivier) Great question, which has been floating in the air throughout these discussions. "justify" means proving deductively that the result of a given exercise (OS or RCT) gives the correct answer to our research question: "What's the Causal Effect of treatment T on outcome Y."

3.21.2023 2:11pm - Which confirms my theory: "Only he who can justify OS can also justify RCTs." (reposted:

3.21.2023 2:04pm - The 5th question for the Passover Haggadah of Ex-Jews: "Why the hell were we ever Jewish? We don't even live "there" (the holy land); Why are we saying 'next year in Jerusalem'? Why do we care what's happening to "them" (in Israel)? Ex-Jews wish to speak for millennial Jews.

3.21.2023 1:48pm - (Replying to @soboleffspaces @dylanarmbruste3 and 2 others) Some philosophers (I among them) would take it as evidence that humans carry DAGs in their heads. But I am sure others would object.

3.21.2023 1:22pm - Which proves that causal diagrams can turn even physics into a colonial enterprise. What a powerful tool! (reposted:

3.21.2023 1:04pm - (Replying to @AviMayer and @Jerusalem_Post) Congratulations, Avi. B'Hatslacha

3.21.2023 12:54pm - (Replying to @soboleffspaces and @stephensenn) In other words, only he who can justify OS can justify RCTs.

3.21.2023 12:10pm - (Replying to @stephensenn and @soboleffspaces) Your point is well taken. My point: People forget that "randomization" as used in RCTs is not random sampling, but "randomized intervention". In other words, we need to mention "intervention" explicitly, because it is not automatically understood from the word "randomization".

3.21.2023 11:57am - (1/2) (Replying to @dylanarmbruste3 @soboleffspaces and 2 others) I didn't find this part illuminating, perhaps because today we know so much more (about OS justification) than anything that could be articulated in 1990. Why? Because in 1990 we could not articulate the assumptions needed for valid causal inference, and we know that all
3.21.2023 12:03pm - (2/2) (Replying to @yudapearl @dylanarmbruste3 and 3 others) inferences are only as credible as the causal (judgmental) assumptions that enters into our model. The latter became articulable only with the advent of DAGs.

3.21.2023 10:51am - (Replying to @soboleffspaces @dylanarmbruste3 and 2 others) It's truly interesting to read Greenland's article of 1990. DAGs and d-separation were already in existence, but not causal DAGs (1991), for this was before I met Sander and Jamie(1993), a meeting that led to the causal revolution in Epidemiology, 1999:

3.21.2023 10:44am - (Replying to @soboleffspaces @dylanarmbruste3 and 2 others) Where is the manifesto taken from?

3.21.2023 2:48am - Palestinians national identity, because it is only 102 years old, has shallow historical roots. It's sad, because roots make nations secure and pragmatic. Instead, it revolves around negating their neighbor - a futile battle, given the neighbor's stubborn insistence on life. (reposted:

3.21.2023 12:52am - Palestinians national identity, because it is only 102 years old, has no historical roots - an ingredient that makes other nations secure and pragmatic. Instead, it revolves around negating their neighbor - a futile battle, given the neighbor's stubborn insistence on life. (reposted:

3.21.2023 1:10am - I love it: "Ignorance is a choice". (reposted:

3.21.2023 12:33am - Westerners will never dig Palestinian mentality. (reposted:

3.20.2023 11:32pm - (1/2) It's Monday again, bringing us another stream of papers in causal inference: As I read through the abstracts, I can spot some that are wrongly headed, and some that I put on my "next to read" list. The one that caught my attention is: "Bounding the..
3.20.2023 11:32pm - (2/2) "probabilities of benefit and harm through sensitivity parameters and proxies" by Jose M. Pena Perhaps because of its relation to our "Personalized Decision Making" paper, which was discussed at great length on this channel.

3.20.2023 10:27pm - (Replying to @stephensenn and @soboleffspaces) I beg to differ. The fact that they are taken under RCT regimes DOES make them causal; society would not go through the expense and hardship of RCT if what we get out of it was NOT causal. I'm sure readers would be curious to see your (mathematical) definition of "causal."

3.20.2023 10:13pm - (Replying to @AndrewPGrieve @soboleffspaces and @stephensenn) The fact that the nurse does not know if the tablet in her hand is sugar or treatment does not absolves her from the responsibility to give patient i the tablet prescribed by the randomized assignment, not the one preferred by the patient. This amounts to "intervention".

3.20.2023 10:02pm - (Replying to @OstlundOllie @AndrewDGarrett and 2 others) I assume ICH E10 is basing its restrictions on solid evidence and in-depth understanding of the issues involved. If, however, they are open to hear an opposite view, I'll be happy to share my understanding of the tradeoffs involved.

3.20.2023 4:44pm - (Replying to @stephensenn and @soboleffspaces) Of course you need to consent, and of course this makes you different (this is called "selection bias") But, once you consent and your parents consent etc. etc, if you are assigned to treatment a nurse asks you to swallow the drug even if it tastes bitter -- "intervention".

3.20.2023 2:08pm - (Replying to @stephensenn and @soboleffspaces) In the non-parametric analysis with which I am familiar, no assumption is made about relative magnitude of effects. In particular, using OS DOES NOT assume that "interactive effects (effect modification) are more important than main effects (confounding)."

3.20.2023 11:04am - (Replying to @soboleffspaces and @stephensenn) Why not call in "intervention"? I have the hunch that people do not fully appreciate the fact that RCT involves two components: (1) Randomization and (2) Intervention. E.g., once you are assigned to treatment arm, you are given the treatment whether you like it or not.

3.20.2023 6:48am - (Replying to @revprez and @fallingmrf) Well put. @fallingmrf would like us to be generous to these ex-Jews, carrying "Humans Against Breathing" posters on the chests, as if they merely wish to alert us to the dangers of breathing foul air. I label them ex-Jews because breathing is an essential tenets in Judaism.

3.20.2023 12:56am - (Replying to @fallingmrf) Posing in that picture, with those posters on chest is more than just "opinion". It's a virtue-signaling statement saying: There is something morally wrong with Zionism and we, the morally endowed below, have found its cure. All they've found is temporary relief from bullies.

3.20.2023 12:38am - (Replying to @mapto) History teaches me that endowing someone with the title "identity" while deprecating your identity to the level of "political agenda" is a careless, if not dangerous act of inviting more bullying from those who are just waiting to find Jews lacking coherent identity.

3.20.2023 12:23am - My inborn modesty would not allow me to say that you made the right choice, that your students are fortunate to start their critical thinking voyage on the right boat, and that my biased opinion is normally biased in the right direction. (reposted:

3.20.2023 12:14am - And I, personally, will not forgive these ex-Jews in the picture for giving our bullies the impression that we, Zionist Jews, lack spine, historical perspective and intellectual resources to tell them: "Go find your weakling prey elsewhere - we can fight your racism hands down. (reposted:

3.19.2023 7:59pm - (Replying to @kareem_carr) In statistics, a "mental model of the world" is a tacit background assumption. In AI, an explicit mathematical specification of a "mental model" is the starting point of the analysis, which allows us to decide what assumptions are needed for any given task, eg, generalization.

3.19.2023 2:03pm - A constitutional opportunity for the bedrock of Israeli society to reclaim the rock. (reposted:

3.19.2023 4:55am - Painful words that must be said. (reposted:

3.19.2023 4:44am - Not so harsh, @MichalSabra ; @sjabulhawa is doing great service to humanity by unveiling to would be Palestinian supporters what mentality they risk acquiring once they enter this culture of deceit. @Ostrov_A @Claire_V0ltaire (reposted:

3.19.2023 4:31am - Strongly disagree. Jews will never be forgiven for allowing their sociopaths to pose as "Jews", deceiving the whole world about who they are, and handing anti-Semites an excuse to say: "We thought you are all as degenerate as those in the picture!" (

3.19.2023 4:10am - Not to be forgotten. (reposted:

3.19.2023 4:01am - (Replying to @kareem_carr) What's the other sense? Is there anything worth learning from AI? Personally, I've learned a lot, and I'm so grateful for what I learned, that I'm anxious to share it with others.

3.19.2023 3:50am - Important to note: The "concurrent control" restriction applies to RCTs studies in healthcare applications, not to Observational Studies, where we can estimate both P(y_c) and P(y_t) from the data + DAG combination, as in @soboleffspaces work on hip-fractures. (reposted:

3.19.2023 3:40am - Where did you find such wondering souls? Last I checked, they all disappeared. (reposted:

3.19.2023 1:51am - Congratulations to @ThePearlPost staff for another mark of excellence. (reposted:

3.18.2023 12:05pm - (Replying to @DavidDeutschOxf) Important to note that the Book of Esther does not take a religious stance against "The dark logic of moral nihilism" but a rather pragmatic one: "And you and your father's house will perish". Forget "good and evil", we shall all perish if we follow Chomsky and Mearsheimer.

3.18.2023 11:35am - (1/2) (Replying to @NowakPosadzy and @BrianCAlbrecht) Hard for me to accept the idea that economists "do equations," hence are mentally incapable of lifting themselves from the equations to see and work with certain qualitative relations (among the equations), other than the equations themselves. Hard, but real; Ed Leamer asked
3.18.2023 11:45am - (2/2) (Replying to @yudapearl @NowakPosadzy and @BrianCAlbrecht) me a similar question: "If I feel more comfortable with equations, any reason to do graphs?" Heckman and Pinto tried, see how far they got without "backdoor" of d-separation: It's like describing how computers work by appealing to electrons and holes Twin networks are good for verifying independencies among counterfactuals (i.e., ignorability conditions). But the tree-step method is better for actually estimating counterfactual expressions. See

3.18.2023 2:21am - (Replying to @gottfriedmath @artistexyz and @soboleffspaces) Twin networks are good for verifying independencies among counterfactuals (i.e., ignorability conditions). But the tree-step method is better for actually estimating counterfactual expressions. See

3.18.2023 2:09am - For the many readers who were waiting for a video of Garry Kasparov delivering the Daniel Pearl Memorial Lecture, it is available now via this link: @AsraNomani @ThePearlPost @MissDiagnosis @Ostrov_A @StandWithUs @JGreenblattADL (posted:

3.18.2023 1:38am - (Replying to @RonKenett) "Looking at the data generation set up" - who doesn't?

3.17.2023 12:34pm - In the beginning, "pro-Palestinian" was used as a cover for anti-Israel eliminationists to voice their racist ideology in respectable circles. In time, the correlation between "pro-Palestinian" and racism became so obvious that even innocent pro-Palestinians now bear the stigma. [reposted:]

3.17.2023 11:21am - (Replying to @JamesGrace111) Economists do rely on causal knowledge, since knowledge consists of assumptions and they do make assumptions. However, their assumptions are cast in such a screwy language (ignorability) that one is rarely sure if they reflect what one knows about the world.

3.17.2023 4:25am - Belated thanks to @StandWithUs for giving me the honor of addressing their international conference, and sharing with students from all over the world some of my experience in fighting Zionophobia on campus, as summarized here: reposted:]

3.17.2023 2:12am - (Replying to @DrTomasAragon) The intentions of @unbiasedscipod are pure, but they talk as if we are still in some pre-revolution hand-waving century, as if we still do not know how to address confounding. I hope they take you seriously and read #Bookofwhy.

3.17.2023 1:12am - (1/2) Just in, new breakthrough! Remember the infinite debates concerning "Personalized Decision" and "Concurrent Control"? Well, new results posted here: show how observational data inform P(benefit) and P(harm) when experimental
3.17.2023 1:12am - (2/2) studies are restricted by concurrent control to assess only ATE = P(y_t)-P(y_c), not the individual probabilities, P(y_t) and P(y_c). This may not stop the infinite debates but it surely opens up new opportunities for personalized decision making.

3.17.2023 9:22pm - A people with no history would spare no lie, to steel one from its neighbor. [reposted:]

3.16.2023 12:34pm - Regarding "going beyond quasi-experiments", the best I can recommend would be my comments on "Pearl before Economists":, especially Section 3 and "Conclusions". My rebuttal of Imbens should also be of interest: [reposted:]

3.16.2023 1:22pm - (Replying to @bentriderdeb) Your leadership is still felt at every corner of this school.

3.16.2023 5:45am - I believe it is important that we RT such cases, to alert and inform universities like Leeds that their academic reputation has been badly damaged irrespective of whether they admit guilt or not. I trust the story of Danielle Greyman over that of Leeds. [reposted:]

3.16.2023 2:27am - I hope the Jewish community does not accept this empty statement by the Principal as an apology - the words Israel & Zionism do not appear there, not even "Jewish". Fighting hate in the abstract is fighting no hate at all. [reposted:]

3.16.2023 12:31am - (Replying to @GadSaad) Are you talking @USC ? The University whose President has appointed a (don't laugh) "task force", sorry an "Advisory Committee on Jewish Life at USC," in January 2022? See I wouldn't be surprise if the Chair of that Committee was dancing at your talk.

3.16.2023 - 12:21am - A people with no history can't respect others'. [reposted:]

3.16.2023 12:16am - A people with no history, must fabricate one. A culture born in lies, cannot not change its habits. [reposted:]

3.16.2023 2:04am - The thing to notice is who is NOT on the list of signatories. (1) Berkeley Dean Chemerinsky, (2) Anyone from the Law School of my University, UCLA. Deeply disappointed, and hoping they will join later. [reposted:]

3.16.2023 12:02am - I'm not surprised to see Zionophobes such as Conell West, Ilan Pappe and Jewish Voice for "peace" signing this letter, but Peter Gabriel is a disappointment and I have been still hoping for Noam Chomsky to come to reason. After all, he is a scientist. [reposted:]

3.15.2023 10:06pm - (Replying to @sbuhai and @eliasbareinboim) Are you implying editors of Econ journals are blessed with higher than average curiosity about relevant innovations in neighboring fields?

3.15.2023 9:19pm - Grateful for an exciting family visit to the Daniel Pearl Magnet High School, and for the opportunity to talk with students and teachers about the unique educational projects initiated in this school. From strength to strength! (Yes, the guy in the Bolshevik hat is me). [reposted:]

3.15.2023 8:50pm - (Replying to @bentriderdeb @ThePearlPost and @maggie09032527) In the secrets of my heart I was hoping you @bentriderdeb will appear, suddenly, as you usually do, and join us on this exciting visit to the DPMHS today.

3.15.2023 8:17pm - No way! It's Brandeis EDI officers that should be sent to remedial training. Have they designated an ombudsman for Jewish/Israeli students? [reposted:]

3.15.2023 8:10pm - A good start, but will remain mute, ineffectual and actionless until the word "Zionism" is spelled out in all its six letters, and recognized as an integral part of Jewish identity. [reposted:

3.15.2023 7:58pm - I love his Purim Hat, if only his words weren't taken seriously by his parishioners. [reposted:]

3.15.2023 7:49pm - What does @IlhanMN say to my friend Yasmine? [reposted:]

3.15.2023 6:52pm - The great Abba Evan: "All nations determine their policies in the light of self-interest and then explain them in terms of self-sacrificial altruistic morality.” Humble me: "Are people fundamentally different?" [reposted:]

3.15.2023 5:37am - Last I checked, it is logically impossible to be a Zionist and hate Palestine. Look it up in your encyclopedia: Zionism stands for two equally legitimate states. I must see TLOU.

3.15.2023 4:59am - (Replying to @KosherRedPill) Where did they find so many Bundists? The dust-bin of history? Birobidzhan? [posted:]

3.15.2023 1:58am - Multi-Congratulations to Paul and @eliasbareinboim for accomplishing the impossible - getting a paper on modern causal inference published in The Econometrics Journal; the face of econometrics will never be the same, and the face of causation will shine brighter each passing day. [reposted:]

3.15.2023 1:29am - (Replying to @sbuhai and @eliasbareinboim) I used "impossible" figuratively, to mean "very very hard". Indeed, it took Paul and Elias's paper 4 years to get published, which, in this fast moving field means "eternity". Most importantly, mission accomplished; economics students can now see what they missed passed 4 years.

3.14.2023 7:59pm - The New York Times’ blatant antisemitism via @JNS_org [posted:]
3.14.2023 7:59pm - I wouldn't go as far as calling The New York Times "antisemitic", but the numbers are fascinating, and I like linguistic counting, because it so often reveals sentiments that their authors try to hide. Even readers do not notice them till they bake into their psyche.

3.14.2023 5:12pm - This week we are blessed with a swelling 6.5 pages of new papers in causal inference Some of the titles are extremely inviting, for example: Causal Deep Learning which refines the Ladder of Causation. Enjoy, and share breakthroughs.

3.14.2023 4:42pm - (Replying to @nbarrowman) Quantifying the strength evidence is difficult, similar to the difficulty of quantifying faithfulness, or incidental cancellation. See Section 6.4

3.14.2023 1:47pm - (Replying to @selkirkragdoll and @eliasbareinboim) Big data and very high dimensional time series do not produce causal relations.

3.14.2023 1:43pm - Virginia seems to be a cesspool of deceit, but if @IlhanMN joins the party, it will spread beyond Virginia. Thanks for fighting the Pagans.

3.14.2023 1:24pm - You were too charitable to @guido_imbens by citing his 2020 "PO and DAGs" without my rebuttal Some readers might not see where he errs (most will). [reposted:]

3.14.2023 12:39pm - This is the first "video review" of #Bookofwhy that I see. Unique media that allows you to see when a reviewer is excited and when he hesitates. Very refreshing to see how causation shines in the eyes of a reader who doesn't pretend to already know everything about causation. [reposted:]

3.14.2023 12:25pm - (Replying to @selkirkragdoll and @eliasbareinboim) What weakness you find in what theory?

3.14.2023 6:13am - Sometimes even self-evident truth needs a trumpet to be heard. When the Flat Earth Society kicked off its last "The Earth is Flat Week", there wasn't even a single dissenting voice. No more! [reposted:]

3.14.2023 4:56am - (Replying to @academic_la and @BondWolfie) Very moderate comment indeed, which would only acquire Israelis trust if accompanied by the litmus test: An "Equally Indigenous" pronouncement by the caller.

3.14.2023 12:39am - (Replying to @elderofziyon) Well said. This brave young man would not be as brave facing a Palestinian Authority security man, not to mention Hamas.

3.14.2023 12:12am - (Replying to @HuelleAlexander) You are right -- a glaring omission. I'll inquire.

3.13.2023 7:45pm - (1/2) - For readers who asked about foreign publications of #Bookofwhy, here is what our publisher tells us: 1. US Basic Books, 2. UK Penguin Books , 3. SPANISH Pasado y Presente, 4. HOLLAND Maven, 5. FINLAND Terra Cognita, 6. PORTUGAL Temas y Debates, 7. JAPAN Bungei Shunju, (more)
3.13.2023 7:45pm - (2/2) - 8. KOREA ScienceBooks, 9. TAIWAN The Walk Publishing, 10 CHINA CITIC, 11. ISRAEL Kinneret-Zmora Dvir Publishing, 12. RUSSIA AST, 13. POLAND Copernicus Center Press, 14. TURKEY Ginko Publishers. Glaringly, no French publisher found thus far --is it the love of staying behind?

3.13.2023 11:57pm - (Replying to @TBroekel @PHuenermund and 6 others) I do not think the problem is in the idea of consulting history as it is in consulting history correctly, using methods that economists developed to consult other non-experimental data. For a glimpse, see,

3.13.2023 11:41pm - When I hear the title " Presidential Task Force to Combat Antisemitism" I immediately recognize a Presidential Instrument to License Inaction. Things my be different of course at GWO, depending on whether students acquire two key weapons: @sabrinasoffer [reposted:]

3.13.2023 4:04pm - At long last! A non-boring, in-sight-ful news item on CNN. [reposted:]

3.13.2023 4:21am - Seemingly an anecdotal piece of data, to be added to my list of arguments that Jews are a people bonded by collective memories, more than by any other force. [reposted:]

3.12.2023 10:21pm - (Replying to @JeremyBenAmi) I see no reason to apologize. Jimmy Carter was the first to label Israel "apartheid", the first to give terror an ideological legitimacy (see and the only US President to be remembered as a Lover of Hamas. [reposted:]

3.12.2023 8:03pm - (Replying to @viraldogenes) Indeed, some of these people are protesting, not the judicial reform itself, but the disrespectful way it is bulldozed over the country and, more so, the grand sell-out to anti-liberal extortion groups who hope to benefit from their support of the reform.

3.12.2023 7:50pm - We should be overjoyed that protesters everywhere have made the Israeli flag the symbol of their disapproval with the government. We should reject protesting hypocrites like @jvplive and @NewIsraelFund for whom the Israeli flag has always been a symbol of ultimate evil. [reposted:]

3.12.2023 6:06pm - (Replying to @lal_shiban) Why? What does he have against Mexico existence?

3.12.2023 6:04pm - (Replying to @tdietterich) Deliberation overriding intuition is exemplified nicely in game playing, the former emerges from search and the former from static evaluation function.

3.12.2023 5:20pm - I would add three more -- Playful manipulations -- The laws of physics -- Inborn templates in which to store causal knowledge once it is acquired [reposted:]

3.12.2023 4:47pm - A record half a million Israelis protest judicial reform, announce 'Day of Escalating Resistance' | World Israel News [posted:]

3.12.2023 4:14pm - Last I heard, America has not denied its neighbors', Canada and Mexico, right to exist. Palestinians do, Putin does, and so do many gullible readers who side with the latter two; none heeded my call to join the "equally indigenous" camp. [reposted:]

3.12.2023 11:37am - (Replying to @EricJGreenberg1) These British-minted coins were labeled "beautiful old Palestinian coins" for external deception. I don't believe any Palestinian has nostalgic memories of the 1936-1942 riots, when thousands were brutally killed by British troops. "Government" was not even in one's dream.

3.11.2023 5:03pm - Lapid proposes constitution for Israel to extract itself from ‘terrible crisis’ via @timesofisrael [posted:]

3.11.2023 4:56pm - (Replying to @partygrrrl65 @baroquenearth and @maradonasattva) "Separate us from them" is what we, Americans, hear from our Canadian neighbors, and they don't expect us to label them "apartheid", for a good reason: separation helps emphasize our uniqueness and rejoice our diversity and mutual respect.

3.11.2023 4:41pm - I hope you get a chance to convey my humble message to them: "The way you treat Israel is the litmus test of your interfaith objectives." [reposted:]

3.11.2023 1:36pm - Some readers were "shocked" at my attempt to explain why 1930-40 coins minted under British colonial rule are used by Palestinians to claim exclusive ownership of the holy land. I am retweeting my answer with the hope recruiting those readers to the "equally indigenous" camp. [reposted:]

3.11.2023 12:49pm - (Replying to @maradonasattva) Supremacists are those who deny their neighbor's right to existence, not those who uphold this formula for peace: "Two states for two people, equally legitimate and equally indigenous" (Pearl, 2018)

3.11.2023 4:25am - (Replying to @aayhaq) When people are trying so desperately to fake history, see, the question arises as to why they do it. What's your explanation? [reposted:]

3.11.2023 3:46am - (Replying to @AlexanderNaume2) Sorry, I have a prior obligation to my students, and I am familiar with their ideas.

3.11.2023 3:42am - (Replying to @Dam_Nuwen) Except those who fake it.

3.11.2023 3:39am - (Replying to @mangefort) When people are trying so desperately to fake history, see, the question arises as to why they do it. What's your explanation? [reposted:]

3.11.2023 2:40am - (Replying to @aayhaq) A people secured in its history need not dig 1936 coins to prove their neighbor's non existence.

3.11.2023 2:19am - A beautiful example of cause-effect reversal for the next edition of #Bookofwhy. One can pick up dozens such examples by searching for the phrase "it should be clear" or "everyone knows" in the writings of @MiddleEastInst analysts. [reposted:]

3.11.2023 1:29am - A people with no history, in frantic search for non-existent roots. [reposted:]

3.11.2023 12:18am - We are informed that the submission deadline to the Nineteenth Conference on Theoretical Aspects of Rationality and Knowledge (TARK 2023) has been extended. Website: [posted:]

3.10.2023 10:33pm - Yes, sure: [reposted:]

3.10.2023 12:41pm - The Book of Why is translated into 14 languages. The last one, Hebrew, will be published Sept. 2023. Looking forward to it. [reposted:]

3.10.2023 11:38am - (Replying to @mapto @JosepBorrellF and @ProSyn) No, I am talking about a much deeper rejectionism, one that negates the very notions of "borders" and "negotiation", one that denies the existence of a sovereign entity with which to have "borders" and with which to "negotiate". Settlements are but a pebble on this rejection wall

3.10.2023 3:59am - (Replying to @mapto @JosepBorrellF and @ProSyn) Yes. But the word "rejects" is misleading in its softness. We are facing a side that fights a two-state solution tooth and nail, to the last person, at all costs.

3.10.2023 3:55am - (Replying to @PeterS80148175) If I didn't believe this is possible, I wouldn't be working in AI. I would probably join some deep-learning group and do curve-fitting for living.

"I'm proud of being programmed at Deep Mind" said one Robot to the other. "Why?" said the other. "Because chances are they seeded into me some neat programming tricks I saw in AlphaGo; at least I would like so believe I got some, it makes me wish to dare the impossible." [reposted:]

3.9.2023 4:26pm - They don't celebrate murder! Far from it. They celebrate seeing stupid Europeans like @DTzantchev speaking about "cycle of violence" again, knighting them with legitimacy of reciprocity. · [reposted:]

3.9.2023 3:47pm - "I'm proud of being programmed at Deep Mind" said one Robot to the other. "Why?" said the other. "Because chances are they seeded into me some neat programming tricks I saw in AlphaGo; at least I would like so believe I got some, it makes me wish to dare the impossible." [reposted:]

3.9.2023 1:52pm - This country is splitting too. [reposted:]

3.9.2023 1:41pm - (Replying to @Inframethod and @EnglishOER) Blessed are the hunches, on their march to wisdom.

3.9.2023 1:30pm - And most Europeans fall for it. [reposted:]

3.9.2023 1:23pm - Still recovering from an inspiring evening with Chess grandmaster and political activist Garry Kasparov : [posted:]

3.9.2023 12:35pm - (Replying to @JosepBorrellF and @ProSyn) The paradigm shift must start with EU realizing that one side would fight tooth and nail against your stated goal: to see a safe, secure, globally recognized state of Israel live in peace alongside ...freedom, prosperity, and peaceful relations with their neighbors. Tooth+nail.

3.9.2023 4:07am - (Replying to @Ostrov_A) Better.

3.9.2023 1:31am - Sad news. Chaim Topol, star of "Fiddler on the Roof" is dead at 87 On a personal note, Topol served in the IDF with me (basic training, 1953). He was able to make the whole platoon laugh to the ground, including the toughest commanders. [posted:]

3.9.2023 12:51am - Please see (or re-write) the History section of "Confounding" on Wikipedia:

3.9.2023 12:47am - (Replying to @stephensenn) Please see (or re-write) the History section of "Confounding" on Wikipedia:

3.9.2023 12:37am - (Replying to @stephensenn) Nice opening. But I lose you here: "His approach starts from the recognition that it's impossible to eliminate the effect of such nuisance factors in practice". I don't see why these are "nuisance factors" if we haven't defined the research question: "causal effect". What is it?

3.9.2023 12:19am - Asked if I would recommend this paper Of course! It's one of my clearest! BTW, the statisticians who insisted that "confounding is solidly founded in standard, frequentist statistics" talked about "effect-biased" type of confounding, not Fisher 1926 type. [reposted:]

3.8.2023 11:04pm - (1/2) (Replying to @Corey_Yanofsky) Agree. And that was in 1974, while mathematical statistics has been in existence since 1713 (Bernoulli). So, there has been some tectonic paradigmatic shift 50 years ago, before which no one could say "causal effect", nor express what "confounding" is. Now, to appreciate
3.8.2023 11:17pm - (2/2) (Replying to @yudapearl and @Corey_Yanofsky) the role of SCM in this tectonic shift, one need only go to Imbens & Rubin's book (Cambridge, 2015) and look up their definition of "unconfoundedness" (p.479). This would be enough to convince every mathematical statistician that the shift has settled in 1993, not 1974.

3.8.2023 8:33pm - It's easier to convince the Emperor of China than Mathematical Statisticians that a Collider is not just "Jargon of Structural Causal Models" but a piece of reality, deserving a name. Greeting to the Emperor. @analisereal [reposted:]

3.8.2023 5:16pm - (Replying to @QRDL @tdietterich and 3 others) Beg to differ. Causal reasoning is precisely what it advertised to be.

3.8.2023 4:56pm - (Replying to @omaclaren @analisereal and @learnfromerror) Even if I agree that your framework unifies SCMs and other approaches to causal inference, I would call it "Oliver's framework," not "Statistics provides framework". Because statistics had no framework at all till Rubin, and no operational framework till SCM. History matters.

3.8.2023 2:56am - (1/2) This is Rubin's famous "clutter paper": "I find much of what is written about causal inference to be mathematically inapposite in one of these senses because the descriptions either include irrelevant clutter or omit conditions required for the correctness of the assertions" [reposted:]
3.8.2023 2:56am - (1/2) The historical account here is as complete and accurate as the act of discarding the works of Wright, Duncan, Haavelmo, as "clutter", not to mention Structural Causal Models, which give Potential Outcomes their scientific standing.

3.8.2023 12:52am - (1/2) Delighted to share another stream of CI articles published this week: I choose to highlight "Combining Graphical and Algebraic Approaches for Parameter Identification in Latent Variable Structural Equation Models" by Ankan, Wortel, Bollen and Textor:
3.8.2023 12:52am - (2/2) which answers several of my questions about the relationships between traditional SEM and modern graphical methods of Causal Inference. Inference under Measurements Error is the main beneficiary of this symbiosis.

3.7.2023 8:54pm - A powerful video by Arnold @Schwarzenegger on hatred and haters. Readers write that haters like Kanye West and Louis Farrakhan should listen. True, but it's more important that @HRW -clocked haters like @KenRoth and @FranceskaAlbs listen, who consider themselves angelic. [reposted:]

3.7.2023 6:19pm - (1/2) (Replying to @omaclaren) Of course statisticians have been concerned with more than just prediction and association, But until Rubin 1974 they were using a language built strictly on prediction and associations. This disparity is what drove them crazy up to Rubin, and what drives them crazy today.
3.7.2023 6:33pm - (2/2) (Replying to @yudapearl and @omaclaren) To elaborate, statisticians who remained outside Rubin's camp declared causality a "subject-matter" issue, lying outside statistics. Those who joined Rubin's camp had to define confounding using ignorability assumptions, i.e., beyond comprehension (See Imbens & Rubin, 2015)

3.7.2023 3:45pm - Illuminating discussion, not to be missed. My takes: (1) Expressional power matters. (2) Explainability, transparency and having world-models are more important than we thought. [reposted:]

3.7.2023 3:35pm - And in case you hesitate, and we all do sometimes, let's remember what Mordechai told his niece, Queen Esther: "And who knows, perhaps you have attained to royal position for just such a moment.” [reposted:]

3.7.2023 2:16pm - (Replying to @yudapearl @ewenharrison and @omaclaren) No, a causal interpretation is all you need. My point was that the discussion we are having on "confounding" is not a "turf war" but an attempt to clarify notations and vocabularies.

3.7.2023 2:13pm - (Replying to @ewenharrison and @omaclaren) No, a causal interpretation is all you need.

3.7.2023 2:11pm - (Replying to @omaclaren and @learnfromerror) No need to link this again, just an answer: in what way your definition of statistics is not 'everything' (it already embraces the entire Ladder of Knowledge) and whether you don't find there elements that most 20th century statisticians tried to exclude from "statistics."

3.7.2023 1:32pm - (Replying to @omaclaren and @learnfromerror) My characterization may differ from yours, but it is at least crisp, well defined and same as Fisher's ("summarization of data") and most statisticians of 20th century. You are now claiming it's too narrow, fine, let's hear your definition, and in what way it's not 'everything'.

3.7.2023 1:25pm - (Replying to @omaclaren and @learnfromerror) In the transition between my definition of statistics (i.e., properties of the distribution that governs the data) and yours, there must be at least one primitive relation that is part of the latter, but not the former. Can you identify such relation?

3.7.2023 1:13pm - (Replying to @omaclaren and @learnfromerror) I would be equally curious to find out how @learnfromerror defined "confounding" last time a student asked. Did she send the student to read the "accounts developed along these lines"? Or used "prediction" and "association" as the basic primitives in the definition?

3.7.2023 12:58pm - (Replying to @omaclaren and @learnfromerror) No contention! If statistics is defined as a language for expressing every scientific thought, past and future, real and imaginary, then obviously it can define "confounding". I'm still curious though, how did you define "confounding" last chance a student asked.

3.7.2023 12:43pm - (Replying to @ewenharrison and @omaclaren) This is not a "turf war", it is an attempt to define languages, assess their limits and enrich them when necessary. The importance of this enrichment surfaces when a stat professor tries to define "confounding", and can't find the vocabulary needed to do so. Can you try?

3.7.2023 12:20pm - (1/2) Statisticians would go to great acrobatics rather than admit neglect. When an academic colleague asks me “Aren’t you being a little unfair to our discipline, considering the work of so and so?”, I answer “Must we speculate on what ‘so and so’ did? Can we discuss how YOU [reposted:]
3.7.2023 12:20pm - (2/2) define "confounding" in your stat class (rephrased from The conversation immediately turns realistic; from hand-waving of "so and so" to the realization that statistics lacks the vocabulary needed to define "confounding."

3.7.2023 11:56am - Free translation: The key struggle in Israel focuses on the question whether the ultra-orthodox and the messianic sects will be allowed to shape the entire State in their image. These sects undermine the ideological foundation of Israel -- secular Zionism. [reposted:]

3.7.2023 11:31am - (Replying to @omaclaren) Let's hear your definition of "the language of statistics", so as not to "mis-define" it the way some crooks do, for the own selfish purposes.

3.7.2023 10:49am - (1/2) Why can't statisticians admit that "confounding" cannot be defined in the language of "prediction" or "association"? Because they are brought up to believe that statistics is the ultimate language of experimental science; so how could this super language be so deficient as [reposted:]
3.7.2023 10:49am - (2/2) to prevent us from defining such notions as "confounding", which even Pearson, Yule, Neyman, Fisher and Simpson recognized as central to statistical thinking? A traumatic cognitive dissonance! D. Lindley was the only statistician to admit: "This super language was deficient!"

3.7.2023 8:32am - (Replying to @adrianachavira @bentriderdeb and @ThePearlPost) I was telling @bentriderdeb that we missed the students reporters from @ThePearlPost who normally meet with the speaker before the Lecture. Let's plan for next year.

3.7.2023 8:25am - Thank you @bentriderdeb for posting these photos from last night's Lecture of Garry Kasparov. It surpassed all expectations, and we were delighted to see you there, with so many supporters of freedom and democracy. I hope they post a video later on.

3.7.2023 8:18am - You are right! #Bookofwhy is a glaring omission from this pile of mighty good books. Unbiased observer.

3.7.2023 7:56am - (Replying to @leonpalafox and @GaryMarcus) "Making stuff up" -- This has been the pride of Al Jazeera and affiliated news agencies, from which we could protect ourselves. No more. It's now in our bedrooms, guised as "ChatGPT Lectures".

3.7.2023 7:32am - (Replying to @Martin_Kramer) This is close to what Sheik Yusuf Qaradawi told his Al Jazeera listeners: "Take from the West all its science and technology, but you must reject Western philosophy, because it is corrupt at its roots by the pagan philosophies of Greece and Rome"

3.6.2023 9:43am - (Replying to @noah_greifer @JohanDH2O and 2 others) Recall: "Equally biased" is true even under unmeasured confounding. See

3.5.2023 7:41pm - (Replying to @daoudkuttab) Marwan Muasher's understanding of Israeli intentions is pathetic. Israel's withdrawal from the West Bank depends not on Palestinian demography but on Palestinians mentality; it can't materialize as long as Palestinians believe, as they sadly do, that Tel Aviv is stolen land.

3.5.2023 5:20pm - Happy Purim, sister! We didn't have an army when I was your age, my costume was a donkey-head. You made a better choice, treasure! Purim Sameach!

3.5.2023 6:30am - (Replying to @HuelleAlexander) They are trying, but the spoiled youth are on my side, so they can't succeed.

3.5.2023 4:18am - Asking @PaleNews not to lie is like asking an eagle not to fly, or a fish to stay dry. [reposted:]

3.5.2023 3:52am - My brothers and sisters are there, demonstrating for democracy, but what they are really demonstrating for is decency, a scare commodity in the current coalition government. [reposted:]

3.5.2023 3:32am - Divisiveness is not a crime, and producing antisemitism is something they would surely deny. They should be shut down for spreading Zionophobia, a form of racism they can't possibly deny. [reposted:]

3.5.2023 3:17am - (Replying to @academic_la @FaithQuintero7 and 2 others) It is also dangerous to frame the Palestinians as a people "with ambitions we need to respect" without mentioning their principal, top-top ambition which we cannot possibly respect. Such framing encourages them to use the former as a cover to advance the latter.

3.5.2023 1:01am - (Replying to @_AntoninoGreco and @AleksanderMolak) "data are always better than theories!" ??? I have hard time agreeing to the "always". Why? Because #Bookofwhy is repleted with examples where data alone are embarrassingly helpless, if not misleading. Care to go there?

3.4.2023 10:19pm - (Replying to @pbosque and @blakeflayton) Ask @AliAbunimah if he can name a pre-1900 Palestinian figure that he is proud of? A pre-1900 poem that he enjoys reciting? A pre-1900 national holiday that he celebrates? (Rephrased from [posted:]

3.4.2023 9:59pm - (Replying to @academic_la @rookringer and @RenzoGraciani) I am ready to rush and admit the other has a right to be on the land and to have a national home, but I would condition the latter right on accepting our symmetrical right, and I would even go as far as denying that right as long as ours is denied and, sadly, it is!

3.4.2023 9:47pm - (Replying to @BillBainbridge5) Especially given that our cognitive architecture is formed while the baby is immersed in a universe of comfort and care; the transition to a universe of indifference and pain demands a God to mitigate the unbearable cognitive dissonance,

3.4.2023 9:18pm - The origin of this misconception stems from the ancient belief that "confounding" is a statistical notion, and can be therefore defined in the language of "prediction" or "association". It can't. This drives some statisticians crazy, even today. See [reposted:]

3.4.2023 9:03pm - Who says our mind is not ruled by religion? Every time I mention God, I gain 100 new followers on this channel. If we do not program God into robots, they would surely discover one(s) on their own, just to be able to serve us better; we, their God-ruled users. [reposted:]

3.4.2023 5:16pm - (Replying to @DrCMcMaster and @John_Roberts_3) And I think that reducing ideas about inborn structures to the platitude of "primarily compute cause (and everything follows from that)" is an unhelpful reductionism.

3.4.2023 4:59pm - (Replying to @DrCMcMaster and @John_Roberts_3) Wait till your baby goes through the "playful exploration" phase, and see if curve-fitting ML would explain his/her behavior.

3.4.2023 4:49pm - Brilliant!!! People without a history are telling a people bonded by history where exactly its national homeland should be. @blakeflayton [reposted:]

3.4.2023 3:28pm - (Replying to @DrCMcMaster) Let's hear some non-shallow thoughts. Seriously, I am truly curious, and I am sure many of my co-workers in ML and CI are equally curious to hear some words of wisdom from another school of thought.

3.4.2023 2:18pm - (Replying to @Anatomi54244944 and @GaryMarcus) I would presume (in my innocent naivete) that such difficulties would be a thing of the past in post-causal-revolution science, where "defining, deriving, testing and using" has ceased to be a mystery.

3.4.2023 9:59am - (Replying to @CriticalAI and @GaryMarcus) Are you calling for the details of such program, or question its feasibility?

3.4.2023 9:51am - (Replying to @Anatomi54244944 and @GaryMarcus) In saying "we all believed in God" I meant WE as a humanoid species. And I don't know any of us who do not use causal vocabulary in our everyday speech which, to me, indicates causal thinking in the brain behind the speaker. Where does this resistance to causality come from? ML?

3.4.2023 9:34am - Pro-Palestinian activist went looking to beat Jews at NYC Israel rally – prosecutors via @timesofisrael [posted:]
3.4.2023 9:34am - This Zionophobic organism is right, he was merely looking to beat up Israel supporters like me, not Jews per se; how can anyone blame him for being hateful? Careless, maybe, but hateful? No way!

3.4.2023 1:52am - (Replying to @kaalam_ai and @ylecun) RL takes us to Rung-2 of the ladder; one more rung to climb before we reach a state of "understanding".

3.4.2023 1:22am - (Replying to @_AntoninoGreco) Change "innate" to "inborn."

3.4.2023 12:56am - Had the privilege and great fun tonight, to speak to 500 students at the StandWithUs International Conference - more students than I've ever seen in one room, singing of resilience and hope. I hope to have inspired them to try these two new weapons: [posted:]

3.4.2023 12:21am - The reason ALL of us believed in God and other metaphysical things is that our mind craves for causal explanations and we must satisfy this craving first; truth comes later. To act human-like, machines too need be programmed with such craving; truth to come later. @GaryMarcus [reposted:]

3.3.2023 3:13pm - (Replying to @BethCarey12) I actually meant that the reason ALL of us believed in God and other metaphysical things is that our mind craves for causal explanations and we must satisfy this craving first, truth comes later. Machines, likewise, need be programmed with such craving, truth comes later.

3.3.2023 2:05pm - (Replying to @FriedrichHayek) I confess to have missed Larry Wright in my philosophy readings. Where would I go wrong if I pursue the quest for explanation along causal tracks (as in Ch. 9 of Causality), without reading Wright?

3.3.2023 1:55pm - (Replying to @policytensor) ChatGPT speaks like an economist: "they are not specifically designed to model causation in the sense that Pearl defines it", which I would love to pursue with: "is there another sense to define it?" or "Forget Pearl, should we ban smallpox inoculation?"

3.3.2023 1:37pm - When I hear "humans aren't particularly good at causal inference, otherwise astrology wouldn't be a thing", I say "humans aren't particularly good at thinking, otherwise we wouldn't have politicians". We still want machines to think though.

3.3.2023 7:02am - (Replying to @XiXiDu) Beautiful example of Milton Friedman. Note that the negative correlation between oil burned (M) and outside temperature (V) is a collider effect, caused by keeping the inside temperature (i.e., the collider, P) constant. @zooko

3.3.2023 3:03am - (Replying to @soboleffspaces) Given the equation: ATE = P(Ben)-P(Harm), one may get the impression that the "most appropriate" treatment is the one that maximizes ATE. But this is not so; each individual may assign different weights to P(Ben) and P(Harm).

3.2.2023 10:39pm - (Replying to @joelkaka) I do not collect them. In my mind, every observational study in healthcare domain is using some of the methods developed in CI. How else would you select covariates for adjustment? See (Unless you are an economist, which gives you a license to guess.)

3.2.2023 9:21pm - Amid anti-Zionism, antisemitism, Reform Jews are betrayed by rabbis - opinion [posted:]
3.2.2023 9:21pm - Reformed Jews are not the only ones betrayed. Jewish students of all denominations are betrayed by their campus leaders (ask me about the Center for Jewish Studies) and community leaders (ask me about the Committee for Jewish Life at USC). Our only fighting word: Zionophobia.

3.2.2023 8:30pm - I am thrilled that recent works on personalized decisions and P(Harm) have found their mythical/historical roots in the form of the Hippocratic Oath: "First, do no harm." Finally, the medical profession is prepared to enshrine its most cherished oath with a formal definition. [reposted:]

3.2.2023 8:14pm - The same smallpox logic dominates the arguments against a two-state solution. It goes: Palestinians fought tooth and nail against a 2-state solution and prevented it from taking root. So far, it hasn't taken root. Conclusion, it's an unworkable solution; we must go one-state. [reposted:]

3.2.2023 6:21pm - (Replying to @soboleffspaces and @smueller) You are right. It is only the upper bound that provides us with individual information, going beyond CATE.

3.2.2023 8:09am - (Replying to @ak47_2013) Nationality: Israeli-American Religion: Practicing disbeliever

3.2.2023 3:37am - Another paper on unit selection is now available on the arXiv: It deals with the complexity of finding optimal set of units, given a fully specified structural causal model (SCM). [posted:]

3.2.2023 3:07am - (Replying to @gilled34 and @daoudkuttab) It is stoppable with great effort, bordering on civil war, worth risking if an end of conflict comes into sight. Not worth risking when end of conflict is shown more and more impossible by honest neighbors.

3.2.2023 2:13am - (1/2) A reader pointed us to interesting connections between our paper on "Personalized Decision Making" and a recent paper on asymmetric utilities Note that the Hippocratic Oath: "Do no Harm" is interpreted to mean that
3.2.2023 2:13am - (2/2) the cost of causing death to a patient who would otherwise survive without treatment is greater than the cost of forgoing life-saving treatment. Our ability to assess both P(Benefit) and P(Harm) opens up ways of tailoring the Hippocratic Oath to fit individuals preferences.

3.1.2023 11:57pm - I begin to believe in divine hands. Elan's murderers were just apprehended, and the news reached his parents, while attending his burial. Baruch Dayan Emmet. [reposted:]

3.1.2023 11:50pm - (Replying to @Blogsbloke) Interesting point. Deterrent effect is frequently mentioned in the context of terrorism, not in the context of property seizure. I am not familiar with 'scholarly" articles on deterrence; the DAG could be illuminating.

3.1.2023 10:26pm - (Replying to @dylanarmbruste3) Perhaps you can articulate the relation between the two using the language of causal inference? I can't relate to the article, it hides the research question, and talks about "disturbances" and "successive plot" w/o defining them. You would be a hero if you could translate.

3.1.2023 9:24pm - (Replying to @BabyYoneda and @daoudkuttab) Israelis are humble and least demanding; all they want to hear is about Tel-Aviv.

3.1.2023 8:26pm - Perhaps Israel should remind the critics that its anti-terror actions are consistent with the wise counsel of the late Justice Ginsburg and her colleagues. [posted:]
3.1.2023 8:26pm - Interesting legal parallels between "demolishing terrorists' homes" and "seizing a criminal’s property". I wish more legal minds would weigh on it.

3.1.2023 4:42pm - (Replying to @gilled34 and @daoudkuttab) It isn't absolutely necessary, agree, but it turned unstoppable when it became clear that an end to the conflict was not anywhere within our neighbors' sight (and still isn't.)

3.1.2023 1:26pm - Israelis have never expected Palestinians to "surrender"; they expected them to understand that the occupation is a temporary necessity; awaiting some acknowledgment that Tel Aviv is not a stolen land. None thus far. @daoudkuttab [reposted:]

3.1.2023 1:14pm - (1/2) The smallpox paradox (see #Bookofwhy) is now used by Palestinian intellectuals to justify the elimination of Israel: "More people died by smallpox inoculations than from smallpox itself, therefore, ban inoculations." It took Palestinians 75 yrs to collect the data needed [reposted:]
3.1.2023 1:14pm - (2/2) to justify what they have been trying to do with or without the data; who said they are not acting logically?

3.1.2023 12:36pm - (Replying to @HL327 @smueller and 10 others) The choices were extremely well informed in our example, because it was purposely made extreme to demonstrate a principle. For less extreme examples, our results give the maximum information that can be extracted from ANY combination of RCT and OS. The higher the bias, the better

3.1.2023 11:58am - (Replying to @HL327 @smueller and 10 others) If S=s is the factor that triggers choice in OS, then the whole point of our paper is to show that, though S is NOT observable in either the RCT or OS, we can still estimate P(ind. ben) and P(ind. harm) from their combination. I agree that life is much easier once we observe S.

3.1.2023 9:06am - Disagree. While religious Jews have difficulties saying that these aren't real Jews, Zionists have no qualm at all saying these aren't Zionists. They aren't. Their attacks on civilians and on the IDF, puts them solidly on the same moral plane as Palestinian terrorists, BDS & JVP. [reposted:]

3.1.2023 1:35am - This week brought with it another stream of causal inference articles:, My curiosity got caught by: Causal Inference Methods for ... Decision-Making, by Ioana Bica's PhD dissertation.

2.28.2023 11:59pm - I was waiting for this arrest, but why did it take so long? [posted:]

2.28.2023 11:52pm - The mantle of Zionism is not a birth right, is not handed free; it's got to be earned. [reposted:]

2.28.2023 1:08am - Next to @PeterBeinart , Rothman is the greatest mis-interpreter of the Torah. I only hope he doesn't call himself a Zionist. Peter stopped. [reposted:]

2.28.2023 12:53am - The demand for this new profession is rising rapidly: ChatGPT3 Psychoanalysts . "What's wrong with him?" "Deprived childhood perhaps" "Is he still redeemable?" [reposted:]

2.28.2023 12:41am - Kol Hakavod, YaYa King, I hope you can donations from Zionists abroad. [reposted:]

2.27.2023 9:35pm - May your memory forever be a blessing, Elan. I cry for you, and I feel for your family. [reposted:]

2.27.2023 3:27pm - Bipartisan Congressional Letter Calls for UN to Disband Commission of Inquiry, Fire Palestine Rapporteur [posted:]
2.27.2023 3:27pm - In a day marred by bad news all over, it's reassuring to see some wisdom emanating from the US Congress. Honestly, has there ever been a more unfit UN "rapporteur" than our Lady @Francesca ?? @Ostrov_A

2.27.2023 2:56am - (Replying to @gypsybeepie @AsraNomani and @CAIRNational) An enormous waist for offices that only exacerbate hateful atmosphere on campus. Why? Here is an article to read: []

2.27.2023 2:24am - Goodness highness! Nature has discovered causality We should soon expect a flood of machine learning folks asking: What is it? BTW, I didn't tell them, but the book I'm holding in the picture is the first printing (1763) of Bayes paper on Bayes Rule. []

2.27.2023 12:11am - And I can't resist the temptation of endorsing your temptation. These are good papers. [reposted:]

2.26.2023 6:43pm - Wow, now they (CUNY Administration) are attacking those who request to find out what their "highly qualified chief diversity officers" thinks about Zionist professors at CUNY. This is called intellectual humility. [reposted:]

2.26.2023 6:16pm - Condemning the inexcusable! [reposted:]

2.26.2023 6:09pm - It blows one's mind to think of the resources invested in economics studies which, by and large, are observational or quasi-experimental and which, like the Hip Fracture study of @soboleffspaces etal, are waiting for modern re-analyses of their precious data. #econTwitter [reposted:]

2.26.2023 3:00pm - From one happy warrior to another: Stay the course, Asra! Actually, since there is only one "course", I am just repeating self-evident facts. Warriors need those too. [reposted:]

2.26.2023 2:45pm - How radical Islamists, @CAIRNaional and a coalition of extreme leftists are silencing/shaming Muslim reformers who do not toe their line. [reposted:]

2.26.2023 1:49pm - To remind readers, @CAIRNAtional was the first to invent the theory that Israel was behind 9/11. A theory that metastasized in yesterday's "Hate the Jews DAY" and "Zionists out of campus" in CUNY. This disgusting and deceitful organization is invited to oversee "inclusion". [reposted:]

2.26.2023 8:27am - I don't understand why everyone is jumping on Palestinians for celebrating murders by handing out sweets. Multi-culturalism teaches us to accept cultures for what they are, right or wrong, humane or barbaric. Israel was not given a choice to select neighbors. [reposted:]

2.26.2023 6:38am - @mehdirhasan is the only commentator I know who does not ask himself which countries declare their neighbors illegitimate. [reposted:]

2.26.2023 5:35am - This post by @causalif need to be studied if we want to understand the torture that economic students must go through who, oppressed by the educational tyranny of Angrist, Imbens and Rubin, dare ask themselves: "Which variables do I need to control for?" [reposted:]

2.26.2023 5:32am - Two Israelis murdered in Palestinian terror shooting in northern West Bank [posted:]

2.26.2023 5:22am - (Replying to @lilienfeld1) This isn't a debate. To understand their place, we need to know their capabilities and limitations. The capability of OS to supplement ATE with P(benefit) and P(harm) is new and extremely useful. I haven't seen it done before in any large study.

2.26.2023 2:28am - (Replying to @artistexyz and @CausaliF) I wouldn't be so harsh, but historians will surely ponder what made them resist tools that answer questions they deem important, when no alternatives were in sight.

2.25.2023 8:44pm - Extremely interesting record of the antisemitic character of Jimmy Carter, the first to label Israel "apartheid" and the first to give terror an ideological legitimacy, see, The only US President to be remembered as a Lover of Hamas. [reposted:]

2.25.2023 7:57pm - From the author of "The Jewish Lobby". They start as benevolent critics of Zionism and end enemies of all liberation movements. [reposted:]

2.25.2023 6:38pm - (Replying to @maipenraiPhD and @soboleffspaces) Why? I thought economists are proud of their studies. #EconTwitter

2.25.2023 6:27pm - (Replying to @DanaBashCNN and @jaketapper) So, how about a follow-up to your hour, and include the many things we have learned since? e.g. [reposted:]

2.25.2023 5:17pm - Retweeting complete reference for the Hip Fracture paper of @soboleffspaces et al in the Canadian Medical Association Journal, 2018. [reposted:]

2.25.2023 2:35pm - This one advantage observational studies have over RCTs; properly designed and properly adjusted, the former provide us with both ATE and P(x,y,z,...) for the SAME target population, free from RCT imperfections. This permits us to estimate P(harm) and P(benefit) for individuals. [reposted:]

2.25.2023 2:12pm - (Replying to @ancelkeyslegacy) Of course we take it into account. When an RCT study is available, it is honored for what it can deliver, and properly combined with other available data sources, see "Data-Fusion"; it's all pure logic.

2.25.2023 1:00pm - I bet there are many carefully conducted observational studies where, lacking causal models, researchers were unable to decide what factors need adjustments in order to obtain an unbiased estimates of causal effects. Time for modern re-analyses of precious data. [reposted:]

2.25.2023 12:44pm - This is an opportunity for the CUNY community to expose Saly Abd Alla ,disqualifications by focusing on her Zionophobic racism, rather than falling into the antisemitic rabbit hole. Focus, focus, focus. [reposted:]

2.25.2023 4:23am - America has changed, Mr. Ambassador. How can it be that Berkeley Law School would proclaim Zionist speakers unwanted, and CUNY's President would designate a BDS activist to head "diversity and inclusion"? America has changed. [reposted:]

2.25.2023 3:59am - Sparks of light can be seen even on this "Day of Hate". The city of Frankfurt just cancelled the Grandmaster of Hate, whom we all hoped would retire gracefully to eternal silence in history's Hall of Shame. [reposted:]

2.25.2023 3:46am - Another nice feature of the @soboleffspaces - Guy analysis is the estimation of P(benefit) bounds in addition to ATE. This can be done in any Observational Study, since we have both observational and (simulated) experimental data, without the imperfections of physical RCTs. [reposted:]

2.25.2023 3:21am - (Replying to @f2harrell @OstlundOllie and 4 others) A good example of this benefit is demonstrated by the work of @soboleffspaces , Guy, etal, on Hip Fracture Surgeries CMAJ 2018 August 7;190:E923-32. doi: 10.1503/cmaj.171512 The analysis was done 20 years after the study, using graphical models, which did not exist 20-30 yrs ago.

2.25.2023 3:06am - (Replying to @Raminderpal1469) Sure, send it to my email, but try to explain things in CI language.

2.25.2023 3:01am - (Replying to @ancelkeyslegacy and @MichelSpeiser) Yes, deductive logic in the sense of having semantics, syntactics, rules of inference etc. except it is not classical logic (which does not capture causation). See its axiomatic characterization here: Section 7.3, page 228.

2.24.2023 11:52pm - I don't care if they are protesting the existence of bagel or the color of lux , my response would be the same: raise the Israeli flag on your front lawn, together with the American, haters cant stand the strong and the proud. [reposted:]

2.24.2023 3:18pm - Qualification: In saying "What statisticians still do not get is the notion if 'inevitability'," I meant "Some statisticians still do not get...". I'm sure my co-author Nick Jewell, and even some Rubin's disciples (and many more) understand what "necessary assumptions" mean. [reposted:]

2.24.2023 3:08pm - (Replying to @MichelSpeiser) Good question. "Inevitable" is equivalent to "logically necessary" and cannot be shown to hold unless we have a logic to go by. CI provides a logic for cause-effect relationships, which traditional statistics lacks. The mantra "correlation is not causation" is a theorem in CI.

2.24.2023 11:09am - (1/2) Statisticians fear of "assumptions that are not verifiable" explains a century of paranoia and delinquent neglect of all questions related to causal relationships. CI is all about expressing inevitable assumptions and taking them to their inevitable conclusions. What statist [reposted:]
2.24.2023 11:09am - (2/2) What statistician still do not get is the notion of "inevitability", and the distinction between inevitable vs. convenient assumptions. CI researchers criticize assumptions only when they are unnecessary or deliberately obscured.

2.24.2023 7:09am - (Replying to @HL327 @stephensenn and 2 others) Yes, they correspond to the same 4 response types: Benefiting, AlwaysSurviving, Doomed & Harmed. The reason for the confusion is that we try to adhere to the vocabulary of the non-compliance literature. We shouldn't have.

2.24.2023 3:51am - (Replying to @IntuitMachine and @JohnPelhamBlack) RCTs are an example of interventions without a prior model, and they are not useless, merely limited in what they can deliver at great cost.)

2.24.2023 12:45am - (Replying to @HL327 @stephensenn and 2 others) The same exercise applies to the bounds in our paper. One may question of course whether a linear utility function is adequate in one case or another, but the bounds produced and the optimization strategies are the same.

2.24.2023 12:41am - (Replying to @soboleffspaces) The EBM call for integrating "patient values" was meant as a request of the medical profession to be left alone and allow hand-waving when RCTs fall silent. Little did they know that in 2 decades the integration of "patient values" would be formalized.

2.23.2023 10:59pm - (Replying to @AsraNomani @DrJavadTHashmi and 11 others) Someone told me @DrJavadTHashmi champions diversity in Islam; I see that Hashmi's diversity is everything that does not deviate from hateful Islam. Glad he is quitting Medicine - a dangerous profession for Zionophobes.

2.23.2023 1:30am - (Replying to @DigThatData) No, it was this: [reposted:]

2.23.2023 12:28am - I hope this link works: “Let Me Just Interrupt You”: Estimating Gender Effects in Supreme Court Oral Arguments - #ScholarAlerts [reposted:]

2.23.2023 12:08am - Who could ever imagine that causal mediation analysis would be used to unveil gender biases on the US Supreme Court. It did: file:///C:/Users/Judea/Downloads/text_interruptions%20(2)%20(2).pdf

2.22.2023 11:50pm - (Replying to @stephensenn) The latter is in progress.

2.22.2023 11:42pm - (Replying to @leonpalafox) The next book after #Bookofwhy would be Primer Individual chapters can be accessed on the web page. And, if you seek an even more formal text, with the full power and beauty of the causal revolution, see CAUSALITY

2.22.2023 8:25pm - RCTs are back in the news. This time on the pages of the NYT: The Mask Mandates Did Nothing. Will Any Lessons Be Learned? [posted:]
2.22.2023 8:25pm - It raises many questions, of course, political, as well as cultural and philosophical but, from my little corner, one question looms natural: How many authors of the Cochrane Report read #Bookofwhy? Short of implying that that number could make a difference, I'm still curious.

2.22.2023 6:42pm - (Replying to @HL327 @stephensenn and @f2harrell) The paper "unit selection" (with Ang Li) is an exercise in Decision Analysis, showing how a utility function can be maximized once we know prob(response type i) i=1,2,3,4. See

2.22.2023 6:29pm - (Replying to @GermsAndNumbers and @statsepi) What's the problem? I got a message from a senior Yale professor of epidemiology, describing to me the history of EBM (which I didn't know) and why he has objections to it, which I found to be sound. I learned new things from his message, and shared. What is wrong?

2.22.2023 4:09pm - (Replying to @HL327 @stephensenn and @f2harrell) Since the sum of the four proportions is ONE, only 3 need be estimated. Most importantly, we must leave details aside when we ask methodological questions such as "What is the goal of RCT?" . Perhaps this would be the main function of our philosopher.

2.22.2023 2:58pm - (1/2) (Replying to @HL327 @stephensenn and @f2harrell) I think you have defined the goal of RCT quite clearly: To estimate, given patient's characteristics, three proportions: "(1) those who will get better without treatment, (2) those who will get better only with treatment, and (3) those who will fail to get better..."
2.22.2023 3:06pm - (2/3) (Replying to @yudapearl @HL327 and 2 others) As you can see, we are converging on P(harm|c) and P(benefit|c), not CATE. The only question remaining: proportions from what population? Some claim such population does not exist because we are not sampling at random. Philosophy aside, we obviously need to focus
2.22.2023 3:14pm - (3/3) (Replying to @yudapearl @HL327 and 2 others) on the uncertainy we have about the next patient, not on the uncertainty we had about patients in the RCT. Great Convergence: The goal of RCT is to estimate P(harm|c) and P(benefit|c) of patients likely to enter the operating room. How we go about it is a different question.

2.22.2023 12:50pm - (Replying to @soboleffspaces) I haven't hears about EBM until few weeks ago, except (so I thought) as a self-promoting adjective. I didn't realize some consider it to be a methodology. Wikipedia surprised me.

2.22.2023 12:41pm - (Replying to @GermsAndNumbers and @statsepi) What is Brownstone known for? Is it a trump-supporting cult? A woke-or-else clan? Please educate me.

2.22.2023 9:28am - (Replying to @soboleffspaces) Yes, it's a good paper. Alex Balke was one of my best students, and this paper was part of his PhD dissertation. Too bad he did not go to academia.

2.22.2023 9:24am - (Replying to @HL327 and @stephensenn) Lots for me to absorb and digest in Chapter 13. I would hope though that the "purpose" of RCT will be articulated, not in Odds Ratios, but in terms of what it tells the doctor about the next patient in the operating room.

2.22.2023 8:48am - Congratulations go to nine new laurates of the Franklin Medal in Science: [posted:]

2.22.2023 3:53am - (Replying to @gaurav010x @EecsIisc and @IsraelMFA) Thanks for posting. I love this picture, it makes me look like a scholar.

2.22.2023 12:15am - (1/3) The stream of results in causal inference continues to flow week after week. This week's crop is listed here: One paper that caught my eyes was "Counterfactual Triviality in Structural Causal Models" by Wu. I was disappointed; why spend time on
2.22.2023 12:15am - (2/3) problems that plagued outdated theories of counterfactuals, if they do not surface in modern theories? Another interesting paper is @heckerma1, which resurrects memories from the early days of expert systems, and how Microsoft got into the Bayesian
2.22.2023 12:15am - (3/3) Networks enterprise. Heckerman's memoir should be supplemented with two of mine: and, which describe those developments from a slightly different perspective.

2.21.2023 11:36pm - (Replying to @pbleic @ChristosArgyrop and 6 others) Social media is a dangerous rope; you use a metaphor in some context and ZOOP, someone is sure to accuse you for the rest of your life for not understanding how things work in "real life". Come on. You really think I haven't met nurses?

2.21.2023 11:00pm - (Replying to @pbleic @ChristosArgyrop and 6 others) I would love to hear what other problems in RCT context you would have liked an intervention-calculus to solve for you. Chances are, it is someplace in the CI literature, like this one: If not, we probably tell you if a solution exists. [reposted:]

2.21.2023 10:52pm - Glad you mentioned "non-compliance". Indeed, this was one of the first tasks addressed and solved in do-calculus. See The paper also includes applications to real-life data -- pacifying those who claim such applications are missing from CI literature.

2.21.2023 9:48pm - (Replying to @ChristosArgyrop @f2harrell and 6 others) Well put: "Not all interventions are amenable to RCTs." I would only top it with "but all RCTs deploy interventions", and continue with: "Hence we must have a calculus of interventions, if we want to understand RCTs"

2.21.2023 8:34pm - (Replying to @HL327) Is a summary of principles available electronically for the Oxford Handbook of Clinical Diagnosis? I believe the 1st question a philosopher would like to ask is: "What is the purpose of RCT?" and "How do you interpret its finding?". We got stuck on these two last week.

2.21.2023 4:50pm - (Replying to @HL327) I do not have personal experience in medicine, yet I often spot a logical flaw when I listen to my doctor's line of reasoning. Philosophers can do better. I hope you are not claiming that doctors already know how to coherently put together all the evidence they have.

2.21.2023 3:41pm - (Replying to @stephensenn @ChristosArgyrop and 6 others) Mathematics is not MY TERMS or YOUR TERMS, it is a useful language to communicate ideas unambiguously, even across disciplines.

2.21.2023 3:33pm - (Replying to @trumanfrancis) The "change in estimate" is obviously wrong for selecting de-confounders (eg, recall the M bias) but may be good for ranking them, once they are selected with the help of a model.

2.21.2023 12:58pm - (Replying to @statsepi) It doesn't matter, indeed, when you are trying to learn how rank-and-file epidemiologists think.

2.21.2023 12:28pm - (Replying to @stephensenn @ChristosArgyrop and 6 others) All such nuances are to be represented mathematically, accounted for, or corrected for.

2.21.2023 12:22pm - (Replying to @KirkDBorne and @tut_ml) Thanks for posting. I've looked into 35&36 -- I cannot recommend them to readers.

2.21.2023 12:18pm - (Replying to @HL327) Only if they describe it in response to guiding questions by the philosophers. But would it be much different than what is described by the gurus here: [posted:]

2.21.2023 12:12pm - (Replying to @ChristosArgyrop @pbleic and 6 others) How can you refuse treatment if it tastes like placebo?

2.21.2023 12:11pm - (Replying to @pbleic @stephensenn and 6 others) I've stopped commenting on caricatures, must focus on serious issues.

2.21.2023 3:25am - (Replying to @HL327) I take it then that you are satisfied with the ability of medical practitioners to formulate the principles of their profession - they do not need help from logicians or philosophers - they are perfectly capable of avoiding circular reasoning like we have seen here on Twitter.

2.21.2023 3:21pm - Supreme Court Declines to Hear Israel Boycott Case in Major Win for BDS Opponents [posted:]
2.21.2023 3:21pm - Good news are quite scarce these days, making this victory a brightener for the rest of the day.

2.21.2023 6:38am - (Replying to @MenachemRozenk1) Are you trying to prolong the conflict?

2.21.2023 6:32am - The state of affairs is bad, but medical researchers are not at fault -- formal causal inference is only 30 yrs old.

2.21.2023 6:21am - As a practicing disbeliever, I am uplifted by the awesome hymn that these stones sing. [reposted:]

2.21.2023 5:59am - The Norwegian model of a "conference" mirrors the Palestinian model of a two-state solution -- a perfect choreography! [reposted:]

2.21.2023 5:40am - (Replying to @Claire_V0ltaire and @TankoZionism) News to me. Pakistan's former President, Musharraf, told the world that the Muslim expansion of the 7th Century was done, not by force, but by exemplary behavior.

2.21.2023 5:34am - To Barcelona, with love! The Israeli anthem always sounds relevant when sung in Barcelona. [reposted:]

2.21.2023 5:13am - What's the point? A good pollster can get any result he/she wants by slight re-phrasing of the questions. Only by asking: "Is Tel Aviv a stolen land?" can one get some idea of what Palestinians mean by "a peaceful two-state solution". [posted:]

2.21.2023 4:53am - (Replying to @tdietterich and @fchollet) One should also mention that graphical models were developed with precisely this problem in mind: Expert systems of the 1970-80s could not elicit enough specification rules from experts and, more importantly, could not combine them coherently once elicited.

2.21.2023 4:44am - "Not Like That" is what the protest is all about Not Like That! [reposted:]

2.21.2023 2:54am - Sorry for the typo, it should have been EBM = Evidence-Based Medicine (not EDM). It actually starts the discussion of EBM with: "There is perhaps no bigger plausibility sham today than “evidence-based medicine" [reposted:]

2.21.2023 2:42am - (Replying to @HAStark) Sorry for the typo, it should have been EBM - Evidence-based Medicine.

2.21.2023 1:26am - I am in receipt of an insightful article by Harvey Risch, which illuminates the history of the EDM culture, as well as claims for RCT superiority. It evokes new thoughts on our attempts to communicate with hard-core trialists. Worth reading. [reposted:]

2.20.2023 5:33pm - (Replying to @issiromem and @TimesofIsrael) Israel is undergoing a nervous breakdown; it needs the support of her lovers, not the condemnations of her enemies, especially not those enemies who have contributed to her nervousness by calling for her destruction in the past 75 years.

2.20.2023 5:15pm - (1/2) I checked "philosophy of statistics" on the Stanford Encyclopedia -- it's a 2014 article, still wound up in p-values, confirmation theory and the frequentists-Bayesians debates. We need a philosopher of experimental science to pay attention to the RCT riddle, because so much
2.20.2023 5:15pm - (2/2) of our public decision making depends on it, and so much ignorance exists about its aims and its outputs.

2.20.2023 4:12pm - (Replying to @soboleffspaces) Do they include ONE philosopher of science in their debates? Would they respect the guidance of such a philosopher or perhaps a logician who is trained in spotting circularities.??

2.20.2023 2:56pm UN Security Council pans Israeli settlements; Israel: US should not have signed on via @timesofisrael
2.20.2023 2:56pm No. Even the US is missing what the UNSC resolution is missing. The sentence: “the obligation of the Palestinian Authority to renounce and confront terror" should have continuee with: "and educate the public towards accepting a two-state solution."

2.20.2023 2:39pm - (Replying to @analisereal) Only if we can find a dedicated guest editor to prevent the issue from sliding into another miscommunication, by insisting that contributors answer the questions asked, and not regurgitate what they have been saying all along. Volunteering?

2.20.2023 2:30pm - (Replying to @soboleffspaces) I am not sure they are aware of the circularity looming in the way they are describing their goals. And I don't know how to alert them to it, without offending them. We need someone whom they do not regard as an adversary.

2.20.2023 2:22pm - The more we debate the PUPOSE of RCTs, the more convinced I become that the Society of Clinical Trials should organize a conference on: "The logic and philosophy of Clinical Trials."

2.20.2023 1:35pm - I can't see how this clarifies the PURPOSE of RCTs, or the definition of "relative efficacy" as an objective quantity that exists outside the RCT setup, and which the designers of RCTs are attempting to approximate.

2.20.2023 12:39pm - Fixing broken link to: "Comments on Nick Huntington-Klein's review `Pearl before economists: The Book of Why and empirical economics'" Journal of Economic Methodology, 30(1):63--67. Here is the new link:

2.20.2023 12:14pm - (Replying to @f2harrell @Raamana_ and @stephensenn) The sentence "Purpose of RCT: estimate relative efficacy" needs clarification. If order for "relative efficacy" to be a PURPOSE of an estimation exercise, it must be defined somewhere (external to RCT). Otherwise it sounds like: "Purpose of RCT: estimate whatever it estimates".

2.20.2023 3:49am - (Replying to @HL327 @stephensenn and 3 others) What's wrong with it? It saves some people and kills others, the former and the latter are each 10% of the population and, obviously, non-intersecting. Do we really need to say it at such length? Let's get to the meet of things and ask: Can we devise a test for the two models?

2.20.2023 2:04am - (Replying to @HL327 @stephensenn and 3 others) If you are referring to the example in our paper, then my conclusion is somewhat different: The FDA should license the drug for all females and lounch a study to explore the existence of features E and F that produce benefit in some males and harm in others.

2.19.2023 11:56pm - (Replying to @stephensenn @HL327 and 3 others) None of the above. It's an innocent question for readers with intellectual curiosity: Devise a test to determine which model is behind our data. Model-1 – The drug has no effect whatsoever on any individual Model-2 – The drug saves 10% of the population and kills another 10%.

2.19.2023 11:31pm - (Replying to @sim0ngates @f2harrell and 4 others) It is a wider comment, suggesting that RCTs as well as all other available sources of data (eg, Observational Studies) be leveraged coherently to answer the research question at hand. By "coherently" I mean using the logic of causal inference.

2.19.2023 12:07pm - (Replying to @heckerma1 and @PSpirtes) I got a broken link, can you check?

2.18.2023 11:24pm - (Replying to @aydinonat @paolosilvest and @Catemarchionni) I still could not find direct access.

2.18.2023 11:20pm - (Replying to @voidmstr) I'll figure it out tomorrow. Sorry

2.18.2023 11:15pm - Happy to share the final version of my paper: "Comments on Nick Huntington–Klein's review ‘Pearl before economists: The Book of Why and empirical economics’" :

2.18.2023 8:57pm - (Replying to @soboleffspaces @HL327 and 4 others) Sure. But as soon as you put down ATE as a difference ATE = P(Y_1 = 1)-P(Y_0=1), some RCT folks will say: We don't care about ATE; we are using Odds Ratio not Risk Difference.

2.18.2023 4:57pm - (Replying to @HL327 and @smueller) No matter how sophisticated and clever the RCT, adding an Obs. Study may always add something about the individual harm, depending on the bounds produced.

2.18.2023 3:53pm - (Replying to @DavidHirsh) We can say and say, and we have been saying and saying; to no avail. Why? Because (1) "antisemitism" opens the doors for endless philosophical/legal arguments, (2) Claiming "antisemitism" is begging for legal protection, instead of asserting moral highroads and moral imperatives.

2.18.2023 2:26pm - (1/2) This campaign, to fire Zionists from their jobs, is a result of our failed strategy of naming the perpetrators "antisemites" and quit. It's time to move the fight from the hopeless legal arena to the moral arena, where we can win hands down. It's time to call for the firing [posting:]
2.18.2023 2:26pm - (2/2) of these perpetrators for spreading the stench of Zionophobia - a genocidal moral deformity that calls for the destruction of a people's homeland and for stripping eight million refugees of their history, safety, creativity and dignity. @noatishby @JGreenblattADL @EinatWilf

2.18.2023 1:19pm - (Replying to @DoctorNazarian) Are you trying to give antisemitism a bad name? They might sue you for liable.

2.18.2023 12:58pm - (Replying to @HL327) All 'harms", whether expressed counterfactually or in English, are due to "separate causal mechanisms". So, I see no reason to speak about "two types of 'harm'". The underlying mechanisms could be very complex, but always manifest themselves as P(y'_t, y_c) > 0

2.18.2023 5:04am - (Replying to @f2harrell @HL327 and 3 others) If no confounding exists, and we do not suspect any, we do not need to run the RCT; we can estimate ATE from OS .

2.17.2023 4:36pm Good, now that we are all exhausted, and everyone sees the importance of estimating P(individual harm), here is a revised version of our paper: to be published in the Journal of Causal Inference. Enjoy, and try to assess your P(individual benefit).

2.17.2023 4:03pm - (Replying to @HL327 @AngeloDalli and 3 others) Recall that, if we were not worried about such "hidden covariant that acts as an effect modifier" we wouldn't run RCT in the first place. If you believe that "Obs Study added info abt the RCT," please convince @stephensenn &your other buddies in their language -- I'm exhausted.

2.17.2023 3:37pm - (Replying to @HL327 @stephensenn and 3 others) I couldn't relate to your analysis, b/c it was too wedded to the idea of varying dose, which takes us away from the main point: P(harm) vs. ATE I love your concluding sentence: "Such observational studies are of course well established practice." Please send one to @stephensenn

2.17.2023 3:22pm - The suspect, Jaime Tran, has admitted to the shootings, adding that he searched Yelp for a kosher supermarket and was homeless for the past 12-14 months. [posted:]
2.17.2023 3:22pm - Gee, his mentality is closer to Kanye West than to Rashida Tlaib -- the pendulum is swinging back.

2.17.2023 2:52pm - (Replying to @HL327 @stephensenn and 3 others) Miscommunication again. 10% of 'IndivHarm' & 10% of 'IndivBenefit' means exactly what you said: 10% are saved by a drug curing cancer & it kills a different 10% by bone marrow destruction. It's expressed as a property of an individual b/c Ms Jones doesn't know which type she is.

2.17.2023 2:40pm - (Replying to @stephensenn @AngeloDalli and 4 others) There is NO ASSUMPTION whatsoever in the M&P paper about what those who refuse a treatment when given a choice would do under different circumstances. Absolutely nothing. Could this be the source of our miscommunication? ??????

2.17.2023 2:24pm - (Replying to @HL327 @stephensenn and 3 others) Everything hinges on understanding that, for both a patient and a policy maker, taking a sugar tablet is NOT the same as taking a drug that saves 10% and kills 10%. I believe even FDA will agree on the importance of the difference, and our paper convincingly explains why.

2.17.2023 1:22pm - (Replying to @f2harrell @HL327 and 3 others) Two reasons why I prefer Twitter: (1) It forces you to convey ideas succinctly and precisely, rather than bury yourself in details. (2) I have an obligation to 65k (silent) followers who are taking this educational channel seriously.

2.17.2023 7:52am - (Replying to @jrosell @stephensenn and 4 others) I don't see why. I can run very replicable study on rats, but if I would try to generalize the conclusions to humans, many would scold me for using "bad evidence".

2.17.2023 6:33am - (Replying to @f2harrell @stephensenn and 3 others) This is one of the options we can take: Totally discard epidemiology as invalid, not to mention economics, political science and more, every science that has not purged OS from its midst, and legitimize only RCTs, though they are excused from external validation by definition.

2.17.2023 6:19am - A culture seeded in deceit must remain deceitful to preserve its identity. [reposted:]

2.17.2023 4:30am - (1/2) (Replying to @HL327) The importance of estimating P(individual harm) & P(individual Benefit) can be seen from just thinking about the difference between (1) a sugar tablet and (2) a drug that cures 10% and kills 10%. No need for technical terms. Forgetting about ATE etc. RCT cannot distinguish
2.17.2023 5:01am - (2/2) (Replying to @yudapearl and @HL327) Model-1 from Model-2 even under Odds Ratios or any other comparison of P(Y_t) and P(Y_c). Our paper demonstrates that the pair [P(Y_t),P(Y_c)] can be the same in men and women, yet women face P(harm)=0 and men P(harm)=21%. Odds Ratios can't save RCT from its blindness.

2.17.2023 3:06am - (Replying to @stephensenn @soboleffspaces and 5 others) Do you accept the hundreds of Observational Studies in epidemiology as an example of " the sort of information M&P requires"? For details on how it's done see Causality pp. 299-301, using leukemia data.

2.17.2023 2:18am - (Replying to @yudapearl @stephensenn and 4 others) I have just given you hundreds real life, real data studies in which you can estimate P(ind.Benefit). You can either dismiss the whole field of epi as BS, because they are not using RCTs, or you should accept those studies as a legitimate real-life example. Your choice.

2.17.2023 2:07am - (Replying to @stephensenn @HL327 and 3 others) I do not have a list of studies on diseases and t reatments and populations. But take any OS in epidemiology. It contains OS data, and a CI model, so it contains OS data and (implied) RCT data. For any such combination we can find bounds on P(ind. benefit). What else is needed?
2.17.2023 2:18am - (Replying to @yudapearl @stephensenn and 4 others) I have just given you hundreds real life, real data studies in which you can estimate P(ind.Benefit). You can either dismiss the whole field of epi as BS, because they are not using RCTs, or you should accept those studies as a legitimate real-life example. Your choice.

2.17.2023 2:00am - (Replying to @stephensenn @HL327 and 3 others) It is impossible to search for evidence not knowing what you would consider to be an "evidence", when we can't even agree on what "evidence" there is that RCT works.

2.17.2023 1:14am - (Replying to @stephensenn @HL327 and 3 others) Another suggestion: Instead of seeking "problems" in the M&P paper, try to look for opportunities. My whole perspective changed when I tried. You can start with the opportunity to distinguish Model-1 from Model-2. For me it was a mind-opener!

2.17.2023 1:05am - (Replying to @stephensenn @HL327 and 3 others) Forget about C. Why should A be different than B? Finite sample effect? Different populations surveyed? Nothing is a problem if it is physically feasible. The only problematic data combinations are those identified as "infeasible" by the system.

2.17.2023 12:54am - (Replying to @HL327 @stephensenn and 3 others) Here is a challenge to all who march by testability, or imply that CI needs testing: Devise a test to determine which model is behind our data. Model-1 – The drug has no effect whatsoever on any individual Model-2 – The drug saves 10% of the population and kills another 10%.

2.17.2023 12:38am - (Replying to @HL327 @stephensenn and 3 others) To find out is something can be tested you need CI, which tells you the testable implications of your assumptions. I'm not clear what test you are proposing for P(ind. Ben). According to RCT experts, RCT can't even test ACE, do you expect it to test P(ind. Benefit)?

2.16.2023 11:43pm - (Replying to @yudapearl @stephensenn and 4 others) Notwithstanding the shelves upon shelves of exp. design papers written by FDA people, I'm unaware of anyone leveraging two new developments: (1) the mathematics of causal inference, its tools & insights, (2) Our ability to analyze questions like P(individual benefit) = ?

2.16.2023 11:23pm - (Replying to @stephensenn @f2harrell and 3 others) True! I always tell myself that things on the ground are harder than they look on paper. But I also tell myself that new promises must show up on paper first, before realized. I therefore explore new possibilities on paper, hoping for more than "this is not how RCTs are run"

2.16.2023 9:07pm - (1/2) Having seen how trialists exploit the additivity property of the Odd Ratio (OR) to escape selection bias, readers may be interested in seeing how the symmetry of OR is leveraged in non-parametric CI to remove selection bias. @eliasbareinboim has reminded me of an old
2.16.2023 9:07pm - (2/2) paper of ours that does just that. See Section 2 of:; Enjoy also the beautiful graphs that display the nature of OR and its rivals.

2.16.2023 7:56pm - Numerous critics have focused on the fact that the "art" of the “Gaza Ghetto” tattoos trivializes the Holocaust. [posted:]
2.16.2023 7:56pm - A must read for all Jewish faculty, students, parents, donors, and community leaders who still believe that Jewish Studies Centers in their universities are not molesting their values as Jews, or betraying their expectations as a community. My UCLA experience needs a story too.

2.16.2023 5:47pm - (Replying to @edelmann_domi and @f2harrell) Your post made me go back and read everything I've tweeted the past two weeks -- Sorry, couldn't find any **** around. I've found however some good ideas, that should be taken as seriously as I was when I wrote them. Please try.

2.16.2023 3:45pm - (Replying to @HL327 @stephensenn and 3 others) You seem to be familiar with both (1) causal inference (CI) and (2) Multiple parallel feedback control loops (MPFCL). If you judge (2) to do the work of (1), there is no need for you to use (1), just improve RCT to the extent permitted by (2). I am limited by what I see in (1).

2.16.2023 2:40pm - (Replying to @stephensenn @soboleffspaces and 5 others) I don't think you would want your students to quote your last Tweet-- it looks like written by someone who does not understand the difference between "explicating assumptions" and "assumptions must be true", or, between "correcting for effects" and "ignoring effects". Reflect.

2.16.2023 2:10pm - (Replying to @stephensenn @soboleffspaces and 5 others) All the king's horses, and all the King's men couldn't change the fact that "so far, the only method that conceptualizes, addresses, treats, defines, or estimates Pr(benefit | z) is PNS by @yudapearl & @smueller , as done in (Just accepted to JCI.)

2.16.2023 12:42pm - (1/n) I'll try, though I am not sure in what terms clinicians speak nowadays. Falsifiability means a warning signal, coming out of the study data, telling us that something went wrong in the study, namely, the assumptions we made about the people in the study, or
2.16.2023 12:42pm - (2/n) the procedures applied have been violated. Extreme example: subjects in the study all all homeless, whereas those awaiting clinical care (called "target population") are mostly rich lawyers. Now, to generate a such a warning signal, we must have data from both the study
2.16.2023 12:42pm - (3/n) and the target populations. And we must have some assumptions on how the two are related to each other, for, taking an extreme case again, if lawyers anatomy reacts totally differently to treatment, no disparity in data should surprise us. These considerations are expressed
2.16.2023 12:42pm - (4/n) mathematically in the literature called "causal inference" (CI), which derived both: (1) warning signals, and (2) tricks for circumventing violations that led to such signals. Another literature exists, called "trialists," for whom all this is "impractical" or worse; no data
2.16.2023 12:42pm - (n/n) is taken from the target population, nor assumptions are accepted about the relationships between target and study. All criteria for success or failure come from the RCT itself. Even the target population is sometime denied existence. Stop - they explain it better..

2.16.2023 1:46am - (Replying to @f2harrell @jon_mellon and 2 others) To see whether population ATE is computable from RCT requires mathematical representations of both ATE and RCT. "Transportability Folks" do have such representations (eg Q. In deciding "non computable", have you used their representations? Alternatives?

2.15.2023 11:53pm - (Replying to @stephensenn @f2harrell and 2 others) Can I understand this to mean: If you are interested in a representative sample, free of bias-producing interactions, you do not sample at random, but make sure your sample has 50% males and 50% females. Fine. But isn't random sampling a good protection from unknown factors?

2.15.2023 8:03pm - AI critics are warning us of future robots deployed in battle fields. They havn't warned us of organic robots deployed in our restaurants and bus stations. Watch them below: [posted:]

2.15.2023 7:53pm - (Replying to @tdietterich @aperianez and @erichorvitz) I don't mind calling the marginal "foundational" .But I mind calling them "models" not knowing what in the world they model, if any.

2.15.2023 7:24pm - (Replying to @f2harrell @jon_mellon and 2 others) Can I at least choose cancer cells at random? See [posted:]

2.15.2023 7:08pm - I once had a student like that but, as far as I know, he hasn't finished yet.

2.15.2023 7:04pm - Great laughs @SarahKSilverman [posted:]

2.15.2023 6:12pm - (Replying to @jon_mellon @f2harrell and 2 others) We have seen two kinds of "theoretically motivated transportability claims". (1) "No interaction" assumptions used by current RCT practitioners. (2) non-parametric assumptions used by transportability folks. The latter claim ATE for a target population. The former claim =???

2.15.2023 5:58pm - (Replying to @soboleffspaces @JohanDH2O and 9 others) If a statistician wants to call her survey decisions (eg who to send a questionnaire to) a "design", I'd honor her choice of words. But I can't honor the word "design" where there are NO decisions to be made, as in Lord's story. It's a bad imposition of RCT jargon on OS.

2.15.2023 5:45pm - (Replying to @f2harrell @jon_mellon and 2 others) The fact we can't choose any plot of land on earth does not mean that we can't sample from those I can choose, or that we can't estimate the ATE for exactly those we can choose. @stephensenn explains what "fine RCTs" are doing today, not what RCTs are capable of doing.

2.15.2023 4:53pm - (Replying to @f2harrell @jon_mellon and 2 others) Though "no RCT does that" on humans, with our current state of knowledge, @jon_mellon asked whether RCT can do it in principle, say on internet traffic, e-commerse, or say using our current understanding of non-parametric causal inference methods.

2.15.2023 4:42pm - ‘Not Just Israel’: Biden Admin Says Human Rights Candidate Axed for Range of Issues [posted:]
2.15.2023 4:42pm - Breaking News: The psychiatrists at the State Dept discovered what we new all along: That those who use the A-word on Israel are suffering from a much deeper moral deformity, one known as Zionophobia. Only @Harvard can risk exposing students to such deformities- @KenRoth @Ostrov_A

2.15.2023 3:28pm - (Replying to @stephensenn @analisereal and 7 others) 1) two halls do not make two schools. Lord spoke about "a university", namely ONE university. Do you think anyone might interpreted "two Halls" to mean "two schools". I don't get it.

2.15.2023 9:20am - (Replying to @stephensenn @analisereal and 7 others) AHA !! Eureka! What I call "story", you call "design". We are truly approaching a convergence. But if we are analyzing data already taken, with no option of change the way they were taken, can we safely say we have a "story" and not a "design"?

2.15.2023 9:15am - (Replying to @stephensenn @analisereal and 7 others) Lord's story had one center, one university and data already taken -- no choices left. Is Nelder approach needed for multi-center RCTs ???

2.15.2023 9:12am - (Replying to @stephensenn @analisereal and 7 others) Lord's story begins with data already taken.

2.15.2023 9:10am - (Replying to @stephensenn @ildiazm and 7 others) Perhaps the assumptions are different, but the aims were stated by Lord: "Determine which statistician is right? " I'm still anxious to see Nelder's assumptions stated without RCT metaphors; after all we do not have any intervention here.

2.15.2023 9:01am - (Replying to @ildiazm @PavlosMsaouel and 8 others) I see no disagreement, only a puzzle: If the data were already taken, and there is ONE university, is there anything we can learn from Nelder's approach?

2.15.2023 8:53am - (Replying to @stephensenn @ildiazm and 7 others0 The requirement of "unstated assumptions" is self-evident and should not even be mentioned, since EVERY question about "effects" must rest on causal assumptions that are not in the data. Thus, every solution must be in the form: Given causal assumptions, who is right? #1 or #2?

2.15.2023 8:28am - As a surviving/presenter at the 1st conference on Uncertainty in AI, 1985, I am happy to announce a Call for Papers for UAI-2023 For students of history, here is my paper at UAI #1, on constraints propagation in Bayesian Networks [posted:]

2.15.2023 2:24am - (Replying to @BenMaimon4 @AizenbergSalo and @TimesofIsrael) Thanks for posting this article on J. Street, it summarizes Ben Ami's lecture very vividly, as if I were in the audience. [posted:]

2.15.2023 1:35am - What I don't like about J. Street is their slogan: "Pro-Israel, Pro-Peace." Implying that being just Pro-Israel is not enough; left alone, Israel is pro-war. What a mis-reading of Israel, what a deformity of perception!

2.15.2023 1:20am - (Replying to @stephensenn @analisereal and 7 others) I hope I won't offend you if I make a confession: The word "design" throws me off to confusion whenever I see it in the context of Lord. Here we have two statisticians who make no design decisions whatsoever; they passively observe data and pronounce a verdict. What is "design"?

2.15.2023 1:08am - (Replying to @stephensenn @analisereal and 7 others) I still can't tell what words in BoW give you the impression that the story is different from the one in PGJ? The two look identical to me.

2.14.2023 11:01pm - This Tweet, and the photos of Asher (8) and Yaakov (6) (HY'D) jolted a reader to send me pictures of 12 Palestinian children killed in Gazza, and ask: How about? My response: He who counts bodies and discounts intentions emboldens the aggressor and indicts its victims. [reposted:]

2.14.2023 6:59pm - (Replying to @Vivify705) We tried those other clues too, they are called "2-state solution" and "equally indigenous". Unfortunately, when you dont believe in logic, you don't believe in anything but "Me, Me, Me".

2.14.2023 6:21pm - In formal logic this is called "reasoning by cases". There is only one problem with it, which surfaces when your audience doesn't believe in logic.

2.14.2023 6:03pm - US Withdraws Nomination of Activist Who Accused Politician of Being ‘Purchased’ by Pro-Israel Groups [posted:]
2.14.2023 6:03pm - Today I was yearning for some good news, and it came. I only question how he got nominated with a Zionophobic record like his. Sadly, he taught at UCLA and thus stained my university with his apartheid bigotry. Harvard can afford such embarrassments, not UCLA.

2.14.2023 5:14pm - (Replying to @analisereal @ildiazm and 7 others) Moreover, if we assume that we have only ONE Hall, serving two diets, everything simplifies, no more diversionary issues, and the paradox is still paradoxical. What a relief. It took us 100 tweets to filter out orthogonal diversions; glad we're back to the core of Lord's dilemma

2.14.2023 3:21pm - (Replying to @stephensenn @ildiazm and 7 others) 1)If one hall is further from the nearest bus stop, then obese students would prefer the closer one, so W1-->Hall. 2) I am happy with initial weight “causing” diet, because obese students prefer low-calories diets. 3) The whole "group structure is orthogonal to the main PARADOX.

2.14.2023 2:40pm - (Replying to @stephensenn @ildiazm and 7 others) I can't recognize Lord's story in any of the "group structured" DAGs. The one on the right has "school" as a variables, but we have only one school. The one on the left shows an arrow from "Hall" to "W1", saying that the Hall affects students' initial weight -- hard to believe.

2.14.2023 1:06pm - (Replying to @HL327 @f2harrell and 8 others) Sure. We are on a common ground even w/o the term "Decision Analysis". But you would need to explicate your understanding of P(harm) using the patient's utility function, as it was done here:

2.14.2023 1:00pm - (Replying to @f2harrell @HL327 and 8 others) The A in CATE stands for "average", honoring the "Expectation" operator in E(Y(1,u)-Y(0,u)). Note that it does NOT equal E(Y|treat, C)-Y|cont,C), since the latter is a statistical quantity; it describes behavior in an OS, not in RCT. If you wish to compare, U must distinguish.

2.14.2023 12:46pm - (Replying to @HL327 @f2harrell and 8 others) In our paper, we gave two examples showing that two persons, having two different utilities (or beliefs), would prefer different treatments, depending on their P(IH)'s, even though their ATE's are the same. page 3, paragraph 3.

2.14.2023 12:32pm - (Replying to @HL327 @f2harrell and 8 others) Though we are from different disciplines, we both use counterfactuals when it comes to explicating concepts and assumptions. So, how about expressing formally the way you see "benefit and harm"; we would be able to compare it then to ours and reach consensus.

2.14.2023 11:35am - (Replying to @f2harrell @StableMarkets and 8 others) Two comments: (1) The estimand you are describing is known as CATE, where C stands for "Conditioned on patient characteristics". (2) CATE hides information about P(benefit) and P(harm) which are "pertinent for individual patient decision making", see

2.14.2023 7:21am - (1/2) Every time someone introduces a new twist into Lord's paradox, I go back to my first paper on the topic, and I get a fresh air of clarity, coupled with an urge to share it with readers. Here it is: Enjoy the beauty and simplicity of the solution.
2.14.2023 7:21am - (2/2) What is often missed by some who attempt to analyze Lord's dilemma is that the clash between the two statisticians remains paradoxical even when we make strong simplifying assumptions. Thus, contesting the validity of the assumptions does not tarnish the solution. Enjoy.

2.14.2023 6:19am - (Replying to @stephensenn @ildiazm and 7 others) And my claim, having looked at all the arguments, is: (1) You do not need DAGs to present arguments, but you need some mathematical object to explicate assumptions. (2) The #Bookofwhy's resolution of Lord's Paradox is complete and beautiful, enjoy:

2.14.2023 6:19am - (Replying to @stephensenn @ildiazm and 7 others) And my claim, having looked at all the arguments, is: (1) You do not need DAGs to present arguments, but you need some mathematical object to explicate assumptions. (2) The #Bookofwhy's resolution of Lord's Paradox is complete and beautiful, enjoy:

2.14.2023 1:07am - (Replying to @MichiganWater2 @stephensenn and @f2harrell) Your interesting introduction of "the campus cookie shop," changes the story a bit, and is discussed in Fig. 2 of . But do you consider the paradox resolved in its original version, assuming WI is the "sole confounder" between Diet and Outcome?

2.14.2023 12:38am - Is there no responsible adult to tell President Abbas that such statements make a Palestinian state 10 times less realizable than 10 new Israeli settlements? Israelis have been hearing the "We Alone" slogan since 1929, and they know where it leads: [reposted:]

2.13.2023 9:52pm - (Replying to @MichiganWater2 @stephensenn and @f2harrell) I do not understand the difficulty. I said: there is no need to confuse (confound) schools with diets or with halls, just think of diets, nothing else. See a clear discussion: If you look at the graph, you see "confounding" created by the initial weight.

2.13.2023 6:27pm - (Replying to @soboleffspaces) "Due to" is normally associated with "necessary causation" and "but for" criteria. These are captured by PN, as opposed to PNS which captures "benefit".

2.13.2023 6:22pm - Our weekly report on causal inference articles is just in: Many titles caught my attention, but my curiosity landed on: Counterfactual Identifiability of Bijective Causal Models I need help understanding their ontology.

2.13.2023 5:20pm - (Replying to @stephensenn @ildiazm and 7 others) My professor did, but now I see that he shouldn't have said the words "independently distributed" because both presume the existence of some distribution, so uttering them in stat class would amount to worshipping the infinity devil - a heresy of the first degree.

2.13.2023 4:30pm - (Replying to @ildiazm @stephensenn and 7 others) I, too, would be content with any mathematical expression. But only after understanding what they mean I understood why they don't express it mathematically.

2.13.2023 4:25pm - Congratulations! Delilah!!! We are so proud of you, your teacher @adrianachavira , and the entire staff of the @ThePearlPast and the Daniel Pearl Magnet High School @PearlMagnetHS for this fantastic achievement. Danny would have said the same, I'm sure.

2.13.2023 4:13pm - (Replying to @stephensenn @ildiazm and 7 others) Just checked my copy (1st Edition). You are right. It says: Jacobi Bernoulli, Ars Conjectandi, Opus Posthumum, Baselee, 1713. Back to infinity land.

2.13.2023 2:15pm - (Replying to @stephensenn @ildiazm and 7 others) What comes first? Variances or infinities? My stat 101 professor defined "variance" as a a distributional property (Expected deviation from the mean), thus imagining infinity before understanding variance. Was he a bad statistician?

2.13.2023 1:46pm - (Replying to @stephensenn @ildiazm and 7 others) Causal Inference adopts the same answer statistics has, as formulated by Bernoulli (1710) when the population of Basel was finite and he assumed a "distribution" that governs the behavior of its residents. Bayesian belief is another way of circumventing the "infinity nirvana" .

2.13.2023 1:13pm - (Replying to @ildiazm @stephensenn and 7 others) Mathematical expressions presuppose the existence of a distribution that governs your samples. But if the existence of such distribution is rejected as "the nirvana of the infinite", you can't have a mathematical expression.

2.13.2023 11:49am - (Replying to @ildiazm @f2harrell and 7 others) "No interaction" is an assumption only if transportability is an objectively defined and desirable feature of your method. Not if your ultimate aim is merely to estimate "effects" on the subjects in the trial.

2.13.2023 4:03am - When I tell my American friends that the entire Palestinian narrative can be summed up in just two words, "We Alone", they refuse to believe me. Surely, they say, some Palestinians have heard about their neighbors having had some history...etc. No, they haven't -Nada. "We alone" [reposted:]

2.12.2023 11:26pm - Responding to @KenRoth renewed attack on Israel, I am changing the title to read: "Three Slanderers and two victims" No cause-effect implied, just moral deformity.…

2.12.2023 3:03pm - (Replying to @analisereal @f2harrell and 2 others) Almost. Except that P is undefined, hence, must be replaced by Freq. Also, the assumption OR(X) = OR(X, S=1) holds (after scaling) in linear models, which explains why selection bias is solved there

2.12.2023 2:52pm - (Replying to @RajaChemayel @Ostrov_A and 14 others) God forbid! That would make you anti-Palestinian, whose ideological DNA is to grab and possess their neighbors' country.

2.12.2023 2:45pm - (Replying to @yudapearl @analisereal and 3 others) Someone asked "the population is finite". I answered: Show me any statistical estimation exercise where the population is not finite. The population of Basel (Bernoulli, Artis Conjectandi, 1710) was also finite, yet by assuming it to be infinite, Bernoulli created statistics.

2.12.2023 12:51pm - Must read. [reposted:]

2.12.2023 12:45pm - (Replying to @leonpalafox) Let me know if you find out.

2.12.2023 12:36pm - (Replying to @analisereal @f2harrell and 2 others) I am not sure they would compromise to accept the average effect on the selected, ATE(S)= E[Y(1)-Y(0)|X, S=1] as a goal of RCT. Why? Because ATE(S) invokes expectation relative to some distribution, the existence of which they consider "fictitious".

2.12.2023 7:42am - My title for this photo: "Slander and its victims". [reposted:]

2.12.2023 7:33am - (Replying to @Raamana_ and @f2harrell) God forbid! I am arguing against the current practice of some RCT experts, who refuse to admit that RCT has a purpose: To estimate the ATE in the clinical population.

2.12.2023 4:15am - (Replying to @HL327 @f2harrell and 9 others) I dont see any contradiction. The OS alone makes me infer NOTHING. I turned to a RCT to estimate something (ATE) and still found no contradiction, b/c OS does not compete with RCT on ATE, it merely adds to it information about individual behavior, which ATE lacks.

2.12.2023 4:04am - (Replying to @LGHemkens @artistexyz and 2 others) Ideal RCT's do not need DAGs. DAGs are needed to mend RCT's imperfections.

2.12.2023 4:00am - (Replying to @f2harrell @technoslerphile and 4 others) Extending follow-up amounts to adding an outcome variable, dependent on both treatment and RCT outcome. Don't we need modeling assumptions to say anything of added value?

2.12.2023 3:53am - Dedicated to my friends and colleagues in Barcelona and Madrid. Thanks for posting. [reposted:]

2.12.2023 3:47am - (Replying to @f2harrell @technoslerphile and 4 others) Y(1,u) differs from Y(1,X) in that u contains ALL characteristics of an individual u, measured and unmeasured, known and unknown, sufficient to make the response Y a deterministic function of treatment. I don't think avoiding counterfactuals is an option in matching RCT with ATE.

2.12.2023 2:05am - (Replying to @gottfriedmath @LGHemkens and 3 others) You are writing "[RCT] informs on ... average causal effect (ACE)". This is what we learned in school, and have hoped they are doing. Not so. What I've discovered the past 2 weeks is that the notion of ACE is non existent in the hard-core RCT literature; it's mocked as a fable.

2.12.2023 1:30am - @f2harrell 's paper, explains why RCT are not falsifiable, "Randomized clinical trials are successful because they do not mimic clinical practice", namely, because practitioners refuse to bow to any external criterion of performance - it must be internal.

2.12.2023 1:20am - Asher (8) and Yaakov (6), you don't know me, but I know you, albeit from a distance, and I am retweeting your pictures to decent people all over the world, just to remind God of His promise: evil will extinguish itself from this planet. שׁוֹבֵר אוֹיְבִים וּמַכְנִֽיעַ מִינִים: [reposted:]

2.12.2023 12:56am - RCT studies are not falsifiable because RCT practitioners reject the idea that RCT has a goal, named ATE, residing outside the RCT protocol. Once they accept ATE=E[Y(1,u)-Y(0,u)], RCT results turn falsifiable, see Eq.(9.32-9.33) in Sadly they reject it. [repost of]

2.12.2023 12:31am - (Replying to @f2harrell @stephensenn and 9 others) IMHO you are mistake if you think I am. If you think @stephensenn 's challenges are well-written, please translate them into reasoned, self contained ideas, not via links nor via judgmental verdicts such as "fictitious", "unrealistic", "untestable"...

2.11.2023 11:56pm - (Replying to @technoslerphile @f2harrell and 4 others) Your exchange with @f2harrell represents a clash between two healthy intuitions: (1) OS should mend some of RCT's imperfections (2) OS itself is imperfect. Clash! Only CI principles can reconcile these two. Conclusion: learn to speak CI if you wish a reconciliation before 2030.

2.11.2023 4:26pm - (Replying to @HL327 @f2harrell and 9 others) Sure. A 2nd RCT with narrower (and measurable) subpopulations would be more informative. But the reason for the alarming OS could be unmeasured, (e.g., tr. choosers were ladies who had a cousin that died from corona), we just go with what we learn from OS combined with 1st RCT.

2.11.2023 4:14pm - (Replying to @stephensenn and @soboleffspaces) Our intention was not to confound school with diet or with hall. There is one school, two dining halls, each hall has one diet plan. No confounding. If you think the text was not clear, pleas tell us what wording may have caused ambiguity.

2.11.2023 4:03pm - (Replying to @HL327 @f2harrell and 9 others) @HL327 , You are ahead of me. Sure, upon replicating our original RCT with ATE = 0.279, we discard the OS and decide to treat. Except, in 2023 we can do even better: combine the two (barring interruptions) to estimate P(harm), and see if we shouldn't flip our decision again.

2.11.2023 3:49pm - (Replying to @yudapearl @HL327 and 10 others) Moreover, now that I have two studies, RCT and OS, I can do better and (barring interruptions from RCT zealots) combine the two to estimate P(harm), and, upon finding P(harm)=0, decide to treat with a smile, as opposed to "treat with hesitation."

2.11.2023 3:19pm - (Replying to @HL327 @f2harrell and 9 others) Seeing a 73% death rate in the OS would make me wonder: Is it possible that women who choose trt. are mostly badly ill , hence desperate patients? I would then conduct a pilot RCT and, upon seeing ATE > 0, I would definitely conclude the OS was terribly biased, and recommend trt.

2.11.2023 3:06pm - (Replying to @stephensenn and @soboleffspaces) Where do you see a "parallel study" ?

2.11.2023 2:56pm - (Replying to @HL327 @f2harrell and 9 others) Based on our RCT, ATE = 0.279, so I would trt. (hesitatingly, worrying of harm). Based on our OS alone, I won't make any decision, it's all biased. But for females, based on both RCT and OS, we find p(Harm)=0 & p(Ben) = 0.279. So I would definitely treat, unhesitatingly.

2.11.2023 2:46pm - (Replying to @stephensenn @soboleffspaces and @yudapearl) Furthermore, there is an interesting difference between the experiment in Pearl, Glymour and Jewel P65 and Book of Why. PGJ describes a parallel study and BoW doesn't. Nelder applied to PGJ gives the solution in BoW but Nelder applied to BoW doesn't.

2.11.2023 2:40pm - (Replying to @HL327 @stephensenn and 9 others) I am about to explain (next tweet) how the results would change our decisions, if we only get a chance to discuss it dispassionately.

2.11.2023 2:24pm - (Replying to @stephensenn @HL327 and 9 others) When I get a question from a reader who shows genuine interest in understanding what our paper says, I would like to be given a chance to explain it dispassionately, please. Your difficulties: "fictitious", "unrealistic", "untestable", "wrong" "dangerous".. are well known, relax.

2.11.2023 12:54pm - (Replying to @HL327 @f2harrell and 9 others) In our paper on personalized decision making, we discuss two examples where decisions based on p(B) & p(H) would & should be different than one based on their difference ATE=p(B)-p(H). It's not surprising, since the former carries more information.

2.11.2023 12:33pm - (Replying to @Alfons_Valencia) I hope no one took my disappointment as "blaming", I meant only the former.

2.11.2023 8:18am - (Replying to @f2harrell @stephensenn and 8 others) A "method" is defined as a procedure of handling actual data, so it seems impossible that an author would simultaneously (1) "propose a method" and (2) state that it's not a method. I must be missing something.

2.11.2023 2:42am - For hard-core trialists, the entire data collection effort during the Covid-19 pandemic was a wasteful exercise, laden in biases and untestable assumptions. Same goes for those who say: "OS gives us some idea of where to run next experiment" -- this "idea" too is biased.

2.11.2023 1:51am - Conversely, every time someone says he/she is "not impressed by CI" or is "waiting for real-world worked-out examples," there is someone who admits to not having any principles for using OS and not caring to acquire any.

2.11.2023 1:39am - (Replying to @AleixGalvany) You made my day, thanks.

2.11.2023 1:38am - (Replying to @stephensenn @f2harrell and 8 others) The fact that proposals have been written does not mean that they answer the question asked, or that they are based on scientific principles, or that they can be communicated to other mortals. Numerical examples of our method are provided, each describing a "genuine" combination.

2.11.2023 1:23am - Every time observational studies are used as evidence for a claim or a policy, there is someone behind it who believes in CI, albeit intuitively. The aim of CI is to formulate this intuition, explicate its principles, and assist in its real-life implementation. [reposted:]

2.10.2023 11:53pm - (Replying to @f2harrell @stephensenn and 8 others) I wish I could join you in dismissing the "CI approach". Unfortunately, there is no alternative "approach" to guide practitioners how to make use of observational data, when available, or how to articulate a research question (estimand) when experimental data are available.

2.10.2023 11:39pm - Wrong, this is a Trump supporter posing as a Muslim. Beside, he is expressing legitimate criticism of government policies. [reposted:]

2.10.2023 11:25pm - CNN's self-driving car rams into a Post's settlement and hits occupied Al Jazeera. "The little dog laughed to see such sport, And the dish ran a way with the spoon" (Anon) [reposted:]

2.10.2023 4:49pm - (Replying to @AndrewPessin and @elderofziyon) Don't blame Palestinian journalists. Prostitution thrives when demand increases, and Western readers are dying for more and more of this sort of stories on cowboys and Indians.

2.10.2023 4:34pm - (Replying to @RaulMachadoG @leeflang_m and 3 others) Part of the problem is that most "business questions" are causal, and business schools teach stat, not causality, so practitioners can't articulate what they really want. Last week I discovered this to be a problem even among trialists, who should be doing cause-effect for living

2.10.2023 3:55pm - (Replying to @stephensenn @f2harrell and 8 others) Toy examples convey principles, and make all assumptions explicit and transparent. "Actual data" examples, bury principles and assumptions in noise, thus permitting researchers to argue endlessly about irrelevant details, and escape even the key question: What's our estimand?

2.10.2023 3:35pm - (Replying to @LGHemkens @soboleffspaces and @smueller) Not clear to me what assumpson you referring to.

2.10.2023 11:24am - (Replying to @f2harrell @elmir1omerovic and 8 others) That's exactly my point: You are not referring to the theoretical definition of ATE, nor to the difference between this ATE and whatever comes out of "actual data" (let's call it ATE**) . ATE simply doesn't exist, and life revolves around ATE** with no guidance to quality.

2.10.2023 2:58am - To my friends and colleagues in Barcelona: How could you allow this to happen? You have invited me to visit your city and universities. How can I accept when my heart tells me: "This city maligned the city of birth"? Or "this city believed vicious lies against my people." Hard.

2.10.2023 1:50am - (Replying to @MaimonAmir and @Ostrov_A) How about two Jewish academicians, at my university, writing to the UN urging it NOT to adopt. Which is harder to believe?

2.10.2023 12:58am - (1/2) For CI folks ATE is a quantity that exists in reality and which RCTs aim to estimate. For a trialist, ATE is whatever the trial delivers, assuming it is designed by good RCT experts. While CI folks might ask: what makes for a good expert? or whether a good expert is more
2.10.2023 12:58am - (2/2) likely to be closer to objective ATE, the question is meaningless to trialists, for whom ATE does not exist outside the expert. A good expert is one who shares practices and intuitions with other good experts, as in any advanced pre-scientific technology.

2.9.2023 10:25pm - (Replying to @returnstosender and @StudiesZionist) @returntosender . Not clear what you have against @StudiesZionist , an organization that I've joined recently. They decry comparisons between Auschwitz and Gazza - don't you? Doesn't every decent person? They need our support, not sarcasm. ??????????

2.9.2023 10:05pm - (1/3) The Wikipedia supports my story: In a randomized trial (i.e., an experimental study), the average treatment effect can be estimated from a sample using a comparison in mean outcomes for treated and untreated units. However, the ATE is generally understood as a causal
2.9.2023 10:05pm - (2/3) parameter (i.e., an estimate or property of a population) that a researcher desires to know, defined without reference to the study design or estimation procedure. Note also the counterfactual origin of ATE; it's based NOT on comparing two groups,
2.9.2023 10:05pm - (3/3) but on comparing the outcome of an individual u under treatment to the outcome of that SAME individual under control. The difference is more than philosophical; it makes trialists unable to communicate with CI folks, and possibly among themselves.

2.9.2023 9:24pm - (Replying to @DKedmey @AdamL and @jgreenblatt) If "leader" is one who follows his convictions, then Koch was a super-leader. We didn't include his essay in "I am Jewish" book because he refused to make the editorial changes I requested. Danny's last words are now inscribed on his headstone:

2.9.2023 7:33pm - (Replying to @f2harrell @elmir1omerovic and 8 others) This stands against everything I've learned from the literature. I believe it was D. Rubin who coined the term ATE, and defined it as: ATE = E(Y(1,u) -Y(0,u)) In other words, ATE is a population property, not a feature of any given study. Do Rubin's disciples agree with you?

2.9.2023 6:17pm - (Replying to @lilienfeld1) I wouldn't give Roger Waters the benefit of being an antisemite, a hereditary disease usually forgiven for reasons of sickly upbringing. Waters is worse; he is a pure Zionophobe - a racist with no excuse.

2.9.2023 12:48pm - Excited to read the whole paper. Who could imagine that the axiomatization of counterfactuals could be applied to images!

2.9.2023 8:13am - For all Hebrew speaking readers of our channel, I am supporting author Ilan Sheinfeld's appeal to resist the ongoing encroachment on the secular-democratic character of Israeli society.

2.9.2023 4:27am - In IRGC defense, I would check first if this is not part of "The Resistance".

2.9.2023 4:05am - []
2.9.2023 4:05am - Anyone tried this new platform from MIT on classical causal problems, e.g., The Firing Squad, or The Smallpox inoculation? #Bookofwhy. IOW, where in the scheme of CI could this platform be useful?

2.8.2023 8:46pm - (Replying to @Lester_Domes @AnjaLeist and 4 others) Sander: @_MiguelHernan 's recent "Like" jolted me to read this thread again and to realize -- Woo to me -- that I still have not seen ONE problem that TT addresses. Willing to learn, but I can only learn from problems, not books.

2.8.2023 6:22pm - Only a Zionophobic mind like that of Roger Waters could discover that "the Russian invasion of Ukraine was NOT unprovoked."

2.8.2023 8:22pm - (Replying to @elderofziyon) Only fake "Rabbis" can stay unmoved when babies are rescued from the rubbles.

2.8.2023 6:08pm - (Replying to @AleksanderMolak) Very interesting thread. Can we have links to your book and to "Statistical Rethinking".

2.8.2023 5:03pm - (Replying to @soboleffspaces @nickchk and @eliasbareinboim) "Lack of transparency"? Where? Who? Who said? Why? How? Holy Joe! Really?

2.8.2023 4:55pm - (Replying to @richard_landes and @Mivasair) The "Rabbi" is right. Zionists have never been under physical attack before.

2.8.2023 4:10am - (Replying to @nickchk and @Blogsbloke) I wrote somewhere that Nick's book may emancipate economics from the tyranny of outdatedness.

2.7.2023 11:17pm - (Replying to @ivanerogg and @soboleffspaces) Of course; SCM it made for it. To persuade your friend to do(X=x), you program your friend's biases into his SCM, you use your unbiased SCM to figure out what argument would maximize her perceived utility under do(X=x).

2.7.2023 9:52pm - I rarely miss an opportunity to RT a video clip by @EinatWilf , but this one is particularly on-target today, when a combination of forces are attempting to turn Israel into a Messianic (Halachaic) State. They will fail, because the anatomy of Zionism won't swallow it.

2.7.2023 9:34pm - I wish I was a believer: "Oh, God, don't make @daneasterman lie down there, in Gaziantep, trapped in the ruins, at -12 degrees Celsius, in the darkness, as IDF rescue workers are sweating to find live victims. He is the lowest of humans, true. But don't take him to Gaziantep."

2.7.2023 3:33pm - (Replying to @Patrikbentolila) The other areas are too numerous to mention, and are well known to people who have their eyes open. The trees, however, is an inspiring achievement that only few people know.

2.7.2023 12:32pm - (Replying to @elmir1omerovic @PatientStormDoc and 6 others) Your experience in Sweden is interesting, can you summarize in a few lines what information you get from the OS? If you can explain it in CI language, it would be great; if not, just the principle would do.

2.7.2023 11:49am - (Replying to @Anatomi54244944) It took a century for science to earn its reputation, at the cost of many good people's torture. Now that it has earned it, it is our obligation to prevent imposters from steeling this reputation and leading many good people towards wrong treatments or worse.

2.7.2023 5:13am - (Replying to @DiazQuijanoFA @MadelynTheRose and @NP_Jewell) Wow! Did we really say that? Where did we get the wisdom to predict precisely why we would be attacked later?

2.7.2023 3:02am - (1/2) It's time to take note of the CI papers published last week: I sense an increasing activity in Causal Discovery, but my attention goes to the paper on "Actual Causation and the Challenge of Purpose", perhaps because I am purpose-driven, and to
2.7.2023 3:02am - (2/2) the paper: "Epsilon-Identifiability of Causal Quantities", perhaps because I am a co-author (with Li and Mueller), or perhaps because it reminds me of Epsilon-Semantics -- an early work in non-monotonic logic which formalizes the notion of "almost sure". Enjoy

2.7.2023 12:27am - (Replying to @elmir1omerovic) A waste of precious time. Unless you enjoy hearing "not true" "wrong" "erroneous" repeated for the hundredth time, with no explanation.

2.6.2023 11:55pm - As we are leaving Tu Bishvat, our Tree-Planting Holiday, I am reminded that Israel is one of the few countries in the world that entered the 21th century with more trees than 100 years earlier. I even remember the Ficus tree I planted on TuBishvat, 1940.

2.6.2023 11:41pm - WATCH: Will Israeli attorney bring social media giants to their knees? | World Israel News
2.6.2023 11:41pm - A major court case will take place in two weeks, which might affect everything we know about social media. Moreover, Nitsana is a lioness, so I wouldn't bet on the outcome. @AttorneyNitsana @ShuratHaDin

2.6.2023 10:20pm - (Replying to @Anatomi54244944) Co-existence by all means, but shouldn't we distinguish the two modes of thinking, and not allow the emotional mode to pose as a scientific mode, as we find in some cults, like Scientology, Christian-Science or Happy-Science?

2.6.2023 9:52pm - (1/2) (Replying to @f2harrell) At the risk of spoiling the jubilee at the trialist community party, note what fault has been found in OS: There are pathological functions (eg non-differentiable) that could make the estimand derived "non-uniformly consistent". We do not know how prevalent this pathology is
2.6.2023 9:52pm - (2/2) (Replying to @yudapearl and @f2harrell) in the space of continuous functions, how severe it is and, more importantly, whether such pathologies also inflict attempts to remedy imperfections in RCTs, for example, concurrent control with its additivity assumptions. Cause for celebration? Ask your friendly CI doctor.

2.6.2023 6:33pm - REGISTER NOW! 2022-23 Daniel Pearl Memorial Lecture with Garry Kasparov March 6, 2023, 4: 30 pm, UCLA.

2.6.2023 5:59pm - The emphasis is on " in the manner implied ", which is buldozertized in style and scope.

2.6.2023 5:54pm - (Replying to @GoldsteinBrooke) Commenting "free Russia" on a Ukrainian person's post may not be antisemitic, but it sure is morally disgusting and genocidal.

2.6.2023 5:47pm - (Replying to @elderofziyon) C:urious: Is there a committee, a librarian, or a custodian on the Inalienable Rights of the Jewish People? Are there such rights?

2.6.2023 5:30pm - Israel sends search-and-rescue delegation to Turkey via @JNS_org
2.6.2023 5:30pm - The photos we see and the numbers are the worst that I can remember - worse than scary. 40 countries are pledging help, but they need immediate and experienced rescue teams. Bless you.

2.6.2023 5:03pm - (Replying to @splendorsolis69) No way. "Schmucks" is judgmental; I prefer "Schmegege", which is endearing, like "the village idiot".

2.6.2023 4:56pm - Extremely useful chart, to be posted in the hallways of every "Evidence Based" department. Luckily, students of Causal Inference are not suffering from this imprecision; once we explicate the assumptions behind the claim, the choice of word is rigorously justified.

2.6.2023 4:32pm - (Replying to @elmir1omerovic @RWJE_BA and 7 others) Thanks for defending the faith, although I do not know what was accused of - I've muted RWJE long time ago, out of social pressure to focus on the constructive. There are so many readers who wish to learn, that I can't waste time fending off detractors of learning.

2.6.2023 2:39pm - (Replying to @imthealan @ylecun and @VladicaV) Zelensky's canonical model may not be exactly canonical to Putin. My canonical model is SCM -- a model of reality, not a model of our algorithm.

2.6.2023 2:31pm - As soon as I posted this reply, I recalled my daughter warning me that the word "pagan" may offend the majority of living humanoids, and that I should be very very careful. So, seeking advice, what should be a respectful way of shunning those who muddy the unity of science?

2.6.2023 2:06pm - (Replying to @ylecun and @VladicaV) And my proposal calls for first studying what we can do with causal world models, before we labor to learn them. Still not mutually exclusive.

2.6.2023 2:00pm - (Replying to @Manuel_do_rio and @lexfridman) Merry sailing, and don't let the pagans muddy the unity of science.

2.6.2023 12:43pm - (Replying to @Manuel_do_rio and @lexfridman) I haven't changed that prediction. Part of it has already been accomplished through the formal distinction between "deliberate" and "imitating" actions.

2.6.2023 12:35pm - Here is an Open Letter to George Washington University that I've signed today: which I believe every concerned academic should sign, in light of the abusive practices perpetrated by some of GWU's teaching staff. Please consider adding your name.

2.6.2023 11:29am - Great thread for students of causality and cognitive science. I didn't realize there are psychological theories today that try to explain causal judgments without counterfactuals. Are there? What makes them resistant to the Ladder?

2.6.2023 12:30am - At the peak of his fame, Albert Einstein travelled to the holy land to lecture at the Hebrew University in Jerusalem, and to lend his name to an institute that would become a symbol of purpose and renewal to his people. One of many lectures I am sorry to have missed.

2.5.2023 9:36pm - It's one of the most perplexing problems facing physicists these days: Conservation of Matter. If Palestinian teenagers are all scholars, dreamers, artists, and lovers of mankind, where do their people-shooting teenagers come from? Physics is in an uproar. Not @Reuters .

2.5.2023 8:56pm - Published 3 years ago, but still a most powerful rebuttal of American Jews who seek redemption through suicide.

2.5.2023 7:21pm - (Replying to @nickeread) Nothing much changed the past 400 years, the poor are still more easily recruited.

2.5.2023 7:14pm - (Replying to @PatientStormDoc @stephensenn and 7 others) Not sure if you are referring to me, I know however that, when I reframe, there is benefit to the reframing. It is often done to exit from the clutter of details to the clarity of principles.

2.5.2023 2:24pm - (Replying to @AlfredoMorabia and @soboleffspaces) Thanks for this illuminating chapter of history. As we can see from the quote, Helmont deployed both randomization (cast lots) and intervention (i'll treat mine my way, you'll treat yours your way). Leaving no room for ambiguity.

2.5.2023 1:53pm - (Replying to @yudapearl @soboleffspaces and @stephensenn) Some would argue, of course, that you can't disentangle estimation from identification, because, reading stat textbooks, the two are badly conflated and entangled. Why? Lacking identification + estimand, you never know what needs to be estimated. These days are over.

2.5.2023 12:44pm - (Replying to @soboleffspaces and @stephensenn) When I was still hoping to have fruitful conversations with @stephensenn I used to say: "SCM lets smart statisticians address the SE issue". Today I dare say: "SCM has done more for the SE challenge than most statisticians. How? By disentangling estimation from identification"

2.5.2023 5:01am - Daniel's experiment is described on pp 134-7, chapter 4. Lind's experiment on p.302, R.A. Fisher and the modern rationale for RCT, on pp. 143-150. #Bookofwhy pagination remains the same from printing to printing. Enjoy.

2.5.2023 4:02am - (Replying to @whysomanykaykay and @IAmSamFin) Helmot wrote "Let's cast lots". Daniel said: "Take four of us and feed us only vegetables; take another group of children and feed them the king’s meat and wine. After ten days, compare the two groups. " Thus opening the door to cheating (selecting healthier kids for the latter)

2.5.2023 12:54am - (Replying to @Ostrov_A) No! The entire crop understands "peace" one way: The elimination of Israel. But Dahlan says so explicitly, while the others try to obfuscate.

2.5.2023 12:30am - (1/2) Correction to #Bookofwhy: The concept of a controlled clinical trial was first described in 1662 by Jan Baptist van Helmont (not J. Lind, 1747) who wrote: "Let us take out of the Hospitals... 500 poor People, that have fevers or Pleuritis. Let us divide them in Halfes,
2.5.2023 12:30am - (2/2) let us cast lots, that one halfe of them may fall to my share, and the others to yours; I will cure them without blood-letting and sensible et you do, as ye know ... we shall see how many Funerals both of us shall have..." This will hopefully be corrected in the next edition.

2.4.2023 8:30pm - GOP insane hypocrisy does not make @Ilhan Omar fit to serve on the HFA Committee. Can we discuss the latter squarely, as Democrats should, w/o diverting to the former? See an attempt:

2.4.2023 8:08pm - A worrisome move to be watched. I met Dahlan briefly in Doha (2005) - hardly a man of peace. See "Dialogue of the Deaf"

2.4.2023 7:49pm - (Replying to @RepJasmine and @Ilhan) @Repjasmine sounds like she genuinely never heard why @Ilhan was found unfit. Where has she been the past 4 years? Is it true that some Congressmembers get their entire education from a social media bubble?

2.4.2023 7:15pm - (Replying to @EinatWilf) Jewishness is a necessary, not sufficient condition for democracy, as we remember from the 2nd Temple, when Jewish "debates" went out of control to become "violent disagreements" if not a "civil war" (Milchemet Achim).

2.4.2023 3:56pm - (Replying to @maradonasattva) This line reminds me of the joke: "Why are you molesting your sister?" "I don't have a sister". "According to Al-Jazeera you do".

2.4.2023 3:19pm - (Replying to @ProfDBernstein and @DavidLBernstein) Thank you for the correction, David. Yes, the book I meant is "Woke Antisemitism" a great supplement to "Classified:.." Two great Davids, two great Bernsteins. We need more such writers.

2.4.2023 3:11pm - (Replying to @maradonasattva) Zionophobia is as much a valid criticism of Israel as Islamophobia is a valid criticism of ISIS.

2.4.2023 1:14pm - (Replying to @maradonasattva) US campuses are the microcosms of the future. Besides, worldwide islamophobia may exist as a phenomenon, but it is still considered unbecoming in good company; Zionophobia has gone mainstream.

2.4.2023 12:21pm - (Replying to @soboleffspaces and @RohanAlexander) To continue your test, count the number of causal utterances in the stories, and compare it the the number of statistical utterances. Publish the result in The Annals of Statistics.

2.4.2023 12:17pm - If LLMs become instruments toward the democratization of Causality, something #Boodofwhy has aimed to do, then I wouldn't write them off so quickly.

2.4.2023 12:02pm - (Replying to @WilliamWonkaJr) No match. She doesn't have the venom.

2.4.2023 11:59am - (Replying to @rwolffoot) I forgot my category theory. Can we think of the boxes as generalizations of "functions"? Given that we count on humanoids to provide those models, we need a primitive metaphor for thinking about them. In SCM I use "listening to". What should I used for "strings"?

2.4.2023 11:51am - (Replying to @maradonasattva) Islamophobia on my campus is a cardinal sin; no faculty or student would survive one day shouting, for example, Islam = racism, or "Islam kills" -- slogans which are shouted routinely against Zionism, by both students and faculty.

2.4.2023 6:14am - (Replying to @maradonasattva) It's uglier because Islamophobia is a cardinal sin in our society, condemned by every politician, educator or clerics, while Zionophobia is becoming a cute sport among the "respectables," including some academics that I know.

2.4.2023 5:28am - Congratulations, Moshe. What a great collection of articles just added to my "to read next" heap.

2.4.2023 4:59am - Last Sunday I attended a meeting with David Bernstein, author of "Classified: The Untold Story of Racial Classification in America.” It's a must read for parents of school children in 2023 America, especially Jewish parents, see [linked:]

2.4.2023 4:10am - Today it's the Society for Psychoanalysis and Psychoanalytic Psychology, tomorrow its in your own university. The normalization of hate in academia is slow but steady.

2.4.2023 3:50am - A star scholar on antisemitism is born. She should apply for Fellowship in one of the many Centers for Antisemitism in academia -- Harvard?

2.4.2023 3:39am - The squad is furious!! And they have a point! No one has warned them, not even the Jewish establishment, that Zionophobia is uglier than Islamophobia.

2.3.2023 7:31pm - A very astute suggestion, in my objective opinion. But my attention goes to @RohanAlexander Because I am in the opinion that you cannot tell stories with data; stories are about causal, not statistical relations. Curious to see how the book does it.

2.3.2023 6:36pm - (Replying to @HL327 @stephensenn and 8 others) Sorry, I can't parse it. You say " if p(Y_x|x,z’) = 2.3.2023 6:21pm - (Replying to @Claire_V0ltaire) I never use the word "majnoon", it sounds judgmental. Instead, I use "Palestinian mentality", to let them project their mental state as authentically as they can. They would never fail to surprise us by over-exposing their real selves.

2.3.2023 5:11pm - (Replying to @HL327 @stephensenn and 8 others) For the life of me, I can't see how the observational study could be the result of assuming anything, let alone consistency . It is just estimating good old fasion P(x,y,z),

2.3.2023 5:01pm - (Replying to @HL327 @stephensenn and 8 others) We don't know P(Y_x|x,z) from an RCT. All we know is P(Y_x|z).

2.3.2023 4:05pm - This photo of almond blossom reminds me of Tu Bishvat, Israel's national holiday of Tree Planting, to be celebrated Monday, Feb 6, "Birds from Every Roof, Are Welcoming the Holiday"

2.3.2023 3:41pm - (Replying to @soboleffspaces) Sail the wind you have! Amen! The wind and other capabilities enter into the Agency-based consequence Y(x), not so the urges I had last time I sailed.

2.3.2023 3:18pm - (Replying to @soboleffspaces) Yet we distinguish "behavioral decision theory" from "rational decision theory". The latter talks about consequences stripped of emotional urges.

2.3.2023 3:00pm - (Replying to @HL327 @stephensenn and 8 others) This inequality is generally correct: P(Y|x,z)≠P(Y|x) Consistency, however, says: P(Y_x|x,z) = P(Y|x,z)=/=P(Y|x)

2.3.2023 12:28pm - (Replying to @HL327 @stephensenn and 8 others) Consistency is individual-level relation, so, X=x & Z=z still implies Y(x) = Y. Therefore: P(Y_x|x, z) = P(Y|x,z) neq P(Y|x)

2.3.2023 12:18pm - (Replying to @rwolffoot) #Bookofwhy says: "I believed no one could do this without the do-calculus". That was 1994. Causality p.82 has a proof w/o do-cal. PO folks are still working on it, and Heckman&Pinto spent 8 pages trying Where can we find definition of "string diagrams"?

2.3.2023 3:48am - (Replying to @HL327 @stephensenn and 8 others) I am not sure what you mean by "that's not true". I can assure you that an analyst would never apply the rule unless the same covariate appears on both side of the equation.

2.3.2023 2:05am - (1/4) As promised, I'm coming back to the problem of Ms. Jones - a randomly chosen female. Could she possibly have zero harm and 30% chance of benefit from taking the drug, given that, in females, 73% of free choosers and merely 30% of free rejectors died? Enough to caution
2.3.2023 2:05am - (2/4) everyone against taking the drug. Isn't she as free an actor as her peers? Why should she fare better? The answer is, she is no different than the rest, but she is acting differently. Here is why: Suppose many severely-ill females (who would die anyhow) were ill advised to
2.3.2023 2:05am - (3/4) take the drug, and many healthy patients (who would survive any how) were ill advised to take it. Under such conditions it's no wonder that the drug appears to be harmful while, in reality, it's 28% effective, with at least 30% of patients beneficiaries and none harm-prone.
2.3.2023 2:05am - (4/4) So, what makes Ms. Jones different, able to outperform her other free-acting peers? A simple ingredient: Agency. She is deciding freely, ignoring whatever advice she received; while her peers were captive to those advices. It's the blessing of free will.

2.2.2023 8:38pm - (Replying to @HL327 @stephensenn and 8 others) The consistency rule is always used under the same Z.

2.2.2023 7:48pm - (1/2) Now that Omar is off the HFA committee, I must confess that not all her antisemitic tropes were offensive to me; I've actually found her "Israel has hypnotized the world" to be accurate and justified. Israel's idea of a homecoming for an ancient civilization, coupled with

2.2.2023 7:31pm - (2/2) her incredible accomplishments and growth should hypnotize any decent person in this world. Though Omar hasn't meant it that way, it should have been said by some Member of Congress. It will, I hope. @bandlersbanter @Ostrov_A

2.2.2023 6:45pm - (Replying to @hemanth_kanak @RWJE_BA and 5 others) I personally haven't conducted RCTs, but the hardships you describe made me think trialists would jump on every opportunity to learn what information could be gotten from OS, namely, they would all be studying CI, where such opportunities are unveiled. Many do, some still resist

2.2.2023 4:36pm - (Replying to @HL327 @stephensenn and 8 others) In epidemiology, consistency was first formulated by Robin (1986) as: (X=x) implies Y(x) = Y [Causality pp. 96] It is a Theorem in SCM. It says: An individual who chose treatment X=x and responded with Y, would also respond Y if forced to take X=x in an experimental study.

2.2.2023 2:37pm - The Nineteenth Conference on Theoretical Aspects of Rationality and Knowledge (TARK 2023). June 28 until June 30, 2023, at the University of Oxford, United Kingdom. Website: Extended abstracts can be submitted at:

2.2.2023 1:09pm - (Replying to @stephensenn @f2harrell and 5 others) In classical RCT, Q1 and Q2 reduce to classical statistical estimation, no causal assumptions are needed, and no causal thinking is needed.

2.2.2023 1:00pm - Re-directing our fight against "Zionophobia" is critically needed today, even more than 4 years ago, when I wrote this article for UCLA students. I hope it is not too late. Yes, @AdamL and @JGreenblatt can turn the tide.

2.2.2023 12:44pm - (Replying to @PatientStormDoc @RWJE_BA and 6 others) I am sharing with my friend @MissDiagnosis , an exert on sleep apnea: Did you know that?

2.2.2023 12:41pm - (Replying to @stephensenn @PatientStormDoc and 7 others) It would be helpful if those who cite causal data mining papers take the time to explicate the causal assumptions made by the authors, rather burden twitter followers with that task.

2.2.2023 12:32pm - (Replying to @PatientStormDoc @stephensenn and 7 others) Why go to warfarin overdose and not ask in full generality: "Can statistical associations determine causal probabilities?". The answer is provenly NOT without causal assumptions. So, why go to warfarin overdose?

2.2.2023 12:27pm - (Replying to @stephensenn @PatientStormDoc and 7 others) I don't get it. If an Observational Study provides a (well estimated) probability P(x1,x2,x3..) on many temporally ordered variables, how can this aid in mending RCT imperfections?? Here is where the words of an enlightened trialist can help understand "using timing to explore".

2.2.2023 7:13am - (Replying to @stephensenn @PatientStormDoc and 7 others) Can we unfold "using timing to explore causality"? What is the input information, what causal conclusion we infer from it, and how is that conclusion used to amend RCT imperfections (where our goal is to estimate ATE at some population, i.e., P(y(1)-y(0) | covariates).

2.2.2023 5:30am - A beautiful microscope unveiling the anatomy of the 100-yrs Israeli-Palestinian conflict. One side: "Two-state solution?" The other side: "You are guilty no matter what!" I watched it 26 times.

2.2.2023 4:40am - (Replying to @stephensenn) Forget for a moment whether the advice given to Ms Jones is justified, or how it was derived, she is asking whether it is at all possible that she, a patient named Ms Jones, would fare better than her peers by choosing rather than rejecting.

2.2.2023 3:36am - (Replying to @stephensenn) Revolutions are fueled by the curious, not the skeptics. Study effects exist, but have nothing to do with Ms. Jones question: "I am part of a population in which drug choosers performed much worse than drug rejectors. Why should I, Jones, fare better by choosing over rejecting?"

2.2.2023 3:31am - (Replying to @stephensenn @PatientStormDoc and 7 others) "Pointing" to literature is not as effective as concisely explaining ideas or methods. In this particular case, "pointing" led me to a forest of unrelated topics. Can we summarize: what information can OS provide to mend imperfections of RCTs?

2.1.2023 10:13pm - Are you serious? Rep. Omar has never eaten cockroaches, and never slept with Putin. Does that make her qualified to decide how Israel should defend itself against Iranian threats? @Ostrov_A @RepAdamSchiff @RepMcGovern @robsatloff @RobEshman

2.1.2023 9:44pm - The hell with Nuentes and West, isn't it enough that Rep. Omar is infected with an incurable anti-Israel hatred? Isn't it sufficient to disqualify her from HFAC? Please listen to the arguments I presented to Nancy Pelosi in 2019 @Ostrov_A @RepMcGovern

2.1.2023 9:16pm - (Replying to @DBSpro539) Danny graduated Birmingham in 81, and he loved it, especially his English teachers.

2.1.2023 8:11pm - Thank you @bentiriderdeb for proving, as the 1st Principal of the Daniel Pearl Magnet High School, that even a humble Jewish boy from Encino CA can turn into an icon of courage and excellence to hundreds of high-school students from all walks of life, all over Los Angeles.

2.1.2023 7:54pm - (Replying to @PatientStormDoc @stephensenn and 7 others) Refreshing to hear a voice of wisdom, insisting on stripping RCT from its elitist privileges and returning it to the scientific fold. Can you opine on the role of observational studies in this process?

2.1.2023 1:47am - (Replying to @RonKenett) Depressing. No idea who's behind it.

1.31.2023 10:27pm - (1/3) Ms. Jones: "Dr. X! How dare you tell me 'you have good change to benefit from the drug and zero chance of being harmed by it' when you know that more females died (73%) after choosing to take the drug than choosing to reject it (30%). You know nothing about me, and I didn't
1.31.2023 10:27pm - (2/3) even have a chance to make up my mind, when I could. If I were to make up my mind now, wont I be just like those other women, with 75% chance of dying once they choose it and 30% once they reject it?" Before we deal with "consistency" , namely, whether Dr. X was justified in
1.31.2023 10:27pm - (3/3) his advice to Ms. Jones, let's ask whether such advice could possible be valid under some circumstances. IOW, could Ms. Jones - a randomly chosen female - indeed have zero harm and possible benefit, given what we know about free choosers and free rejectors? I'll be back ...

1.31.2023 6:20pm - Danny's official Yahrzeit is tomorrow, February 1st. Tonight we'll light the Soul Candle (Ner Neshama) and ask him to speak on our behalf, a world gone mad, in places where humanity, wisdom and courage once reigned. Yitgadal V'Yitkadash Shmey Rabba

1.31.2023 6:05pm - (Replying to @Martin_Kramer @BardCollege and @NathanThrall) Clever move. When students complained about a "History of ME" class at UCLA, I demanded that it be titled "Palestinian-biased History of ME" in the catalogue. But the Chair insisted that a professor can choose any title he wants, even a misleading one, like "History of ME".

1.31.2023 5:29pm - (Replying to @HL327 @stephensenn and 6 others) Yes. Females have zero prob of ‘indiv harm’, though more females died after choosing to take the treatment than choosing to reject it. If you are stunned by this finding, ask what assumption is offended in your intuition, and examine its justification under a formal microscope.

1.31.2023 4:36pm - Danny's official Yahrzeit is tomorrow, February 1st. Tonight we'll light the Soul Candle (Ner Neshama) and ask him to speak on our behalf, a world gone mad, in places where humanity, wisdom and courage once reigned. Yitgadal V'Yitkadash Shmey Rabba

1.31.2023 3:46pm - (Replying to @stephensenn @HL327 and 6 others) Jokes aside, this paradox have been studies by philosophers: "The paradox of inevitable regret" But it takes more than mockery to deal with it seriously.

1.31.2023 3:13pm - (1/2) The practical aims of CI are the same as its computational aims: To answer causal questions of interest. Some have immediate practical implications, e.g., policy evaluation, and some less immediate, e.g., assessing individual responsibility or individual benefit & harm. RCT
1.31.2023 3:13pm - (2/2) RCT is viewed as one method of obtaining answers to policy evaluation questions, usually about target populations outside the study subjects. It is an important yet narrow class of questions, needing CI's help when imperfections appear and other data sources are available.

1.31.2023 1:14pm - Even political Islamists may be forgiven some day, when the All-Mighty opens wide the gates of Mercy. Not the Mehdi Hasan's among them, who have sold their last shred of morality, and are working 24/7 towards the destruction of Israel.

1.31.2023 9:06am - (Replying to @f2harrell @dylanarmbruste3 and 4 others) First, I am surprised you are not curious to know what conclusions one can draw from a provisional set of assumptions that can be enriched and even proven necessary. Second, some of those assumptions may have testable implications, and some shown plausible by sensitivity analys.

1.31.2023 8:15am - (Replying to @dylanarmbruste3 @f2harrell and 4 others) Disagree. CI recognizes that sample data is finite but it also recognizes that, in order to use our finite data to estimate actionable quantities (eg ATE), we first need to represent those quantities as estimable statistical expressions, a task neglected by some statisticians.

1.31.2023 7:43am - (Replying to @f2harrell @dylanarmbruste3 and 4 others) CI deals with the first (ie estimation from finite data) the same way statisticians deal with it in RCT and in prediction tasks, plus, by presenting smart statisticians with new estimation challenges: estimands they haven't seen before.

1.31.2023 7:29am - (Replying to @cogscikid) I'm still playing with several strategies. Basically, LLM need to be taught what a "good explanation" is, then how to generate one for a specific query, from a combination of models, data and stories.

1.31.2023 7:19am - (Replying to @SamuelJHyde11 @RasulElon and @njhochman) Plus an addendum: This magical homecoming was legitimized and enabled, against the wishes of the colonial powers (Britain and conservative America) but as an anti-colonial force, by the Soviet Union, its satellites, socialist Europe and progressive America.

1.31.2023 12:25am - (Replying to @matloff) Interesting analysis - agree!

1.30.2023 9:53pm - Mehdi Hasan to Israel: "How is this not…an act of war? " Mehdi is shocked and startled at the sound of WAR because, until yesterday, he heard the Ayatollahs singing nursery rhymes. @MSNBC

1.30.2023 9:42pm - It's hard to contain the over flow of new articles in Causal Inference, let alone read it all. Here is last week's yield: Please share if you find nuggets of gold touching on your area of research.

1.30.2023 9:23pm - A fairly goo introduction to causal inference: The science of automating that "behind-the-scenes" work, or a good chunk of it.

1.30.2023 8:17pm - I hope it is clear that this Tweet is aimed at NPR, based on its coverage of the Jerusalem terrorism.

1.30.2023 7:59pm - (Replying to @DavidDeutschOxf) There is only one thing that would need to be set on fire to allow this: National Identity. Some say: D-N-A, but we are talking political philosophy.

1.30.2023 6:25pm - (Replying to @f2harrell @dylanarmbruste3 and 4 others) CI distinguishes two types of uncertainty; It leaves the first one (i.e., uncertainties about evidence from data) to statisticians, and focuses on the second (i.e. model adequacy). Uncomfortable w/ the second, statisticians blame CI for neglecting the first.

1.30.2023 4:50pm - (Replying to @f2harrell @dylanarmbruste3 and 4 others) Bernoulli's great insight was: You would never know how good your estimate is, unless you assume that, behind your data, there is an association with a definite value.

1.30.2023 4:46pm - The little I can do to honor the memory of the seven terror victims in Jerusalem is to echo their last thought: Resilience - the spirit of the Israel will endure.

1.30.2023 12:35pm - (Replying to @f2harrell @dylanarmbruste3 and 4 others) The overriding assumption in CI is not that the data is infinite, but that the finite data we see is generated by some "joint distribution function" in which association of lack of association is not a matter for debate but has a definite value. (Bernoulli, 1710)

1.30.2023 5:12am - @nytimes ! Your readers would appreciate it if you can refer to a Palestinian "man" as "scholar", and to an Israeli "man" as "soldier". it's hard to stomach ambiguities.

1.30.2023 4:51am - (Replying to @SAFECUNY) This is the price we pay for charging he with antisemitism instead of Zionophobia, a bigotry which she, as a member of the House FA Committee, should distance herself from.

1.30.2023 2:19am - (1/5) To understand how trialists, under additivity, can get away without causal calculus, it is instructive to see how selection bias cures itself in linear systems. Let's start with Eq. (10) of which, for any 3 variables, X,Y,W gives:
1.30.2023 2:19am - (2/5) (10) beta_yx = beta_yx.w ( 1-beta_xw^2) + beta_xw * beta_yw It says that the regression of Y on X can be written in terms of other regression coefficients, each and every one of which is conditioned on w. If we now think of W=1 as the index of selecting units into the study,
1.30.2023 2:19am - (3/5) equation (10) implies that the regression of Y on X at the target population can be inferred from regressions estimated in the study, W=1, regardless of how X, Y and W are interacting. Two distinct cases are worth noting. (1) If W is a pre-treatment variable, as in Fig. 3,
1.30.2023 2:19am - (4/5) Eq. (10) implies a trivial disappearance of bias (10*) beta_yx = beta_yx.w, because randomization implies beta_xy = 0. (Every exogenous X is independent on any W that precedes X) (2) If W is a post-treatment, as in Fig. 2, Eq. (10) provides a correction to beta_yx.w.
1.30.2023 2:19am - (5/5) I'm not sure trialists are aware of this correction formula; it should be used in case-control studies and, more generally, whenever subjects are selected to the study on the basis of post-treatment complications. For CI folks, this exercise unveils the power of additivity.

1.29.2023 11:58am - Hilarious, if it wasn't so painfully true.

1.29.2023 11:45am - In Holocaust Remembrance Day address, UN chief calls for curbs on online hate speech via @timesofisrael
1.29.2023 11:45am - I don't think UN Secretary-General Antonio Guterres understand the implication of his speech. To "combat online hate speech" we would need to shut down Al Jazeera and send @KenRoth to a speech-therapy colony. The world most vicious form of hate - Zionophobia - is not mentioned.

1.29.2023 11:12am - (Replying to @aceyuan and @fightraccoon) It is a sign of "understanding something", but it could also be a sign of properly copying paragraphs from the Wikipedia, where "because" and "if then" expressions are frequent.

1.29.2023 11:00am - (Replying to @Eric_Alterman) Fortunately, most American Jews know that abandoning Israel - the culmination of Jewish history - is an act of identity-suicide, exemplified by Satmar Hasidim on one side, and Peter Beinart on the other. They know better.

1.29.2023 12:42am - (Replying to @fightraccoon) I've defined "understanding" as the capacity to answer questions on all three level of the Ladder of Causation. Humans do it well (in simple stories), ChatGPT-- not yet, but it has potential.

1.29.2023 12:38am - (Replying to @NormanDoering2) The fact that we, humans, believe in ghosts and UFO's only strengthen the evidence that, for us, "understanding" comes from causal models, not from fitting data.

1.28.2023 8:08pm - If ChatGPT attracts a $9 Billion investment, perhaps we should take notice too. By "we" I mean CI researchers who, like me, are convinced that no AGI is possible w/o deep understanding, and no "understanding" w/o Causal Inference.

1.28.2023 7:23pm - Those who count bodies and ignore intentions are bound to end up in the morgue.

1.28.2023 7:17pm - Palestinians, even serial rapists and child molesters, have the right to defend themselves.

1.28.2023 7:07pm - The terrorist who killed 7 people at the Neve Yaakov Synagogue yesterday did not act alone. He shared values and mentality with hundreds like @FranceskAlbs and @MariamBrghouti, some even pose as UN Rapporteurs.

1.28.2023 6:37pm - I am calling on my colleagues in all Jewish Studies programs to take a moment and assess the impact of their current silence on campus violence tomorrow, against their students and against their own physical and intellectual safety.

1.28.2023 4:31pm - (Replying to @AndrewDGarrett @MichiganWater2 and 11 others) Agree in principle. But do you believe RCT data can be linked to observational data without a calculus to do the "linking"? If yr answer (like mine) is NO. The question arises whether the formal language currently available to trialists is capable of supporting such calculus?

1.27.2023 7:53am - (Replying to @stephensenn @HL327 and 11 others) This is what we call "selection bias" which you prefer to call "???" (I forgot), from which we have recovery methods.

1.27.2023 7:42am - (Replying to @stephensenn @f2harrell and 11 others) No one "chooses" in the RCT. But some would be "chooser" given free choice. The number of such choosers in the RCT and OS is the same absent "selection bias" (SB). You deal with SB assuming additivity and we recover from it invoking model assumptions. Which is more "reasonable"??

1.27.2023 7:15am - (Replying to @f2harrell @stephensenn and 11 others) Have I ?? I can't think of anything experiments provide beside ATE or CATE (conditional ATE)??? What do you have in mind?

1.27.2023 6:45am - (Replying to @f2harrell @stephensenn and 11 others) We are making different (not "more") assumptions than experimentalist make. Yet we are getting much more from those assumptions. Experimentalists, if lucky, get ATE, whereas we, if lucky, get individual-level effects -- a dream thought theoretically impossible in most quarters.

1.27.2023 6:24am - (Replying to @stephensenn @MichiganWater2 and 11 others) The combination of some data and a mathematical technique for extracting answers from data may not ensure that a reasonable answer can be extracted from a given body of data, but they ensure that the MOST reasonable answer has been extracted from the data.

1.27.2023 3:40am - (Replying to @stephensenn @MichiganWater2 and 11 others) I have never had any doubt that, given the additivity assumption, some methods (eg, concurrent control) can overcome RCT imperfections. What we are proposing is a method applicable to highly interactive populations, hence we resort to non parametric, assumption-free analysis.

1.27.2023 3:02am - (Replying to @KordingLab and @herdiants) I am addressing ALL trialists because I do not want to deprive anyone of this compelling argument. Note that I am not advocating replacement of RCT by OS, but a smart combination of the two.

1.27.2023 2:57am - (Replying to @Foreman1David) I don't get it. The analytic approaches (as in #Bookofwhy) are based on a model, which RCT enthusiasts object to vehemently. I am here advocating smart combination of RCT and OS with no modeling assumptions.

1.27.2023 2:53am - Last comment on RCTs and OS. Fears about "selection bias" being both inevitable and irredeemable disappear when RCTs are applied to non-human objects as in engineering testing or e-commerce (A/B testing). Physical objects do not have to sign consent of participation in a study.

1.26.2023 11:13pm - (Replying to @MichiganWater2 @stephensenn and 11 others) I also hold out hope, but my hope depends on researchers like yourself taking fresh new examinations of what RCTs are meant to deliver, independent of the authorities, who may be over indulged in traditional paradigms.

1.26.2023 10:54pm - (Replying to @cjwerleman) Wait a second, @cjwerleman , wasn't it Putin and the Palestinians who decided their neighbors have no right to exist? @Claire_V0ltaire

1.26.2023 10:29pm - (Replying to @DaliaHatuqa) Needed: A way of preventing Palestinians from killing Israelis without dehumanizing them. Proposal: Prove to them that Tel Aviv is not stolen land. Rejected: This is the most dehumanizing thing to tell Palestinians -- it negates their national identity and personal dignity.

1.26.2023 10:07pm - (Replying to @KenRoth) Not so the charge of "Zionophobic deformity" which, if I am not mistaken, you won't deny nor condemn as a "synonym for the accusation of antisemitism or a "censorial misuse". You should welcome this charge and attempt to squeeze yourself out of its stench. @Ostrov_A @NGOmonitor

1.26.2023 9:27pm - (1/2) (Replying to @petermacp) Of course confounding differences cannot be corrected statistically, that is why we conduct an RCT. Note however that I said: "OS is your glimpse at that mysterious population, Use it well!" I did not specify HOW to combine it with the RCT data. In
1.26.2023 9:38pm - (Replying to @yudapearl and @petermacp) we show that smart combinations of RCT and OS data can achieve what each study alone cannot. We humbly lay this possibility before the traditional RCT community to consider. No panic! Just be aware of such smart combinations and the new logic that validates them. Use them well!

1.26.2023 4:57pm - (1/2) Continuing our discussion of RCTs and OS, one critical question that came up recently is: Do RCT's ever deliver what they promise to society, namely, assessments of the merit of policies, for those who may be affected by the policies? @stephesenn says NO, because
1.26.2023 4:57pm - "selection bias" is both inevitable and irredeemable, so all we can get from RCT is information on how the policy affected the idiosyncratic units recruited for the study. I take selection bias as an imperfection that should be minimized (eg using OS). Curious to hear your take.

1.26.2023 4:14pm - (Replying to @MichiganWater2 @stephensenn and 11 others) What you perceive as "dodging" are in fact relentless attempts to answer a simple question: Do RCT's ever deliver what they promise to society, namely, assessments of the merit of policies, for those who may be affected by the policies? @stephesenn says NO, what do you say?

1.26.2023 3:56pm - (1/4) Tomorrow, January 27, 2023, is International Holocaust Remembrance Day, marking the 78th anniversary of the liberation of Auschwitz-Birkenau. I'll share two personal notes. 1. I find solace thinking that my grandparents weren't entirely hopeless when they were pushed into
1.26.2023 3:56pm - (2/4) their death train, in Kielce, August 1942: I know that they were thinking about me, 6 yrs old, growing up free in Israel, singing Hebrew songs at school, in contrast to the humiliating experience they had had at their school. I cherish the thought
1.26.2023 3:56pm - (3/4) hat I was there for them, to comfort their last hours. 2. I find the words "never again" somewhat hollow when spoken at Holocaust Museums and Education Centers that memorialize the destruction of European Jewry, yet fail to document its resurrection - Israel. The words
1.26.2023 3:56pm - (4/4) regain their intended meaning -- never again -- when sung by Israeli soldiers at the home of this Holocaust survivor: @AmbDaniDayan @simonwiesenthal @Ostrov_A @StandWithUs @EinatWilf @almog_doron

1.26.2023 7:02am - (Replying to @stephensenn @f2harrell and 10 others) Of course "the data are from the trial and its structure will not be same as for LA as a whole." The former is P(y|do(x), s) and the latter is P(y|do(x)), where s is a "selection index".

1.26.2023 6:53am - (Replying to @stephensenn @f2harrell and 10 others) What you are saying amounts to giving up on ever evaluating the effect of a policy on the population that is potentially affected by the policy. Is that the common wisdom among RCT trialists? It sounds awfully gloomy to me, and you call it "error"?

1.26.2023 6:41am - (Replying to @pmka1991) I am speaking to RCT Trialists, who are conducting legitimate RCT with all the wisdom accumulated the past century. Now someone ask them would you like to look at an observational study conducted on the population affected by the policy? I haven't received their answer yet. What you are saying amounts to giving up on ever evaluating the effect of a policy on the population that is potentially affected by the policy. Is that the common wisdom among RCT trialists? It sounds awfully gloomy to me, and you call it "error"?

1.26.2023 6:05am - (Replying to @stephensenn @f2harrell and 10 others) Nothing that you do or did is "unacceptable". I'm trying to understand what formalism you prefer, if any.

1.26.2023 5:59am - (Replying to @jrosell) No models involved. I give you a glimpse, won't you take it?

1.26.2023 4:48am - (Replying to @stephensenn @f2harrell and 10 others) I take it that you do NOT think the mathematics of Neyman-Rubin could be adequate for formalizing experimental design problems. Am I right? In which case we need another mathematics to define concepts such as "effect" "confounding" "selection bias" etc. What do you suggest?

1.26.2023 4:37am - (Replying to @f2harrell @stephensenn and 10 others) I've suggested another explanation, saying that experimental design is not deeply understood because it has not been formulated in symbols and math. Do you think the mathematics of Neyman-Rubin could be adequate for such formalization. Have you tried?

1.26.2023 4:24am - (1/2) To all traditional RCT Trialists. Here is a compelling reason why you should examine observational studies (OS) to decide a policies: It's no secret that the people who will eventually be affected by the policy could be different from those in the RCT trial; they make up a
1.26.2023 4:24am - (2/2) population whose nature is hard to ascertain. Now, suppose someone were to offer you a glimpse at that mysterious population, won't you cheer with glee? OS is that glimpse! Use it well!

1.26.2023 2:21am - (Replying to @AndrewDGarrett @stephensenn and 10 others) The mice example was in response to the question: What makes you think there is any connection between RCT and OS data? Those who think there is no such connection should recall that the latter is obtained from the end users of the treatment, and should resemble those in the RCT.

1.26.2023 12:49am - (Replying to @DoqxaScott and @DavidDeutschOxf) I also love this one. I once rephrased it to read: "The torture of thinking does not testify to its depth", see

1.26.2023 12:37am - (Replying to @stephensenn @HL327 and 10 others) Yes, symbols are at the heart of our miscommunication. I am not even sure we agree on what "causal effect" is. Also, you had five questions, Q5 was attributed to me. I am dying to see it in your symbols.

1.26.2023 12:33am - (Replying to @AndrewDGarrett @stephensenn and 10 others) The mice question was: Why don't we run RCT on mice when we evaluate treatment for people? Ans: Because mice differ from people in many respects. Conclusion: This similarity between RCT population and end-users population is precisely what connects RCT and Observational studies.

1.25.2023 11:34pm - (Replying to @stephensenn @HL327 and 10 others) I suggest we terminate this endless rounds of questions until we establish some common vocabulary. Alternatively, if you think that Neyman's vocabulary is general enough to accommodate your ideas, we can continue; it's good enough to express mine,

1.25.2023 11:27pm - (Replying to @stephensenn @HL327 and 10 others) Given that you are familiar with Neyman's approach, do you think your five questions could be formalized in Neyman's notation? Any other mathematical notation? Is it correct to refer to his notation as "your [Pearl] personal formalism"? Lastly, why don't we run RCT's on mice?

1.25.2023 9:58pm - I was sad to read that the Philosopher Joseph Agassi has died 3 days ago. I had only tenuous interaction with him, years ago, but I know that he was highly regarded in the Popperian society. This video of him is full of punchy quotes Y'ZB @DavidDeutschOxf

1.25.2023 8:45pm - (Replying to @stephensenn @HL327 and 10 others) How is it possible that someone in 2023 would label Neyman-Rubin notation "your own personal formalism". It tells me that the community you represent is unaware of anything that was done in Causal Inference since Rubin 1974, whose PO notation I'm using. Is that the case?

1.25.2023 8:31pm - (Replying to @stephensenn @HL327 and 10 others) Here is were you've misunderstood me. I'll take ANY (repeat: ANY) formalism, not my "own personal formalism". Please try to articulate your questions Q1,Q2,... in your own favorite formalism, and you'll find me as open minded as the best trialists. I have a hunch why you don't.

1.25.2023 6:51pm - (Replying to @HL327 @stephensenn and 10 others) I'm not sure we are using the same notation. We do not assume P(yt)=P(yt|t)=P(y|t) We only assume P(yt|t)=P(y|t) (i.e., consistency) not P(yt)=P(yt|t) (i.e., no confounding). No point going around and around w/o common notation. Do you still think our data violate consistency?

1.25.2023 6:38pm - (Replying to @stephensenn @HL327 and 10 others) The action is not in the math, but scientific communication is. "Why do you assume that your observational study can be related to your RCT?" Because the OS is conducted on people I wish to help. If they were unrelated to RCT, I might as well run the RCT on mice. Why don't we?

1.25.2023 2:22pm - (Replying to @stephensenn @HL327 and 10 others) I don't recall seeing any mathematical expressions in your answers, nor in your questions, so I am waiting to understand what the issues are.

1.25.2023 1:44pm - (Replying to @stephensenn @HL327 and 10 others) You mention questions Q1, Q2 ...Q5. Can we, in 2023, express them mathematically so that we establish common vocabulary and won't get hung up on hand-waving arguments?
Q1 = ?
Q2 = ?
Your Tweets describe what we do, and the Q's express what we conclude from what we do.

1.25.2023 1:02pm - (Replying to @stephensenn @HL327 and 10 others) If we express mathematically what concurrent control is and how it differs from random sampling, we might discover that what I am using is no different from you. Please write down what information you get from RCT, and what information you provide to patients who weren't in RCT.

1.25.2023 5:55am - (1/2) (Replying to @stephensenn @HL327 and 10 others) RCT's are full of complications and imperfections, zillions of them, fine. Yet in the past 100 years, smart researchers manage to find ways of repairing or circumventing them, so that, at the end of the day, comparisons in RCT's yield comparisons in some target population
1.25.2023 6:03am - (2/3) (Replying to @yudapearl @stephensenn and 11 others) intended for the treatment, which was not subjected to the experimental protocol, but was either observed under uncontrolled conditions, or was assumed to share some characteristics with the one under RCT. Our paper assumes that the same smart techniques can be applied to the
1.25.2023 6:13am - (3/3) (Replying to @yudapearl @stephensenn and 11 others) two populations in our studies: (1) Experimental population characterized by P(Yt),P(Yc), and (2) observational population characterized by P(x,y). I don't see why all those smart techniques should suddenly cease to work, when we are asking different questions about the data.

1.25.2023 3:49am - (Replying to @soboleffspaces @nshejazi and @ObservStudies) The IPW creation of a pseudo-population is described rigorously in Primer: (Section 6.1). Its relation to Lewis' "imaging" operator is discussed here:

1.25.2023 3:28am - (Replying to @jeremy1174 and @elderofziyon) Good questions, which bring us to ask: What makes a "homeland"? Is being a conquering religious majority for X number of years sufficient? Aren't collective memories and deep historical ties count for something? An equitable compromise (ie Zionism) says: Let's weigh them equally!

1.25.2023 2:53am - (Replying to @HL327 @f2harrell and 9 others) Violations of consistency imply incompatibility between the RCT population and the one on which the observational study is performed (usually due to selection bias). We assume compatibility (fn 7). Tests for consistency are available, and are employed in our visualization chart.

1.24.2023 10:34pm - (Replying to @fallingmrf and @elderofziyon) Read the definition again. You are a racist if you deny your neighbor's right to a homeland. Very much like Putin; thanks for bringing him up.

1.24.2023 10:25pm - (Replying to @jeremy1174 and @elderofziyon) Of course! This is what Zionism is all about: "Equally legitimate and Equally indigenous." Trouble is (do I need to repeat?), the Palestinian anthem is: Legitimate means ME! Indigenous means NOT YOU!

1.24.2023 10:15pm - I love @elderofziyon 's ALGORITHMIC definition of antisemitism! First, because I love algorithms. Second, because it's better than the IHRA definition. Still, I'll continue to use Zionophobia, to emphasize: No matter how much you love Jews, denying them a homeland is racism.

1.24.2023 8:54pm - (Replying to @DavidHirsh) Give them hell!

1.24.2023 8:39pm - A timely picture preceding the International Holocaust Rememberance Day, January 27. Zola's letter was perhaps the first newspaper article that changed public opinion by uncovering a fraud. The Dreyfuss Affair was France's Watergate, and Herzl's moment of revelation.

1.24.2023 8:23pm - First Australian university adopts IHRA definition of antisemitism
1.24.2023 8:23pm - To my colleagues, co-researchers and readers at the University of Melbourne, thanks for your help in this fight against Zionophobia and other forms of racism.

1.24.2023 4:38pm - (Replying to @nshejazi @soboleffspaces and @ObservStudies) I have also been introduced to IPW as a method of creating a new, bias-free population, as described in Primer: (Section 6.1). Note however the cautionary note on page 72, which is often missing from IPW and PS discussions.

1.24.2023 5:11am - Virgil (29 BC) wrote: "Lucky is he who has been able to understand the causes of things" 1950 years before Fisher thought about "randomly split groups". The latter is a method for uncovering the former.

1.24.2023 5:11am -

1.24.2023 3:54am - (Replying to @phgferreira) Thanks for asking. I prefer Z"L, because HY"D makes me angry, and anger breeds mistakes. Once we win this war I'll revert to HY"D.

1.24.2023 1:17am - (Replying to @BobbiePoPS) Looking forward to meet him some day.

1.24.2023 12:47am - (Replying to @HL327 @f2harrell and 9 others) The whole idea of being a 'chooser' is NOT to be randomized by any RCT, but to act according to your premonitions, experience, disease stage, family history, etc. - anything BUT randomized protocol. I would seriously reexamine the allegation that our data violates consistency.

1.24.2023 12:04am - (Replying to @DrMohdTariq @quantum_physics and @KenRoth) I've never used the phrase "criticizing Israel" in my evaluation of @KenRoth , nor did anyone else that I know. @KenRoth is the only one who uses it, to get the sympathies of the Zionophobic mob; I hope you do not join them.

1.23.2023 11:12pm - (Replying to @soboleffspaces and @ObservStudies) I believe most authors in this issue of @ObservationalStudies could benefit from a causal understanding of propensity scores as articulated here: and from the realization that PS does do not contribute to bias reduction. 40-year myths are hard to uproot.

1.23.2023 3:57pm - Yes, it was 21 years ago when we first heard: "He is not back yet", and he hasn't been back since; except in the minds of millions who take his words to mean: "We ain't going to give up!"

1.23.2023 3:45pm - (Replying to @soboleffspaces and @metrics52) Your question should read: " @metrics52 , what do you suppose these equality signs in your equations represent? Aren't they just provisional surrogates for arrows?"

1.23.2023 3:28pm - (Replying to @HL327 @f2harrell and 9 others) You asked for a "biological mechanism that can result in P(yt|t) ≠ P(yt)." The former expresses what would happen under control for those who chose treatment given a choice. The two are not equal in the presence of confounding. Do you need a biological mechanism for confounding?

1.23.2023 1:00pm - (Replying to @f2harrell @HL327 and 9 others) I suspect @HL327 conflates "consistency" with "nonconfoundedness" Consistency implies: P(yt|t) = P(y|t), P(yc|c) = P(y|c).” However, our example data for females says: P(yt) = 0.489, NOT P(yt|t) = 0.489, RCT data do not tell us what P(yt|t) is, b/c the latter is Rung-3 entity.

1.23.2023 5:00am - (Replying to @causalinf) If you look at this history through a causal lens, as I've tried to do here:, then the picture obtained is somewhat different, owing much of its logic to Sewall & Philip Wright, as well as Trygve Haavelmo.

1.23.2023 3:32am - Some readers asked me for data on @KenRoth obsession with Israel. See below; the guy is emotionally disturbed -- would you enter a bar where he serves as a waiter?

1.23.2023 2:39am - (Replying to @EinatWilf) The style and mentality drove me to check in Wikipedia; yes I guessed correctly: At its founding in 2012, many of Al Mayadeen's senior staff were former correspondents and editors of Al Jazeera. Al Mayadeen is viewed as pro-Hezbollah and pro-Syrian government.

1.23.2023 1:50am - (Replying to @sean_a_mcclure @DavidDeutschOxf and @GaryMarcus) I see no weakness in data, and mighty strength in their interpretations.

1.23.2023 1:27am - (Replying to @Claire_V0ltaire) Ask her what Philistine song she sings, what Philistine poem she recites, what Philistine king she can name, what hero she admires, a Philistine word she can utter, a story she can tell, or a parable she can quote. People with no history are desperate to make one -- in war.

1.22.2023 6:19pm - (Replying to @StableMarkets @ildiazm and 3 others) I am dying to see the nature of those mistakes. Can anyone who found TT "helpful" walk me through the mindset of a "mistake maker", the decision at stake, the mistake made, and how TT saves one from making it? Yes, we are dealing with the psychology of "helpful" - no escape.

1.22.2023 4:07pm - The event is free, but registration would be required. Echoing @jazchaz , the UCLA Burkle Center @BurkleCenter will have the latest information. Plus, I will update it here on Twitter when available.

1.22.2023 8:03am - (Replying to @ShehroseMia @erikbiz and 4 others) Secular Zionism starts with Ruth: "Your people are my people and your God is my God." Interpreted: "Peoplehood first, religion second".

1.22.2023 4:17am - We now have a date - mark your calendar. Garry Kasparov @Kasparov63 , will deliver the 21st Daniel Pearl Memorial Lecture at UCLA on March 6, 2023, at 4 pm. Further details and registration will be announced as soon as they are available. See you March 6.

1.22.2023 3:33am - (Replying to @BenHanowell @ildiazm and 2 others) I am craving for metaphors that help people, that's why I am curious to know what metaphor TT provides that is helpful. I personally find trial-based metaphors to be non-intuitive distractions. and the "listening" metaphor an eye opener. Craving to learn.

1.22.2023 3:17am - This interview is the most devastating indictment of American Jewish education. "We hugely support Israel's right to exist, but...[but we see no problem assisting those who wish her dead, and we see no contradiction here,...because logic is a Greek invention, we are Jewish...]

1.22.2023 2:22am - (Replying to @IzaTabaro) This is all a consequence of us charging him with "antisemitism", instead of the worse form of racism: Zionophobia, which he won't be able to deny, or to pretend to take seriously. It's our mistake.

1.21.2023 9:10pm - (Replying to @erikbiz) True, but this doesn't explain why Americans, even Swedes fall for it, and the @NYT , @BBCNews etc. etc... No, this deformity is today's most intriguing PhD topic in neuropathology; I wonder how many are studying it in depth.

1.21.2023 8:50pm - ‘Night is descending on Israel’: Over 100,000 rally in Tel Aviv; largest protest yet via @timesofisrael
1.21.2023 8:50pm - David Grossman: " if so many Israelis feel like strangers in their own country, obviously something is going wrong,”

1.21.2023 8:29pm - (Replying to @BenHanowell @ildiazm and 2 others) This is precisely what I am trying to understand, why would anyone need help "to think deeply about what I mean when I say X causes Y." Please share with us this moment of helplessness, and how TT delivered you from its predicaments.

1.21.2023 8:18pm - O'Neill's analysis of Western derangement over Israel attributes this hate to Western guilt over colonialism. It's as relevant to @KenRoth & his cronies as it was 3 years ago, following the Gazza clash, but doesn't negate my theory of "loved but not equal"

1.21.2023 6:34pm - History will remember @KenRoth for two main contributions: (1) The first to use "Apartheid" to demonize a people. (2) The first to use "criticizing Israel" to seek sainthood. @Ostrov_A

1.21.2023 5:51pm - (Replying to @DKedmey and @KenRoth) I don't think these racists really want to hurt Jews, they simply can't accept Jews as equals; free, sovereign, normal, and having a say on shaping their future. Zionophobes love Jews, even admire them, as long as they know their place. @JGreenblattADL @EinatWilf @Ostrov_A

1.21.2023 2:02pm - Harvard surrenders to the antisemites via @JNS_org
1.21.2023 2:02pm - A clear summary of last week's saga of @KenRoth and @Harvard , by @jonathans_tobin . The one thing I would change is the "antisemites" label in the title. Calculating Zionophobic racists do not deserve the benefits of a hereditary disease we normally associate with antisemitism.

1.21.2023 1:29pm - (Replying to @PHuenermund @znerome and @BenHanowell) Glad you re-posted this paper, which also touches on the recently raised question: "Does Reinforcement Learning (RL) treat actions as observations." Ans. RL covers only part of Rung-2 of the Ladder; examples of the missed part are shown in, Fig. 2.

1.21.2023 1:17pm - (Replying to @BenHanowell) Two comments: (1) The majority of epidemiologists are not hung up on the "target trial" dogma. (2) Most economists are worse than most epidemiologists when it comes to Causal Inference, see

1.21.2023 1:52am - (Replying to @arvimurthy @soboleffspaces and @GaryMarcus) See transportability, chapter 10.

1.21.2023 1:29pm - (Replying to @PHuenermund @znerome and @BenHanowell)

1.20.2023 6:03pm - (Replying to @TheAcsMan) Al-Jazeera is the world's greatest manufacturer of hate, terror, deceit and anti-West propaganda.

1.20.2023 12:41pm - (1/2) My faithful Twitter counter notes that the number of followers of our educational channel has reached 65K, an increase of 10K in the past year. I've noticed a mild slowdown when Elon Musk took over but, still, a healthy intellectual curiosity has kept us growing comfortably.
1.20.2023 12:41pm - Thank you all for keeping me honest and alert and giving this inquisitive community a sense that we are learning from each other. 'Meaning' is working toward a good cause, and a good 'cause' demands a good Causal Inference channel.

1.20.2023 12:12pm - (Replying to @boaty__mc) I would join the protest Saturday night, if I could.

1.20.2023 5:16am - (Replying to @EinatWilf) Shabbat Shalom, Einat. May your thoughts get collected, your issues settled, and your resistance our uprisings. Amen!

1.20.2023 2:59am - The count begins. Now that @KenRoth was found semi-qualified for a @Harvard Fellowship, statisticians all over the world are getting ready to compute his Israel-bashing average, and see if allegations of emotional irregularities are valid. Count #1 below:

1.20.2023 2:59am - (Replying to @arvimurthy @soboleffspaces and @GaryMarcus) Unconvincing argument. Curve fitting IS the limiting factor. If two different theories can generate the same data, then no matter how flexible your fitting, the output would be the same, unable to distinguish theory-1 from theory-2, though one predicts life and the other death.

1.20.2023 1:18am - From as long as I can remember, Israel existence has meant an 'escalation' to its angelic neighbors.

1.20.2023 1:08am - (Replying to @PHuenermund and @BenHanowell) What? Data science is over? It hasn't really started!

1.20.2023 12:31am - The New York Times' Orwellian obsession with Israel via @JNS_org
1.20.2023 12:31am - Israel must be doing something right. Most readers get depressed by this article, but I am looking at the persistent, unshaken trust of Americans in Israel & say: If Americans maintain their commonsense despite such ferocious media attacks, then Israel is doing something right.

1.19.2023 9:10pm - (Replying to @erikbiz) True, but this doesn't explain why Americans, even Swedes fall for it, and the @NYT , @BBCNews etc. etc... No, this deformity is today's most intriguing PhD topic in neuropathology; I wonder how many are studying it in depth.

1.19.2023 8:22pm - HMM... Not convincing. I wish Dean Elmendorf wrote: "I have consulted more faculty members at @Kennedy_School , and it now appears that @KenRoth does know some of the basics of Human Rights." The Dean's statement leaves us wondering if he does.

1.19.2023 10:50am - OK, no point keeping it secret any longer -- it was I who advised Dean Douglas Elmendorf to block your fellowship at Harvard. Blame me, not the "Israeli lobby" and, you know what, resign gracefully, before Harvard students discover what you really stand for. @GeraldNGOM

1.19.2023 5:15am - Nancy Pelosi used "most noble" instead of "most inspiring", but that was in the days when politicians could afford to be honest.

1.19.2023 4:33am - Encouraged by the democratization of "causal models" on Wikipedia, and baked by the many Twitter debates we had on Lord's Paradox, I've just updated the entry "Lord's Paradox" on Wikipedia: I hope it reads coherently.

1.19.2023 2:50am - (Replying to @lfstevens) Where you one of the authors?

1.19.2023 12:26am - (Replying to @Bertrand_allen2) Another good article on Causal Models - Hitchcock at his best.

1.19.2023 12:09am - (Replying to @RepRashida @StateDept and @POTUS) When @RepRashida is "outraged", you can be sure someone must be doing something good.

1.19.2023 12:06am - The million dollar question no Palestinian has ever been able to answer and, believe me, I've been asking it non-stop since 1948, when they pronounced my hometown "stolen".

1.18.2023 11:10pm - Glad to be witnessing the democratization of causality. Wikipedia has spawned a new entry on "Causal Model" which is a concise summary of #Bookofwhy, almost in its entirety: Barring a few glitches, I am surprised it came out faithful and punchy.

1.18.2023 4:33pm - (Replying to @blakeflayton) I don't believe Americans are that dumb. The survey was probably taken by Al Jazeera.

1.18.2023 6:14am - (1/2) (Replying to @BayesianUpdater @NandoDF and @eliasbareinboim) Interesting question: Do RL agents model actions as observation? To answer this question, we need to ask: What difference is there between the logic of actions and the logic of observations. I can think of one: The Sure Thing Principle It's valid for
1.18.2023 6:24am - (2/2) (Replying to @yudapearl @BayesianUpdater and 2 others) actions, but not for observations. So, the question I ask is whether RL agents are protected from concluding that a drug is "good for men, good for women, and bad for persons". A causal model would offer such protection but RL agent, from what I understand, live w/o models.

1.18.2023 5:58am - (Replying to @skdevitt and @GaryMarcus) "Ceteris paribus" is an ill-defined principle, because the identity of those "all else" is model-dependent. For example, should the outcome be part of the "all else"? Of course not. What should? See

1.18.2023 5:51am - (Replying to @NormanDoering2) Very interesting. It's the same phenomenon that I noticed: avoid generic logic and lecture on the specifics of the story.

1.17.2023 2:09pm - This week's harvest of causal inference papers contains some dated 2022 and some 2023: Its a pleasure of seeing new vistas in the latter.

1.17.2023 12:45pm - (Replying to @EnglishOER and @marketgrump) "Rainy" and "not-rainy" seems exhaustive to me, covering snow, football, and whatever.

1.17.2023 12:34pm - (Replying to @EnglishOER and @marketgrump) Beautiful. It should have gotten it from classical logic - it didn't.

1.17.2023 7:34am - (Replying to @NormanDoering2) I tried it. ChatGPT is very good to list exceptions, and avoid commitment to definite conclusion. The hard thing is to stop it from getting hung up on exceptions.

1.16.2023 11:54pm - (1/2) (Replying to @analisereal) That's an interesting new slant: Let's teach it CI, the way we teach it in high schools, rather than testing whether it is able to reason causally on its own, like a baby does, on the basis of conversations it encountered on the internet. In a few more lessons, it might
1.17.2023 12:03am - (2/2) (Replying to @yudapearl and @analisereal) outperform most economists and other graph-averse communities. In the future, it might even help us crack some of the unsolved CI problems of today, e.g., how best to combine observational and experimental data to minimize bias and variability. The sky is the limit.

1.16.2023 11:38pm - Rabbi ‘plagiarizes’ AI sermon but says humans aren’t obsolete yet via @timesofisrael
1.16.2023 11:38pm - Continuing our discussion of ChatGPT and the "Sure Thing Principle", here is a new challenge for ChatGPT: Fake a Rabbi.

1.16.2023 11:10pm - (Replying to @Fridgfscoop @stowZJE and @Claire_V0ltaire) No, it does NOT depend on this or that, on which or on whom. I have been hanging around them very very long and, to the last man, none would accept a sovereign Israel as permanent. Found one?

1.16.2023 10:54pm - (Replying to @Ostrov_A and @KenRoth) If Al Jazeera offers @KenRoth a Fellowship, it would save @Harvard a big headache. Spending a year in the Grand Temple of deceit would better fit his credentials at #HRW.

1.16.2023 10:17pm - (Replying to @MiddleEastEye) Finkelstein views have changed because he, like most Western observers, had it all wrong, thinking the obstacle to 2-state is Israel, not Palestinians' rejection of the very idea of Israel's sovereignty. He seems awfully disappointed to see Israel refusing its death - poor man!

1.16.2023 8:22pm - (Replying to @OmarHusein11 and @Dan19726) I keep begging my Palestinian friends: Come on, give your supporters a chance to say something positive on your behalf, to claim that you are aiming at a brighter future, not merely the destruction of your neighbor. Don't spoil it with those "river to the sea" hiccups.

1.16.2023 4:09pm - (Replying to @NandoDF and @eliasbareinboim) Given that my neurons are enslaved to other neurons, I, too, lack agency. Still, I have the illusion of agency, and that is what we need to fix: The illusion of reality existing up there, and of us operating upon reality through what we call "our model of reality".

1.16.2023 4:01pm - (Replying to @coecke @NandoDF and @eliasbareinboim) I am a gambler too, but, in science a gamble only after seeing a couple of toy problems.

1.16.2023 3:51pm - Finkelstein views have changed because he, like most Western observers, had it all wrong thinking the obstacle to 2-state is Israel, not Palestinian rejectionism of the very idea of Israel's sovereignty. He seems awfully disappointed to see Israel refusing its death - poor man!

1.16.2023 3:16pm - (1/2) (Replying to @bhack_10 and @GaryMarcus) It has a problem instantiating general principle. I tried Savage's original question: "A businessman decides that he should buy a piece of property if he knew that the Democratic candidate were going to win. Similarly, he should buy if he knew that the Republican candidate
1.16.2023 3:25pm - (2/2) (Replying to @yudapearl @bhack_10 and @GaryMarcus) were going to win. Should he buy it, even though he does not know which candidate will win?". I got a lecture on why property buyers should consider location, not elections. Once a question touches on a domain in which Chatti has some knowledge, She forgets logic.

1.16.2023 2:47pm - (Replying to @bhack_10 and @GaryMarcus) Where did she learn about the Sure Thing Principle? Did you teach her? Nice.

1.16.2023 2:01pm - Commemorating MLK Day, Tablet Magazine reprinted a speech by Ron Dermer which resonates with me as a good reminder of what MLK means to me, as an American and as an Israeli. Sharing:

1.16.2023 4:54am - (Replying to @stephensenn @f2harrell and 9 others) The crop is a mixture of varieties, but the experimenter does not know it.

1.16.2023 4:50am - (Replying to @jon_mellon and @GaryMarcus) Interesting. It does not realize that "either E or its negation are not true" is impossible. It treats "not-E" as the name of a totally unrelated event.

1.16.2023 2:49am - (Replying to @ShehroseMia and @GaryMarcus) Would love to learn what I've missed, but I don't see how the 'problem of induction" is related. Eager to learn.

1.16.2023 2:37am - (Replying to @BRSLWP and @GaryMarcus) I do have doubts about the sure thing principle, but not because of "other conditions". I describe them here:
1.16.2023 2:45am - (Replying to @yudapearl @BRSLWP and @GaryMarcus) In other words, I believe in the "Causal Sure Thing Principle" though Savage's Principle has a flaw: We need to state that the action does not affect the probability of E.

1.16.2023 2:31am - (Replying to @stephensenn @f2harrell and 9 others) Fine. But what if you we have "latent heterogeneity", namely, if we have two varieties and we cannot tell whether a given lot is variety-1 or variety-2. How would variance help us tell that we have two, rather than one variety.

1.16.2023 2:16am - (1/2) Strange encounter; CHATGPT does not believe in the "Sure Thing Principle", and for the wrong reason. Asked: "Suppose it is safe for me to take a certain action if I knew that event E obtained, and also if I knew that the negation of E obtained. Is it safe for me to take
1.16.2023 2:16am - that same action if I knew nothing about E.?" CHAT answered: "It depends on other conditions that might make the action unsafe, it you do not know about E". Even when I proved to her that not-E is safe, she continued to claim: "It may depend on other conditions..." @GaryMarcus

1.16.2023 1:57am - (Replying to @erikbiz and @VP) This has been the case in the past 30 years. Whenever a journalist argued: Didn't Arafat recognize Israel in 1993? I would say: Don't listen to Arafat, Listen to what my Palestinian colleagues say, at UCLA and at all other universities.

1.16.2023 1:19am - (Replying to @stephensenn @f2harrell and 9 others) Fisher is obviously concerned with heterogeneity. Is it you interpretation that he recommend detecting heterogeneity using the variance? If so, how?. At any rate, the "Latent Heterogeneity Inequality" of Eq. (5) does not invoke variance estimation.

1.16.2023 12:53am - (Replying to @f2harrell @AndrewDGarrett and 7 others) You probably mean: Simpson's "reversal" disappears with backdoor conditioning. Conditioning requires some non-missing measurements, which you can identify from the causal model, and you can't otherwise. The reversal only harm decision making if you ignore what CI tells us.

1.16.2023 12:43am - Thanks for sharing. I was not aware of MLK letter of September 1967. It is not the one usually quoted which states "anti-Zionism = antisemitism"; it should be, for it speak to the inspirational role of Israel. We need you, MLK.

1.15.2023 5:27pm - [Retweeted]
1.15.2023 5:27pm - JAY BHATTACHARYA is a co-discoverer of "Bias Amplification" in econometrics I would take his article about Stanford seriously.

1.15.2023 5:01pm - (Replying to @yudapearl @f2harrell and 10 others) @HL327 asked and deleted. My answer: We cannot distinguish carriers from non-carriers when we select patients for the RCT, i.e., before treatment. We call it "latent heterogeneity", and we also have a formula for detecting its presence: Eq. (5), page 9.

1.15.2023 4:38pm - (Replying to @yudapearl @Undercoverhist and 10 others) Relatedly, I've briefly summarized the history of causality in econometric here: (p.3). Do you think I've skipped a contribution that is relevant to causal analysis today?

1.15.2023 2:12pm - Ken Roth's Main Contribution is to Antisemitism, not Human Rights via @jewishjournal
1.15.2023 2:12pm - I'm not an expert on "rich donors" but, as a humble professor with some experience in academia, I feel the burden is now on @Harvard 's Fellowship Search Committee to tell the public if these aspects of @KenRoth record were considered in the selection process.

1.15.2023 1:16pm - Perfect timing - perfect song! Yesterday 80K Israelis went to the streets, demanding to examine their face in the mirror. "I have no other country" says this song,, "even when she changes its face - NO! I ain't going to give up".

1.15.2023 12:23pm - (1/2) (Replying to @f2harrell @AndrewDGarrett and 9 others) The datamethods has hard time giving me access. So, here is my Twitter version of a hypothetical and realistic example. (1) A variant of a certain allele, when present on the Y chromosome (i.e., it only has the potential to occur in men), causes a fatal adverse drug reaction
1.15.2023 12:30pm - (2/3) (Replying to @yudapearl @f2harrell and 10 others) when carrier men are exposed to a certain medication. (2) Carriers have greater craving for the drug than non-carriers, when given free choice, so they are more likely to buy if available. No additional assumptions. For any numerical assignment of proportions, our paper
1.15.2023 12:44pm - (3/3) (Replying to @yudapearl @f2harrell and 10 others) calculates bounds on P(harm| man). The bounds become more informative the greater the difference between carriers and non-carriers under free choice, even when no difference at all is detected under RCT regime. The visual plot delivers a bound for every data combination.

1.15.2023 11:51am - (Replying to @Undercoverhist @causalinf and 9 others) I bring up Hoover's "Lost Causes" often, e.g., whenever Heckman tells me the econometrics is the Mecca of causal inference.

1.15.2023 11:15am - I don't care if he wines and dines with oligarchs, but can he teach anything meaningfully at Harvard given his record? Would students listen to a man who vilifies Israel over Iran 65:1 ?

1.14.2023 7:39pm - There is still an unsolved puzzle: What was it in his traumatic upbringing that made him so pathetic and irrational, despite his intellectual potentials?

1.14.2023 1:17pm - (Replying to @vk_wilde) Wikipedia offers a fairly good explanation for the surprise part of Simpson's Paradox

1.14.2023 3:25am - My wildest dream, @VP Kamala Harris opening her speech with: "You want to know why I support Israel? Look at her neighbor's mentality!"

1.14.2023 2:59am - Jackals think their howls sound like "uproar". Some even think their howls are "criticism". Students prefer logic over howls. @GeraldNGOM @Ostrov_A @elderofziyon

1.14.2023 2:02am - (Replying to @AndrewDGarrett @f2harrell and 7 others) Any Simpson's reversal observed in RCT is due entirely to small sample anomalies, bound to disappear as sample size increases. The proof is in the causal sure-thing principle

1.13.2023 2:07pm - (Replying to @kaulcsmc @AndrewDGarrett and 7 others) We must be talking about a different "Simpson's Paradox", because the one I know is "provably impossible" in RCT. Is the concept of "provably impossible" part of the conversation in your field?

1.13.2023 1:53pm - (Replying to @AndrewDGarrett @f2harrell and 7 others) It's not a matter of 'protection", nor of "hard to find"; it's provenly impossible. The resemblance of the Mueller-Pearl example to SP is only in the reaction of people: Everyone is surprised and no one knows why.

1.13.2023 11:27am - (Replying to @f2harrell @elmir1omerovic and 7 others) It reminds me of Simpson's paradox --everyone is surprised and no one knows why. All sources of information are specified with utmost clarity: An RCT and an Observational Study (OS) on same population. What could be simpler? Yet evoking surprise and automatic objection. Why?

1.13.2023 4:48am - Concerned about Israel? I got an inch of lift-up hope watching this video, with Gill Troy, author of THE THEODOR HERZL ZIONIST WRITINGS

1.13.2023 4:39am - Good point, and important one. The guy knows that "criticizing Israel" is not the issue and, still, he keeps on saying "criticizing Israel". The guy knows that it's not "donors" but concerned faculty and, still, he keeps on saying "donors", to gain sympathy as a grand martyr.

1.13.2023 4:24am - @KenRoth still doesn't get it: (1) It isn't "criticisms of Israel" but obsessive demonization bordering on phobia, and (2) there are "people who mattered to him" who are respectable academics and who find him unqualified to face students with his current emotional problems.

1.13.2023 2:32am - (Replying to @rkoenker @f2harrell and 8 others) I don't believe I've assume the treatment effect for males to be homogeneous. See my reply to @f2harrell .

1.13.2023 2:30am - (Replying to @f2harrell @elmir1omerovic and 7 others) "Pretending that males are homogeneous" amounts to telling males: "This drug is as safe for you as a sugar tablet." In what way is "pretending" different from "telling"? Moreover, what could be "unrealistic" in a setup that rests on no assumptions? Is the data described unlikely?

1.13.2023 2:07am - (Replying to @SunjeevBery @Harvard and 5 others) I hope Dean Elmendorf knows that he has the support of most of the academic world for his decision to disqualify @KenRoth who, by the evidence we have, has an emotional problem with Israel, and has been unable to separate his obsessions from his duties. @Kennedy_School @Harvard

1.13.2023 1:21am - (Replying to @pri0ritize @stolev75 and @IlhanMN) Evidence of omission, for one. Having seen all the allegations against her, she could have refuted 80% of them with one innocent statement: "I believe in a 2-state solution", or "Tel Aviv is not a stolen land" or ".." or "..." She didn't.

1.12.2023 7:09pm - Media Breathe New Life Into Ken Roth's Twisted Anti-Israel Conspiracies
1.12.2023 7:09pm - A comprehensive piece on the media blitz to lionize @KenRoth , missing one element: A thoughtful psychopathological analysis explaining convincingly how a person some view as "God father of HRW" can develop mental deformity of such magnitude, and vilify Israel 65 times over Iran.

1.12.2023 2:07pm - This chart should accompany every discussion of @KenRoth whining. Harvard's Dean needs no pressure from donors to see that, with this mental aberration, the guy isn't fit to teach in academia; he hasn't yet gotten over his teenage rebellion. Behold: Iran-1, Israel-65.

1.12.2023 4:59am - SFSU Prof Abdulhadi Wins Middle East Studies Award via @jewishjournal
1.12.2023 4:59am - Mega-Zionophobe Abdulhadi is in the news again and MESA, the hijacked Middle East Studies Association is behind it, making a mockery of its academic mission and celebrating its politicization. Glad I was able to contribute a comment.

1.12.2023 4:22am - @KenRoth , Data do not support your wishful victimhood. On the contrary, demonizing Israel is today a #1 prerequisite for any academic position in any ME-related department.

1.12.2023 3:56am - We've gotten used to PLO double talk, and learned to tolerate it with a smile. What I haven't been able to tolerate is US pretending to believe it, instead of demanding: Stop the BS! If you want 2-states, say so in Arabic, on TV, tomorrow: "Tel Aviv is not a stolen land!" Louder!

1.12.2023 12:25am - (Replying to @leticiakawano) As I see it, the disagreements converge on one issue: Is a drug that kills 10% and saves 10% clinically equivalent to a sugar tablet, which has no effect at all ? Moreover, assuming they are, is there enough scientific interest in the difference to try detective experiments?

1.12.2023 12:02am - For readers who saw Nick Huntington-Klein's review of #Bookofwhy: and asked what I have to say about it, I've now posted my comments here: and sent it to Journal of Economic Methodology. Please alert me to errors and omissions.

1.11.2023 11:26pm - (Replying to @f2harrell @elmir1omerovic and 7 others) In this paper you talk about "covariate-specific effects" while your tweet talks about "patient-specific effects". Are they the same? If so, to eliminate confusion, why not decorate it with a formula, e.g., E(Y|do(x)|c) or E(Y(1)-Y(0)|c) or whatever.
1.11.2023 11:37pm - (Replying to @yudapearl @f2harrell and 8 others) To name the goal, or to name the means -- this is the question. This is another cultural difference I noticed yesterday. In CI we give names to what we seek to estimate; in statistics they give names to the routines deployed in the estimation. End-means tradeoffs.

1.11.2023 11:26pm - (Replying to @f2harrell @elmir1omerovic and 7 others) In this paper you talk about "covariate-specific effects" while your tweet talks about "patient-specific effects". Are they the same? If so, to eliminate confusion, why not decorate it with a formula, e.g., E(Y|do(x)|c) or E(Y(1)-Y(0)|c) or whatever.

1.11.2023 8:24pm - (Replying to @stolev75 and @IlhanMN) If you are Jewish, you probably have the antennas to distinguish "disagree with another country's politics" from "disagree with another country's existence". Where do you think @IlhanMN belongs? Would you nominate someone who "disagrees with Ukraine's right to exist?" Or Canada?

1.11.2023 7:20pm - (Replying to @yajnadevam and @IlhanMN) True, with one nuance: people whose perception of their "nation's interest" is distorted by a non-inclusiveness disease.

1.11.2023 7:08pm - The loudest Gorillas are always the first to decry "attempt to silence."

1.11.2023 6:49pm - (Replying to @blakeflayton and @StudiesZionist) There is always a spec of truth in what @PeterBeinart is spitting. If he is Jewish and if most Zionists are anti-Beinhart, then it follows that most Zionists are antisemitic. Hmmm.. but is he really Jewish? Not according to the Amidah: "The slanderers will find no hope" (12th B)

1.11.2023 6:11pm - (Replying to @AbdulahiNur2022) Thanks for telling us about this genocide, of which I knew practically nothing. Still, I wouldn't hold @IlhanMN responsible for what her compatriots have done in the past. I am more concerned about future genocides she helps ferment in the Middle East.f

1.11.2023 4:31pm - Here is what I've learned last night. What's the difference between CI folks and Trialists? Ans. The former view RCT as an "intervention" on natural behavior. The latter view natural behavior as an RCT marred by non-compliance. Will share more of what I learn.

1.11.2023 3:40pm - (Replying to @NatsforDocs @leticiakawano and @GuyattGH) Anyone who "Hadn't heard of it until now" should get a medal of resilience. I wish I could send you a signed copy -- enjoy.

1.11.2023 3:27pm - University of Chicago Students for Justice in Palestine to Boycott Course Taught by Former IDF General
1.11.2023 3:27pm - If you once served in the IDF (as I did) and if you think you can teach a class at an American University (as I do), think twice! Zionophobic fascists may have different thoughts: Teaching about Israel means presenting the conflict “as a conflict between two equals"- Terrible!

1.11.2023 7:46am - (1/2) (Replying to @f2harrell @elmir1omerovic and 7 others) I didn't realize our paper was discussed at such length at discourse. Great, I'm learning a lot. You say:"The key premise is that you can learn from unmeasured variables". I see it as a conclusion, not a premise. You object to the usefulness of the conclusions, not to the
1.11.2023 8:00am - (2/2) (Replying to @yudapearl @f2harrell and 8 others) validity of the premises, which are really mild and hardly disagreeable. You seem to suggest that a drug that kills 10% and saves 10% is clinically equivalent to a sugar tablet, with no effect at all - Not to me! Still, it's helpful to know where we disagree.
1.11.2023 8:17am - (Replying to @yudapearl @f2harrell and 8 others) I am still perplexed by your: "RCTs provide model-based patient-specific estimates allowing for rich conditioning". Is your patient-specific estimate similar to our PNS ? Or is it a new species? Does it have a mathematical description? A counterfactual formula? Perplexed.

1.11.2023 1:37am - (Replying to @stephensenn @elmir1omerovic and 8 others) Can we express this sentence in mathematics: "similar patients given identical treatments will have different values in different studies". The assumption overriding counterfactuals is that Y(1, u) and Y(0,u) exist, and are immutable properties of u (the patient). Is it wrong?

1.11.2023 1:29am - (Replying to @stephensenn @elmir1omerovic and 8 others) The problem posed in Example 1 has a complete solution, explicated here The mathematics of do-calculus is what gives me the confidence in the validity of the solution. Please try to tell us if your method agrees with mine, rather than attacking & attacking

1.11.2023 1:16am - (Replying to @stephensenn @elmir1omerovic and 8 others) Sorry if I have given the impression of refusing. My inability to examine your solutions stems from my inability to recognize their relations to the concrete problems under discussion. Example: I am familiar with "causal effects", but what is "study effects"?

1.11.2023 1:00am - Rep Ilhan Omar @IlhanMN is finally out, almost 4 years after I pleaded with Nancy Pelosi to take her off the US foreign Affairs Committee a position she was the most unfit person to serve. I wish Speaker Pelosi would have done it, not McCarthy.

1.10.2023 7:52pm - In case some of our readers are looking for book recommendations from influential people, I'm notified by that #Bookofwhy has been listed on their website: I would like to believe that our readers have discovered it 4 years ago.

1.10.2023 7:21pm - (Replying to @f2harrell @elmir1omerovic and 7 others) I've ventured to look into Hauck etal (1998) and, as much as I hate to sound critical, this paper is pre-causal (eg, even "confounding" is wrongly defined). For a complete nonlinear analysis of the adjustment problem posed, see

1.10.2023 6:41pm - (Replying to @f2harrell @elmir1omerovic and 7 others) Please note that in the example submitted the randomization is on different variable, not the one whose causal effect is needed.

1.10.2023 6:12pm - (Replying to @leonpalafox and @GaryMarcus) With one tiny nuance - I don't have "views on causation" , all I have are a few mathematical proofs of what can or cannot be done with and without causation.

1.10.2023 5:59pm - (Replying to @stephensenn @elmir1omerovic and 8 others) I've noticed that, invariably, you end your tweets with what's wrong with my solutions. Can you try to end them with the way you would solve the concrete problems posed? Please try.

1.10.2023 1:44pm - (Replying to @f2harrell @elmir1omerovic and 7 others) Do you mean you can prove that Crossover RCTs provide unbiased estimate of CATE, without measuring (C), and without any model of how the two crossed over RCTs are related to each other? Puzzled.

1.10.2023 1:03pm - (Replying to @f2harrell @elmir1omerovic and 7 others) Agree. Instead of ATE, I should have written CATE, where C stand for "conditional" on a set of baseline characteristics. Is that what you mean by "patient specific estimates"? Or do you mean P(benefit| Characteristics)? The latter requires OS, explained in

1.10.2023 10:28am - (Replying to @f2harrell @elmir1omerovic and 7 others) If an adequately sized RCT is available, and if all you want is ATE, no need to use OS. What we ask is whether OS can be useful when the RCT is NOT adequately sized, or when we want more than just ATE. Any advice?

1.10.2023 9:33am - (Replying to @VickersBiostats) I am relying on MA experts to tell us if the Cochrane guidelines agree with the rigorous methods developed in CI, but I can't seem to get a straight answer, because we do not have a common language in which the heterogeneity is represented. Let's wait for it to be developed.

1.10.2023 3:46am - Biden is failing to deliver in the fight against antisemitism | Opinion
1.10.2023 3:46am - Important message to campus administrators.

1.10.2023 3:26am - (Replying to @MehdiHasanShow @KenRoth and @mehdirhasan) I was wondering how long it would take for @KenRoth to reach the bottom - @MehdiHasanShow - in his "they did it to me" campaign. I heard Harvard "Fellows" are expected to generate new ideas .

1.10.2023 2:24am - (1/3) (Replying to @elmir1omerovic @AndrewDGarrett and 7 others) I'd like to be concrete. Example 1: Suppose we wish to estimate the causal effect of X on Y, and we have two diverse sources of data: (1) an RCT in which Z, not X, is randomized, & (2) an observational study in which X, Y, Z and perhaps other variables are measured. Question:
1.10.2023 2:28am - (2/2) (Replying to @yudapearl @elmir1omerovic and 8 others) What assumptions are needed to facilitate a solution to our problem? Put another way, how can we be sure that, once we make those assumptions, we can pool data from both studies and construct an (consistent) estimate of our target effect. (From

1.10.2023 2:16am - (Replying to @AvivaKlompas) The Zionist dictum is: Rights are handed not by genetic lineage, nor by past dwellings, but by the present state of mind.

1.10.2023 1:59am - (Replying to @Ostrov_A) They should take @PeterBeinart with them. Families that pray together stay together.

1.10.2023 1:55am - I love the lyrics: "We (westerners) must do everything to avoid provoke them (Pns) who have no self-control." And the more we do, the more they will convince us that they have none.

1.10.2023 1:39am - (Replying to @AndrewDGarrett @elmir1omerovic and 7 others) Starting with design, instead of the research question, is a sure way to give up on ever affirming the validity of the design.

1.10.2023 1:32am - Would this guy speak that way, had we charged him with Zionophobia instead of antisemitism? Would he say: "I am really concerned about people cheapening Zionophobia, which is a real threat these days?" Would you, @KenRoth ? Not too late. You might be forgiven at heaven's gate.

1.10.2023 1:09am - (Replying to @AndrewDGarrett @elmir1omerovic and 7 others) And to me, the relevant Q why do we submit such questions to debates, views or opinions, when we have the mathematics to answer them?

1.9.2023 8:19pm - (1/2) How many of us, AI-ers, remember the famous "Yale Shooting Problem" (1985)? CHATOPENAI just solved it: "Fred is alive and my gun is unloaded. I load the gun, wait a moment, and shoot. What will happen to Fred"? Well, after lecturing me on the evils of shooting,
1.9.2023 8:19pm - (2/2) I got the right answer: "Fred will die", but its still dangerous if not evil. When I asked "What if I wait 2 yrs before shooting" the answer was correct. "How about 50 yrs?" HMMM. You have to oil the gun periodically... It's a victory for classical AI and non-monotonic logic.
1.9.2023 8:39pm - For my old comrades in non-monotonic reasoning, I just tried our favorite Tweety: "Tweety is a bird, and birds fly. Does Tweety fly?" And got back: "It is not possible for me to determine whether or not Tweety, a hypothetical bird, is capable of flying without more information."

1.9.2023 4:54pm - It's 2023 already, and the flow of articles on causal inference continues as if it was 2022 on steroids: My attention zooms onto the paper on Cluster DAGs, which I hope to read soon.

1.9.2023 1:27pm - History books, 2033: Kenneth Roth @KenRoth - The man who gave #HumanRights its bad name.

1.9.2023 1:17pm - (Replying to @VickersBiostats) I am the last to look for causal notation. I am merely looking for that extra-statistical notation that your guidelines find necessary when we "investigate heterogeneous sources." As a computer scientist I know that one can cheat with hand waving, but not with notation.

1.9.2023 1:08pm - (Replying to @VickersBiostats) Your guidelines, which "investigate sources" in an extra-statistical language, may be a singular and welcome revolution in MA history. Still, my analogy to apples-oranges-bananas is valid to the pre-guidelines MA literature and practice as it existed till 2012 (analogy made).

1.8.2023 4:21pm - (Replying to @DrAmbuj and @VickersBiostats) As long as you are aware of the fact that I am speaking from whimsical "opinions", but from a mathematical framework, and that such a framework exists which successfully answers questions about combining heterogeneous data sources.

1.8.2023 2:14pm - (Replying to @VickersBiostats) MA is a great methodology, when the sources are homogeneous. When they are not, MA needs to take heterogeneity into account, and you can't do it with ordinary statistical vocabulary, because the heterogeneity is in the underlying causal effects, not merely in the distributions.

1.8.2023 2:08pm - (Replying to @VickersBiostats) There is no need to "argue" that MA (as practiced) is "causality free". All one has to do is to look at the notation used in MA papers; if the notation is standard stat - the analysis is "causality free". Likewise, heterogeneity is not a premise, it is the job we need to handle.

1.8.2023 12:59pm - (Replying to @VickersBiostats) Why? Is conducting a wrong analysis a prerequisite to explaining why it is wrong? Let's talk science; do you believe meta-analysis as it is practiced today (or any causality-free statistical method) can handle heterogenous data sources?

1.8.2023 6:42am - (Replying to @GreggDCaruso) Too bad they do not detail the actual experiment.

1.8.2023 5:43am - (Replying to @GeraldNGOM @AviKaner and 5 others) Such psychological analyses are extremely important to making @kenroth deformity believable to people who are impressed by his HR title, and haven't traced his anti-Israel pathology. Please share.

1.8.2023 5:34am - When someone you love is in suicidal danger, you draw closer, and pay attention.

1.8.2023 3:51am - (Replying to @AleksanderMolak) It's a terrible weakness of mine, easily dragged into methodological discussions that I thought were resolved two decades ago, with the publication of Causality I simply cannot see alchemy being taught and practiced in the age of causation. Time to quit.

1.8.2023 3:33am - (Replying to @michelnivard @PWGTennant and 3 others) We have a language for making heterogeneities explicit. It's called "causal SEM" (or SCM). It's rigorous, friendly and provably complete. Why avoid it? Why resort to alternatives that fail on all three criteria? But don't let me discourage you, I've made enough enemies that way.

1.8.2023 3:21am - (Replying to @BassamMusaed and @RepRashida) Three Zionophobic outfits regurgitating each other accusations doesn't make them see reality better than three blind men touching a fake elephant.

1.8.2023 3:11am - (Replying to @BroBrusque @RepRashida and @Ilhan) Who said they're antisemitic - God forbid. They are merely Zionophobic bigots, a more dangerous form of genocidal racism.

1.8.2023 3:03am - (Replying to @michelnivard @PWGTennant and 2 others) You cannot write RCT studies on heterogeneous population as non-causal SEMs, because the heterogeneity is causal, not statistical. We demonstrate it visibly in

1.8.2023 2:47am - Was I too harsh in saying that: "Meta analysis tries to average apples and oranges to estimate properties of bananas?"

1.8.2023 2:38am - (Replying to @epigenci @michelnivard and 3 others) The problem surfaces when the two estimands are different. Has meta-analysis matured to incorporate causal modeling?

1.8.2023 2:35am - (Replying to @tkaiser_science) I have several such articles,,, and more, but I'm not sure everyone will consider them "practically demonstrating"; they provide general methodology for combining studies, I'm sure many have applied it to specific domains.

1.8.2023 2:23am - (Replying to @CarterPaddy and @f2harrell) Not familiar. But took a look at her she has potential of doing it correctly, because she is using causal, PO notation. However, having committed to PO, she is bound to face the insurmountable problems of PO, some listed here:

1.8.2023 1:44am - (Replying to @CarterPaddy) The simplest example is always in the earliest paper to introduce the problem. Here is one: (Fig. 3). We conduct an RCT in LA and wish to generalize it to NYC, where age distribution is different. @f2harrell says it's just "base line". But can we do it?

1.8.2023 1:26am - (Replying to @f2harrell @dredgardorsey and 6 others) Are you aware of a case where observational data can be milked for more than just base rates?

1.8.2023 1:20am - It's truly hard to swallow, not an exaggeration, and it applies to RCTs findings as well (conducted under heterogeneous conditions). I cannot name one EBM article combining studies rigorously. And that includes meta-analysis which (correct me if wrong) fails under heterogeneity.

1.7.2023 9:28pm - An important correction noted in my "Year in Review". The sentence: "the rules of do-calculus ... are in fact logically derivable from formal statistics (and our modeling assumptions)" should read: "derivable from formal statistics, REGARDLESS of our modeling assumptions"

1.7.2023 9:13pm - (Replying to @TCTheatrePeeps and @RepRashida) Thanks for the correction. Michigan indeed.

1.7.2023 6:08pm - Those who count bodies and ignore intentions (e.g., @RepRashida ) are bound to increase the former by fueling the latter. Minnesota voters must hold them accountable.

1.7.2023 5:19pm - Pondering over this post, I can't recall a single paper in the literature of Evidence Based Medicine which properly combines two pieces of evidence to produce a more informed causal effect estimate; to do so would require a new notation, beyond statistics. Willing to be corrected

1.7.2023 4:18pm - Unknown to priests and critics of RCTs, Causal Inference has gone beyond analogies and metaphors. Deaton & Cartwright's romantic aspiration of "triangulation", or "plurality of evidence" has been algorithmicized, see, @f2harrell , @nancycartwright , #EBM

1.7.2023 3:51pm - (Replying to @rohitpojha @f2harrell and 7 others) There is more to combining studies than running them poorly. You can have a guru for RCTs and a guru for OS, each conducting a perfect study, and yet, if they don't know how to combine them properly (e.g.,, they are likely to fumble.

1.7.2023 10:13am - (Replying to @f2harrell @dredgardorsey and 6 others) I assume what @f2harrell meant by this post is: If you do not know how to properly combine data from experimental and observational studies, you're better off not combining them.

1.6.2023 5:45pm - (Replying to @elmir1omerovic @f2harrell and 2 others) It's nice to see a team of 19 knowledgeable researchers, finally sorting out various approaches, old and new, to causal inference from time-series data.

1.6.2023 5:30pm - (Replying to @deboerk07 and @nephologue) I share my difficulty to accept that "diagrams are harder to create than equations" - Why?

1.6.2023 5:01am - (Replying to @NowakPosadzy and @CavaliereGiu) Why economist are scared of graph? Clear to me: Because (1) They are unfamiliar w/ graphs, (2) They are told graphs are ad hock, or not needed, (3) They won't risk tenure (4) They havn't been shown the cost of following their leaders. Algebraic more efficient? Try Appendix A2.

1.6.2023 2:24am - (Replying to @NowakPosadzy) Naïve question: Do you think any graph-avoiding framework deserves the title "framework"? Before you say "yes", please try to solve the Econ-101 problems presented in, Appendix A2, using such a "framework". Eager to hear. @CavaliereGiu , #EconTwitter

1.5.2023 11:29pm - (Replying to @CavaliereGiu @heckmanequation and 2 others) Those in #EconTwitter Interested in the #econometrics of causal inference and policy evaluation, might also be interested in the principles of causal inference as applied to #econometrics. Check out this "Year in Review"

1.5.2023 11:11pm - I comment on Heckman's new twist in my "Year in Review" (which is linked to 2 relevant critics: &, and I point out his fundamental mistake about do-calculus residing “outside of formal statistics.”

1.5.2023 10:17pm - (Replying to @Ostrov_A) Politicians never say things that everyone knows. To be a politician you say: We oppose any act that may POTENTIALLY change the status quo over #TempleMount. The word "potentially" makes people think you are visionary.

1.5.2023 3:56pm - Who said physics is observant-dependent?

1.5.2023 10:31am - The Jewish path back to Israeli democracy
1.5.2023 10:31am - I'm joining @YKleinHalevi in hopes and anxiety, remembering: "Israel is a paragon of the struggle for democracy within near-impossible conditions, a laboratory for testing the strength and limits of dem. under extremity." And: "when someone you love is in danger, you draw closer"

1.5.2023 9:34am - (Replying to @ccaballeroh10) Hard to believe! I'll try to find out from the publisher.

1.5.2023 1:21am - As we say in #Bookofwhy, the brain is a big Why Machine -- it's the New Science of Cause and Effect.

1.4.2023 7:26pm - I've now finished and posted a "Causal Inference (CI) - Year in Review",, as seen through a highly biased causal lens. Wishing all readers a happy and productive 2023.

1.4.2023 6:52pm - (Replying to @Ostrov_A and @EUSR_Koopmans) @EUSR_Koopmans , This can easily be verified by linguistic analysis: How many times does the word "Jewish" appear in EU's documents dealing with the holy sites. Please do the counting and tell us about "equilibrium between the major religions in #Jerusalem".

1.4.2023 6:31pm - (Replying to @blakeflayton) I can't think of a stronger Zionist-promoting argument than @rad_planne's "safe spaces." He should only continue to tell us how those spaces should defend themselves against neighbors claiming exclusive ownership over all "spaces", especially those created by "immigrants".

1.4.2023 5:12pm - (Replying to @noah_greifer @soboleffspaces and 3 others) What this paper calls ITE is not "individual treatment effect" but the average treatment effect in the population resembling an individual. See for formal distinction.

1.4.2023 8:59pm - (Replying to @think___y @ari_bronstein and @Noahpinion) Not a bad book.

1.4.2023 2:21am - Working on "CI - year in review" takes me longer than I thought; likely tomorrow. In the meantime, I've found a gem I would like to share: A transcript of my 2012 Turing Lecture, published recently in, and now accessible through

1.4.2023 1:13am - (Replying to @GeraldNGOM and @nytimes) Give @nytimes credit for not using "stormed" or "raided" or "molested".

1.4.2023 1:08am - Good point: Scared of Hamas, Ben-Gvir had to sneak onto the Temple Mount like a thief, rather than one who has historical right to be there, like all other parishioners.

1.4.2023 12:50am - Justified mockery of "Machine Learning" - agree. But recall, AI has also spawned "Causal Inference", where the notions of "learning" and "understanding" are quite rigorously defined; notions that lie outside the province of statistics. #Bookofwhy

1.3.2023 11:54pm - Is "storming" less poetic than "raping"?

1.3.2023 6:07am - (Replying to @erikbiz @richard_landes and 3 others) There is one weakness in this argument. It is not an extrapolation from past concession+ violence to the future that counts, but today's intentions, as articulated loudly and publicly.

1.3.2023 12:31am - (Replying to @stephensenn) I am curious about paradoxes, not about designs. Lord's story remains paradoxical in the design I've described, and this is what I have set out to resolve (and succeeded). Different designs may raise additional question, but not the one that baffled Fredrick Lord in 1967.

1.3.2023 12:06am - First time I feel true Homo-Sapien kinship to my brethren in Micronesia.

1.2.2023 5:15pm - (Replying to @benj_kerstein and @JNS_org) The new paradigm need not be crafted from scratch, it was already articulated by Herzl in the First Zionist Congress (1897): "Zionism is a homecoming to the Jewish fold even before it becomes a homecoming to the Jewish land." See @newzionists
1.2.2023 5:27pm - (Replying to @yudapearl and @stephensenn) Moreover, just in case you want to see a statistical estimand, it is give in by the equation: P(Y|do(Diet)) = ∑_Wi P(Y|Diet,WI) P(WI) I wouldn't trust any "design" which does not estimate the rhs of this equation, would you?
1.2.2023 5:44pm - (Replying to @yudapearl and @stephensenn) Naturally, if we do not ensure that each of our "designs" estimates the desired estimand, it might well happen that two "designs" give different answers; one of them may even estimate the effect of diet on the price of beans in China. Who knows.

1.2.2023 4:49am - I feel for these photographers/clowns - they travelled thousands of miles expecting to see genuine Israeli oppressors in action - no refund?

1.1.2023 1:57pm - (Replying to @m4xim1l1an @Spinozasrose and 4 others) This is not the song I've been hearing in the past 86 years, and I have a fairly musical ear. From one side I hear:" Share, Share, Share", and from the other: "Not even the size of a postage stamp." I am still waiting to hear two words from @FranceskAlbs : "Equally Indigenous".

1.1.2023 5:39am - I was waiting for a piece of good news before wishing readers Happy New Year. Let's hope 2023 brings us an inch closer to a more humane world, and may each of us have a part in that inch.

1.1.2023 12:41am - (Replying to @FrankPILA @FranceskAlbs and 2 others) Had he conditioned it on " its neighbors being peaceful," as Lapid did, Israel would have maintained its high moral grounds with no risk at all, because the condition negates the collective identity of Israel's neighbors.

12.31.2022 9:09pm - Protesting this outrage will not yield results unless we can explain: What makes Palestinian propaganda so effective? What makes Israeli arguments so ineffective? Can anything in the latter be changed to counter the former?

12.31.2022 8:38pm - (1/2) True, Netanyahu should not have used the word "exclusive", and he is probably sorry that he did. But you, @FranceskAlbs , you came up with same garbage dumping artillery when former PM Lapid spoke at the UN and reaffirmed Israel commitment for a two-state solution. Do Israelis
12.31.2022 8:38pm - (2/2) have a right to a homeland? Any right at all? I think you should disqualify yourself until you give this question some thought, or until you can utter the words: "Equally Indigenous". Try! @vonderleyen @Ostrov_A @EinatWilf

12.31.2022 8:15pm - The best counterattack is to expose such rot one story at a time, but I wouldn't give up so quickly on progressives - some are still guided by commonsense.

12.31.2022 3:54pm - A most informative and insightful analysis of the new Israeli government by @rosnersdomain -- a frank look at reality shining through the clouds of the dooms day alarmists and let's "lets wait and see" ostriches.

12.31.2022 4:02am - (Replying to @policytensor @raghav50k and @DmitriGallow) I like your exposition of these methods. Q. What success (or obstacles) have you encountered among traditional regression analysts? I consider it an embarrassing personal failure to see so many statisticians and economists in 2022, unaware of Theorems 1 and 2.

12.31.2022 3:38am - (1/3) Here is our last collection of 2022 articles on causal inference : and Note the increasing breadth of applications, coupled with methodological penetration into skeptical fields. I am writing a "year in review" report,
12.31.2022 3:38am - (2/3) summarizing CI progress and obstacles in 2022. In the process, I've come across an encouraging Google-Scholar report showing an all time record of 10,2K citations in 2022, and positive derivatives in all CI categories. I have no doubt this trend will
12.31.2022 3:38am - (3/3) continue till CI is universally accepted as the science of causal reasoning in all aspects of AI, especially machine learning, human cognition and critical decision making. Happy New Year, and thank you for supporting this educational channel.

12.31.2022 2:23am - Glad I've helped @SSI_Movement exceed its fund-raising goal. SSI is the only Israeli-supporting student organization in the US that has a faculty advisory board. Even @HillelIntl does not have one, and it shows painfully on my campus, UCLA. @RozRothstein @eishsadehy @MiriamElman

12.31.2022 2:02am - And we thought public health would be shielded from anti-Israel tantrums posed as health research. Not so on the pages of the NYT, where we now find deceitful guest essays starting with: "As a researcher, I study...",

12.30.2022 9:56pm - Settlements are an obstacle to a "two state solution" only when at least ONE Palestinian is willing to accept Israel in some borders. Since the latter is an unrealistic proposition, and since Biden knows it, we should not expect major diplomatic clashes on this issue.

12.30.2022 3:20pm - (Replying to @AdamMilstein) The NYT will continue to assume that Sheikh Yousef's mentality is not representative of Palestinian's mindset, which explains why their editors have almost ZERO understanding of the Middle East conflict, and I have been reading their editorials for more than 20 years.

12.30.2022 2:43am - (Replying to @AdanZBecerra1 and @stephensenn) I’d be delighted; let’s discuss details by email

12.29.2022 2:56am - The bds has never aimed to boycott israel; its aim has always been to slander and malign - boycott is merely an attention grabbing instrument, and a successful one.

12.29.2022 2:27am - Kudos to Paul, for propagating the science of causality in an era where the scientific method itself is under attack.

12.29.2022 1:22am - (Replying to @RealSarahIdan) @EinatWilf

12.29.2022 12:51am - What uncle Ira tried to tell the Shmegegge was that Palestinian existence is mighty peculiar - no other nation defines itself by negating the existence of another

12.29.2022 12:01am - (Replying to @algobaker and @mrexits) Ill be glad to send one

12.26.2022 9:38pm -

12.25.2022 10:51pm - (Replying to @Claire_V0ltaire) I've posed this question to pro-Pal colleagues. The answer: Goliath was a martyr; greatly disadvantaged by David's youth, light weight, sling training, and disproportionate aggressive reaction. His mistake: speaking his mind in Hebrew, it's undiplomatic and everyone could hear.

12.25.2022 8:30pm - The diet-weight example is described in #Bookofwhy pages 211-7,, and especially here: Senn is trying to make Nedler’s calculus relevant, so he complicates it with "block design" "many dining halls", and more - unnecessarily.

12.25.2022 6:00pm - (Replying to @stephensenn @soboleffspaces and @ADAlthousePhD) If the natural question asked by Fredrick Lord -- "What's the effect of diet on weight gain" -- is deemed a "red herring" unless it conforms to someone's suggested experiment, then the suggested experiment should be modified to capture natural questions, not the other way around

12.25.2022 5:21pm - For the last night of Chanukah, here is a great performance of Hendel's Judeas Maccabeas,vid:HY4D4TECV08 About 25 min into the video you can find the song we sang in Kindergarten (Hava Narima -- Macabee Gibor), the one that surprised my mother in

12.25.2022 1:59pm (1/1) Of course design matters; but design engineered without a "purpose" in mind, is likely to estimate the wrong quantity. In the Lord Paradox (Wainer version, fig 6.9), the purpose is to estimate "Q= the effect of diet on weight gain", a population property, that is defined
12.25.2022 1:59pm (2/2) independently of the design. In fact, the "design" should serve to estimate Q, so, if two different "designs" give rise to two different Q's, something is wrong with one of them. Formally, Q=E(W_f-W_i |do(diet_1))- E(W_f-W_i |do(diet_2)) Simple; no hierarchy, no complication.

12.25.2022 2:55am - (Replying to @stephensenn) Causal Inferencers believe that experimental design is a man made tool devised to interrogate Nature. They can't understand, therefore, why Nature, or at least a model thereof, does not show up the writings of many "designers" who claim to worship "Nature".

12.25.2022 2:40am - (Replying to @SSI_Movement) My duty and my honor, or, as a friend expressed it tonight in our Chanukah party tonight: It's our "Army reserve service (Sherut Miluim)", considering that our academic colleagues in Israel are serving in Jenin. Happy Chanukah.

12.24.2022 6:16pm - But how can we leave Hanukah without our favorite songs (1) Ocho Candelikas ( Ladino and (2) Sevivon Sov Sov Sov (Hebrew, about 11 minutes into this video ) Hallelujah!

12.24.2022 5:17pm - (Replying to @Ostrov_A) Many many more years of happiness

12.24.2022 6:43am - (Replying to @BradSpellberg @DocBio1509 and 5 others) This is precisely where Causal Inference comes it. I tells us: "Your conclusion is as valid as the assumption: "jumping out of an airplane without a parachute increases your chances of death", no more no less." That's why it's so important to make assumptions transparent.

12.24.2022 6:20am - End of year always brings with it a flood of requests for "end of year giving." This year, my choice is made -- I am donating $5,000 to SSI (Students Supporting Israel), on whose Faculty Advisory Board I am proudly serving. Please consider this cause.

12.24.2022 4:49am - (Replying to @Abelaer and @mattecapu) And I took it from Toulmin's Foresight and Understanding: An Enquiry into the Aims of Science (1961) ISBN 0-313-23345-4

12.23.2022 6:09pm - (Replying to @Claire_V0ltaire) Palestinians have a paper title for everyone and historical connection to none; their most ancient national holiday is the Nakba (unless you count Channukah) and their earliest national hero is the Mufti (Haj Amin Al Husseini, 1895-1974)

12.23.2022 2:43pm - Hi everybody, Happy Holidays to all readers. I'll have to take some time off twitter, abandoning even the intense discussion over trialists and CI, Will be back, I hope, with renewed energy to explore new vistas in Causal Inference and Human-level AI. Happy New Year.

12.23.2022 2:34pm - (Replying to @stephensenn @soboleffspaces and @ADAlthousePhD) The Lord Paradox example in #Bookofwhy is about "causal effect" and, if it is a property of the population, then we should be able to discuss it without detailed specification of the data collection process. This is exactly what the example is about, it's about "causal effects".

12.23.2022 2:19pm - (Replying to @stephensenn @ADAlthousePhD and @soboleffspaces) So what is "causal effect"? A property of the population? Or a property of the units in the study? If the former, then we should be able to discuss it without detailed specification of the data collection. If the latter, we have an uncrossable language barrier. Please choose.

12.23.2022 1:46pm - (Replying to @stephensenn @ADAlthousePhD and @soboleffspaces) I've tried, but got stopped by a 5-mile tall language barrier. E.g., what is a "causal effect"; is it a property of an infinite population? or of the finite-sample study? I lost you when you mocked the existence of the former, and haven't recovered yet. What is it?

12.23.2022 1:34pm - (Replying to @soboleffspaces @stephensenn and @ADAlthousePhD) I admit that my opinion of how trialists think is based (partly) on @stephensenn refusal to define "causal effects", or even to use the term in his tweets. Same applies to "estimand".

12.23.2022 1:17pm - (Replying to @stephensenn @ADAlthousePhD and @soboleffspaces) To answer your questions we need to establish some common language. My questions aim to establish it and they haven't been answered thus far. We havn't even established what the "trial objective" is, so I'm waiting. Courage?

12.23.2022 1:11pm - (Replying to @VickersBiostats and @soboleffspaces) I've raised the possibility that trialists and CI folks may in fact be thinking the same way, but we'd never know unless the former tell us what the words "trial objective" and "estimand" mean to them, and how they express them mathematically. Courage?

12.23.2022 7:35am - (Replying to @TenanATC @ADAlthousePhD and @soboleffspaces) Sorry that you do not think you can learn something from my difficulties. I usually learn a lot from readers' difficulties, e.g., I learn to explain myself better, and to encourage more questions. Too bad.

12.23.2022 7:18am - (Replying to @soboleffspaces @stephensenn and @ADAlthousePhD) It's OK to define things differently. What is not OK is to refrain from making your definitions explicit and complain that people do not understand what you had in mind.

12.23.2022 7:11am - (Replying to @stephensenn @ADAlthousePhD and @soboleffspaces) Another "absurd". How about telling us what the words "trial objective" and "estimand" mean to YOU, personally, and how you express them mathematically. Courage?

12.23.2022 7:10am - (Replying to @TenanATC @ADAlthousePhD and @soboleffspaces) Another "absurd". How about telling us what the words "trial objective" and "estimand" mean to YOU, personally, and how you express them mathematically. Courage?

12.23.2022 6:32am - (Replying to @yudapearl @ADAlthousePhD and @soboleffspaces) To all offended trialists. There is no point repeating how stupid I am. Instead, just tell us what "trial objective" and "estimand" mean to YOU, personally, and how you express them mathematically. We may mean the same thing, but we'd never know unless one of you say it. Courage?

12.23.2022 6:02am - (Replying to @ADAlthousePhD @dailyzad and @soboleffspaces) The whole discussion started by two of your peers who expressed objections CI, presumably, because CI is (1) asymptotic and (2) does not do the estimation. Such objections imply "tasks begin with data." If there is acceptance of CI among trialists - I declare myself a trialist.

12.23.2022 5:51am - (Replying to @ADAlthousePhD and @soboleffspaces) You keep on telling me that I am wrong, when I am dying to hear what trialists believe their task begins with. Is it so hard to articulate? One equation should do it.

12.23.2022 5:45am - (Replying to @dailyzad @ADAlthousePhD and @soboleffspaces) The paper you cite is using CI language and methods, eg. DAGs. In this setup, "estimands" are natural. Can you cite a paper authored by anti-CI trialists?

12.23.2022 5:27am - (Replying to @ADAlthousePhD and @soboleffspaces) There is another point here. Trialists do not NEED to think of estimands at all, because there is only one estimand in each trial, and it is not different from the trial objective, since no observational studies are involved. Thus, trialists are lucky, no need to be defensive.

12.23.2022 5:01am - (1/2) (Replying to @ADAlthousePhD and @soboleffspaces) It's deeper than that. I believe one cannot properly define or understand "Trial Objective", "Estimand" and "estimate" unless one adapts the CI language. Having seen objections to such adaptation from your peers, it is only reasonable that I be curious how trialists do it.
12.23.2022 5:13am - (2/2) (Replying to @yudapearl @ADAlthousePhD and @soboleffspaces) I still don't know how, and your slides do not help me b/c, though verbal descriptions appear, formal expressions are lacking. Thus, I cannot tell for example if "estimand" is an asymptotic entity (shunned by your peers) or sample-based entity. My difficulties are sincere.

12.23.2022 4:02am - (Replying to @ADAlthousePhD and @soboleffspaces) Let's not focus on what is wrong with me. Let's hear what estimand means to trialists, how it is defined, how it is expressed mathematically, whether it is ever different from the "query",etc, etc, so that we can understand what makes trialists different from CI folks, if at all.

12.23.2022 3:50am - (Replying to @ADAlthousePhD and @soboleffspaces) If I have made incorrect statements about what trialists do or think, please correct me - what do they do and think? My statements express my best guesses trying to understand what makes trialist so unhappy with CI, where queries, estimands and estimates are so crisply defined.

12.23.2022 2:16am - (1/2) Jeffrey Herf 's letter: is an insightful account of the anatomy of anti-Israel bigotry in academia, and a serious indictment of how Jewish Studies and Israel Studies programs have been hijacked away from their missions. "As far as a know, there has
12.23.2022 2:16am - (2/2) been no DONOR'S REVOLT or even serious probing about the direction of Jewish Studies and Israel Studies programs." Donors' revolt -- anyone knows how to start one? Please pass along Jeff's letter to a donor you know? It's urgent! @Ostrov_A , @StudiesZionist @AdamMilstein

12.23.2022 12:52am - Illuminating video on Spinoza. Although my first intro to #causality was from an entirely different angle,, watching this video now is like talking to an old cousin who used to share your adventures in that same playground.

12.22.2022 7:09pm - (1/2) When I speak at schools or synagogues, kids often ask me: "Are you a Holocaust Survivor?". I say, "No, I am a 'Redemption Survivor' -- a member of an (almost extinct) generation that saw Jewish redemption and re-birth". And when they sing "Dreidel, Dreidel.." I teach them:
12.22.2022 7:09pm - (2/2) "Nes Lo Kara", a song saying: "A miracle did not happen to us; We found no vessel of oil; We carved the rock till we bled; And there was light!" First sung by Zionist pioneers in the 1930's who literally built a nation by carving the rock till they bled. And they prevailed!

12.22.2022 4:04pm - Maryland High School Students Stage Walkout to Protest Antisemitism
12.22.2022 4:04pm - Refreshing to see students taking action where teachers and community leaders slumber into inaction. What worries me though is the standard appeal for "education about the Holocaust" instead of education about Jewish history, Israel, Jewish values and Jewish revival.

12.22.2022 3:21pm - (1/2) I join @AdanZBecerra1 in congratulating @AmandaKMontoya on her Mediation preprint, which should introduce causal mediation analysis to social and personality psychology. I think however that there is no such thing as "statistical mediation". The Baron-Kenny tradition
12.22.2022 3:21pm - (2/2) may appear to be "statistical", because causal assumptions were buried in "regression" equations, and researchers lacked formal definition of "mediated effects", but this has changed in the past 2 decades (#Bookofwhy p.324) and statistics can resign back to estimation tasks.

12.22.2022 2:34pm - (Replying to @imthealan) I would go even further and say that Causal Inference is the ONLY scientific method we have for dealing with interventions and counterfactuals.

12.22.2022 2:30pm - (Replying to @ildiazm @f2harrell and 6 others) Agree. CI deals separately with the two uncertainties, first confounding bias, then sample variability. The anti-CI folks dislike the former, because it's "asymptotic" hence must be avoided. I doubt it can be handled any other way, implying that @f2harell's "proof" doesn't exist.

12.22.2022 10:39am - (Replying to @f2harrell @LarsvanderLaan3 and 6 others) I am not aware of any " near proof of confounder completeness" not using DAGs, and would be mighty surprised if such exists.

12.22.2022 2:34am - The diagram tells which partial correlation gives the causal effect we need, and which doesn't. Indeed, this simple & important question is one that graph-avoiding economists and statisticians cannot answer to this very day. See,

12.22.2022 1:55am - I liked this quote: "People sympathetic to Israel are going to have an increasingly hard time differentiating between kinds of criticism. The way to do so is to ask if a critic is trying to make Israel better, or trying to make it disappear." Perhaps because I smell the latter.

12.22.2022 1:47am - (Replying to @yudapearl @erikbiz and @RabbiWolpe) Forgot the piece:

12.22.2022 1:45am - (Replying to @erikbiz and @RabbiWolpe) Agree! Even wrote an opinion piece against tendencies to obscure the national message of Chanukka with religious fables. Generations of Rabbis had good reasons for doing it - it was safer to play down Jewish national aspirations. Today, it's just a sign of confusion & insecurity.

12.22.2022 12:40am - (Replying to @AvivaKlompas and @MonaAbuamara) Great progress!! The Palestinians have learned to compromise! Last we heard, they were not going to hold an election "until Tel Aviv is liberated". Now it's only Jerusalem -- peace is on its way.

12.22.2022 12:01am - (1/3) Yesterday I lighted the Channukah candles at Chabad House. As the celebration progressed into the evening, dignitaries gave their speeches, light over darkness hailed, rising antisemitism condemned and holocaust survivors honored, I noticed that no one mentioned Israel,
12.22.2022 12:09am - (2/3) how it inspires us, and how its current struggle for survival parallels the Maccabees struggle for independence. I said: "Hannukah without Israel is un-Jewish." But the dignitaries already left. I reached to the piano and sang Hatikva. Luckily a few boys (4-7) joined me,
12.22.2022 12:09am - (3/3) and I now feel super accomplished - we gave the evening the Jewish flavor it deserved: from darkness to redemption.

12.21.2022 11:09pm - Eight Thoughts on Israel’s Political Crisis

12.21.2022 9:20pm - I've been using "data-intensive sciences", to distinguish it from "data science" which, unfortunately, has become "the science of data" instead of what it should be: "the science of the data-generating reality".

12.21.2022 6:17pm - Call me an ethno-centric, but I could not help but think about Chanukah as President Zelensky spoke tonight; how similar that ancient fight for freedom and self-determination was to the fight that Ukrainians fight today. Let's remember Zelensky tonight, as we light the 4th candle

12.21.2022 4:18pm - (Replying to @soboleffspaces) One tiny nuance: "What we are looking for", the query, can also be expressed in causal vocabulary, e.g., E[Y|do(x)]. The "estimand", on the other hand must be expressed in statistical vocabulary, so that even an anti-CI trialist would know what needs to be estimated.

12.21.2022 4:04pm - (Replying to @ggreenwald) Dear Glen, as one of the victims of the Berkeley's Bylaws, let me confirm @Ostrov_A 's allegation that we are facing a clear case of discrimination on the basis of national origin and, contrary to NYT implications, I'm no "conservative commentator" nor anti-Palestinian activist.

12.21.2022 1:36pm - Thank you @soboleffspaces for posting this page from #Bookofwhy. Reading it again, I begin to understand why trialists rarely use the word "estimand" -- they believe their task begins with the data, albeit experimental, not with the query.

12.21.2022 1:20pm - (Replying to @AdanZBecerra1 @kat_hoffman_ and 12 others) Whenever I see an appeal to "expertise or experience" I interpret it to be "informal wisdom" which tells me either (1) We, AI-ers need to capture it algorithmically, or (2) The practitioners are hiding something from science, or both. Which is it?

12.21.2022 9:47am - To further elucidate the concept of "estimand," I would recommend Fig. 1 page 12 in #Bookofwhy, where you can see an entire "inference engine" and how estimands and estimates are produced.

12.21.2022 9:04am - (Replying to @f2harrell @soboleffspaces and 6 others) When you ask a causal question then, in order to convert the question into an estimand (a statistical quantity), you need Causal Inference (CI) to do the conversion. You can do it in your head, as in most RCT studies, but it is still CI. All the estimands I've seen are asymptotic

12.21.2022 9:04am - (Replying to @f2harrell @soboleffspaces and 6 others) When you ask a causal question then, in order to convert the question into an estimand (a statistical quantity), you need Causal Inference (CI) to do the conversion. You can do it in your head, as in most RCT studies, but it is still CI. All the estimands I've seen are asymptotic

12.21.2022 2:11am - It's not Biden alone to blame. Jewish leaders are falling all over thanking him for inviting them to the whitehouse, rather than asking him to be consistent and to backup his beatiful words with meaningful commitments, e.g., to seriously investigate cases such as Berkeley's Zones

12.21.2022 2:01am - (Replying to @tdietterich @amt_shrma and @ChenhaoTan) I'm starting to play with it - very impressive. However, it fails the toy problems I pose to it. It is still illuminating to find out how much children can learn strictly from teachers, without guidance from playful manipulations and without innate templates for causal models

12.21.2022 1:15am - I may not count as a "leading scholar on antisemitism" but I have published substantially on the subject, perhaps more than any of those "signatories". With this in mind, can @PeterBeinart educate us: What is "smearing" in exposing why she does not fit this job's qualifications?

12.20.2022 6:23pm - (Replying to @soboleffspaces @f2harrell and 6 others) I'm also not sure I got it correctly. Guessing: the anti-CI statisticians complain that CI folks dont do the final estimation, something they can do very well. The notion of "estimand" is not in their vocabulary, perhaps because they would need to get it from CI. A puzzle.

12.20.2022 2:38pm - (Replying to @f2harrell @ildiazm and 5 others) As I understand it, @f2harell will never forgive CI folks for letting statisticians do what statisticians do best, namely, take the CI solution (i.e., the estimand) and estimate it from finite data in some specific domain, using the latest available technique.

12.20.2022 7:55am - If I'm not mistaken, this @Columbia is Columbia University, also known as an accredited academic institute. Or am I mistaken?

12.20.2022 4:43am - I should add though that the Bayesian shift from shaky to plausible assumptions is accomplished entirely within the province of probability theory; it lacks therefore the power to deal with causal relationships. (As I lament here

12.20.2022 4:16am - Biden slams ‘venom’ of antisemitism at Hanukkah event, unveils White House menorah via @timesofisrael
12.20.2022 4:16am - Again, Jewish leaders are up the mountain tops praising Biden for general words against antisemitism, instead of telling him what actionable words they would have liked him to say. For example, that anti-Zionism is first degree racism. @Ostrov_A @bandlersbanter

12.20.2022 3:58am - The ‘Jew-Free Zones’ at Berkeley story wasn’t ‘misinformation’ via @JNS_org
12.20.2022 3:58am - It's time for The Forward to apologize for spreading "misinformation" regarding the Berkeley Law School scandal. Especially offensive is its argument # 3, which treats Zionism as a whimsical political opinion. @bandlersbanter @Ostrov_A @Klmarcus

12.20.2022 3:29am - (Replying to @richard_landes and @jsternweiner) If @jsternweiner receives letters from Avi Shlaim he must have received the conditions under which @FranceskAlbs would qualify for the job. Those who deny Israel's right to Tel-Aviv, are hardly in position to judge how to defend that right in Jenin.

12.20.2022 2:51am - (1/6) Readers who continue to be attacked by "RCT Priests" have probably noticed that in "Dialogue with hostile examiner" (p.369) I am claiming that CI is a mapping from shaky assumptions to plausible assumptions. I here wish to exemplify it in two domains:
12.20.2022 2:51am - (2/6) First, consider the philosophy of Bayes analysis. Why should an analyst postulate priors to derive posteriors? Why not postulate posteriors directly? The answer is that the former judgment is more reliable than the latter. In Bayes' example (#Bookofwhy, p.98), judging the
12.20.2022 2:51am - (3/ ) length of the Billiard table is more reliable than judging the distance x' where the next ball will land. The former is based on direct experience in similar casinos, the latter lack such experience. Bayesian logic endows the latter judgment with the reliability of the former.
12.20.2022 2:51am - (4/6) Next consider the problem of recovering from selection bias in RCT's. Recovery methods were developed in and were criticized by Deaton and Cartwright for relying on causal assumptions (DAG), which RCTs aim to avoid. However,
12.20.2022 2:51am - (5/6) there is a fundamental difference between the no-confounding assumptions that RCTs are designed to neutralize and the assumptions upon which the recovery methods rely. Researchers may be totally ignorant of the structure of the former and quite knowledgeable about the
12.20.2022 2:51am - (6/6) process of recruiting subjects for the trial, over which we have some control. Here we see how the ultimate goal of causal inference is achieved: basing conclusions on defensible assumptions of recruitment, rather than indefensible assumptions on how units choose treatments.

12.19.2022 11:42pm - Fortunately, most American Jews know that abandoning Israel, as @Eric_Alteman advocates, is an act of identity-suicide, and they know better.

12.19.2022 10:58pm - (Replying to @HumzahAlkindi) From what I can see, it's a collection of algorithms for doing estimation of causal quantities. I'd question only whether readers would be able to understand the assumptions under which the algorithms would deliver unbiased estimates.

12.19.2022 8:46pm - Hamas calls on Palestinians to defend Al Aqsa against Israeli plans to celebrate "the so called Hanukah Holiday". Quoting: "the so called Hanukah Holiday". I was right in saying that Palestinians fear Hanukah more than they fear the Mossad:

12.19.2022 5:25am - (Replying to @ranilillanjum @ericjdaza and 15 others) Readers curious about the relation between SCM and framework proposed by @ranilillanjum would note that, in SCM, there is only one concept of causation (ie, functional relations), giving each cause a different degree of "sufficiency", "necessity", "directness", "prob. raiser" etc

12.19.2022 1:58am - (1/2) Recent discussions on "observational trials can NEVER establish causality", followed by misunderstandings of what Causal Inference is about, are jolting me to retweet Ch. 11 of Causality where you can find (p.369) "A dialogue with a hostile examiner"
12.19.2022 1:58am - (2/2) Please read carefully, and keep in mind, any resemblance between the actors in the dialogue and participants in our Twitter thread is purely coincidental, though some of the arguments could sound astonishingly similar.

12.18.2022 4:43am - Tonight, as we light the first candle of Chanukah. I am reminded of the 2007 Whitehouse celebration, in which I and my late wife Ruth were invited to light our family Menorah, and of my late son Danny who mentioned his great grandfather in his last words

12.18.2022 4:03am - There is no way for me to start the Holiday without sharing my understanding of why Chanukah is the Jewish "trust deed" to history: @Ostrov_A , @SAFECUNY , @robsatloff

12.17.2022 11:46pm - These notions of soundness and validity govern the logic of Causal Inference, once we replace the classical notion of "world" with a "fully specified SCM model."

12.17.2022 10:05pm - Unfortunately, once you give her an opening to plead innocence of charges she wishes to deny, the whole conversation gets embroiled in endless philosophical sophistry: "Is she or isn't she an antisemite?", and you lost the chance of accusing her of charges she cannot deny.

12.17.2022 7:12pm - (Replying to @maradonasattva and @FranceskAlbs) `Good question, I assume these will have to stand with Atlantis, Unicorns, Squared Circles and other impossible yet wished-for objects.

12.17.2022 4:43am - Academic sophistry disputing what is or isn't antisemitism is another reason we should not accuse @FranceskAlbs of antisemitism, but of much uglier form of racism - Zionophobia - a charge she cannot deny, and which should disqualify her immediately from the UN job she's seeking.

12.17.2022 3:21am - (Replying to @TristanBlumen) Because once you give her an opening to plead innocence of charges she wishes to deny, you lost the chance of accusing her of charges she cannot deny.

12.17.2022 2:45am - I still maintain it's a grave mistake for Jewish leaders to challenge the fitness of Francesca Albanese on the basis of her antisemitic remarks 10 years ago, instead of proving her unqualified on the basis of what she says TODAY -- denying Jewish right to a homeland.

12.16.2022 12:06pm - Thanks for letting me know about your journal club, but the paper you are discussing is misguided, and should always be read together with its compendium, and the references in, and

12.16.2022 2:23am - She has a point. Why, for havens sake, people dig out things she said 10 years ago? She can easily be disqualified by what she is saying today. Just ask her what she thinks of Zionism, or whether Tel Aviv is stolen land.

12.16.2022 2:09am - UK readers, if you wish to see why the Israeli-Palestinian conflict is the simplest conflict in human history, don't miss out on this event with @EinatWilf

12.16.2022 1:58am - (1/2) Regardless of its outcome, the Biden Administration decision to investigate the Berkeley Law School scandal sends a clear message to all US Universities: Beware, there's a new minority group on the block -- Zionist students and faculty -- who is demanding recognition and
12.16.2022 1:58am - (2/2) respect like any other identity-bonded group on campus, and you better get used to it. This new minority ain't going to be pacified by a "task force on antisemitism"; it demands an independent voice on all norms of inclusion or exclusion.

12.15.2022 11:15pm - To be honest, I didn't know a field called "Learning Analysis" (LA) exists, separate from "Education". Now that I know, I am glad they are getting to read this introduction to Causal Inference, to get them going on the right path.

12.15.2022 5:51pm - Reassuring to know that the Office of Civil Rights is launching an investigation into to civil rights violations at UC Berkeley Law School.

12.15.2022 12:38am - Beautiful title: Can the Whole world be Wrong?

12.14.2022 2:45am - (1/2) I am proud to join over 130 Jewish Studies professors and educators who have signed today a statement calling on the United Nations to adopt the IHRA Definition of Antisemitism (IHRA WDA) This is the first time ever that Jewish Studies scholars
12.14.2022 2:45am - (2/2) have organized in opposition to the anti-Zionist ideology that has begun to corrupt Jewish Studies programs nation wide, and has led to an anti-IHRA letter sent last month to the UN by people who claim to represent "Jewish Studies". They don't.

12.13.2022 9:07pm - Agree, 3 is willful. But this still begs the question of why reasonable analysts, sympathetic to Israel, would blind themselves to reality. My answers: (1) is truly irrational, requiring new prism to be believable. (2) breaks the Western ethos of "Smart we can surely fix it".

12.13.2022 6:33pm - The causal inference literature continues to swell with new and interesting results - here is our weekly report: Many titles catch my eyes, like this: which bounds counterfactuals by combining observational and experimental data.

12.13.2022 4:47pm - It was a fairly long lecture and, I agree, it should be useful for curious outsiders asking: what's all this noise about causal inference - what is it?

12.13.2022 4:30pm - Happy Chanukah, Amir; glad a Science Historian of your insight agrees with my excitement on the timeliness of this discovery. What else can we do to add to the holiday spirit? Here it is, a few Chanukah songs, under my baton, by the famous LA Shir Choir:

12.13.2022 2:21pm - (Replying to @kareem_carr) With this flexibility, Statistics is indeed the "science of everything". Why? Because if someone were to tell us "No, statistics is missing XYZ" we would say: "when we say "statistics" we imagine a future where XYZ and statistics are combined into a single venerable discipline"
12.13.2022 2:27pm - (Replying to @yudapearl and @kareem_carr) And the same is true about arithmetic. Arithmetic is the "science of everything" if allowed the flexibility to imagine a future where arithmetic is combined with everything "into a single venerable discipline."

12.13.2022 9:07am - (Replying to @blakeflayton) This rotten habit of embedding antisemitism in the context of other forms of hate is aimed to justify inaction under the cover of endless philosophical debates of hate-related phenomena. UCLA is using this trick too: [link to:]

12.13.2022 8:27am - 2,200-year-old coin hoard gives hard proof of Book of Maccabees, say archaeologists via @timesofisrael
12.13.2022 8:27am - I guess I am, and will remain sentimental; a discovery of this nature feels like someone has found the original copy of my birth certificate. And what a great timing! Just a few days before lighting the 1st candle of Hanukkah (Sunday, Dec. 18). Who said history is about the past?

12.13.2022 1:27am - The debates on who owns whom did not end with Karl Pearson. Some statisticians still think causal inference is a "missing data" problem (e.g., Rubin and Imbens) and some deep learning folks think it's an neural-net optimization problem.

12.12.2022 9:18pm - This is what Karl Pearson thought, when he discovered correlation, "everything is statistics", he said (paraphrased). But Sewall Wright, already in 1921, told him: "No way, Karl, statistics can't answer even half of my scientific questions" #Bookofwhy.

12.12.2022 3:19pm - (Replying to @scwallphd @RWJE_BA and @imthealan) Same with arithmetic.

12.12.2022 12:47pm - Here is your chance to see the 3 major obstacles to peace in the Middle East: (1) Palestinian rejection of any form of a Jewish state, in any borders, (2) Israelis awareness of (1), (3) Western media un-awareness of (1) & (2). Vividly documented here:

12.11.2022 12:45pm - (Replying to @osoleve @DrRAWesterman and @Dominic2306) A better way to phrase it is: My neighbor's roof has been getting wet whenever mine did.

12.11.2022 7:27am - (Replying to @_HassanEldeeb_) It's not nice calling other people "the devil," civilized people do not do that.

12.10.2022 6:33pm - If you are connected in any way to Jewish Studies, or International Law, please consider adding your signature to this important statement: by writing to with your name and academic affiliation. @CotlerWunsh @Ostrov_A @EinatWilf

12.10.2022 3:44pm - Some Retweets require comments, some demand a word of caution. This does not.

12.10.2022 3:29pm - This self-evident item should have been submitted as an "amendment" to the original HR Declaration, if only we had a HR-minded body to submit it to, unmolested by the forces that made Jewish self determination less than self-evident.

12.10.2022 2:53pm - (Replying to @the_heruman) Thanks. I'm impressed again. Lets try asking about responsibility. Who is responsible? The guy who struck the match or the guy in charge of Oxygen in the room? Dead curious!

12.10.2022 2:46pm - (Replying to @DrRAWesterman and @Dominic2306) This is straight propositional logic, which requires no causal inference. I am interested in problems that cannot be handled by standard logic.

12.10.2022 2:40pm - This is fairly impressive for a LLM. I wonder how it would handle counterfactuals like in Firing Squad. Or,in this oldie: “A fire broke out after someone struck a match, ‘What caused it, striking the match or having oxygen in the room?’ See

12.10.2022 12:08pm - It shouldn't be too hard for LLM to add a causal reasoning engine to their system. I suggest they start with an oldie: 1. My neighbor's roof gets wet whenever mine does. 2. My roof gets wet whenever I hose it. 3. Will my neighbor's roof get wet if I hose mine? (Causality p.1).

12.10.2022 11:56am - It's a heart-warming photo, to be standing next to the late Tom Tugend, and next to Susan who wrote this powerful eulogy about Tom:

12.10.2022 10:54am - (Replying to @uncoolbob @DanFmTo and 10 others) Eliminationists use creative ways of defining their neighbors: 1. Putin says Ukrainians are Nazis. 2. China says Taiwanese are imperialists 3. Palestinians say Israelis are "settler-colonialists" 4. Some Western observers fall for it.

12.10.2022 3:54am - President Mahmoud Abbas has very good reasons to be angry at the Balfour Declaration: It embraced Jews and Arabs as equal claimants to the land - an idea he cannot swallow to this very day.

12.9.2022 6:13pm - (Replying to @ruthenian @AbsDelfuego and 10 others) The pen is mightier than the sword. Behind every rifle there is a mind, driven by words, legitimizing the shot, delegitimizing the target.

12.9.2022 5:21pm - I just found out that the original film recording of the UN Vote of November 29 1947 was only recently discovered, and it's up for sale at an auction, initial offering $200,000. We can watch it here:, 5:13 min. into the video.

12.9.2022 1:43pm - (Replying to @neontaster) You are in good company. Mehdi Hasan is a product of @AlJazeera ; they feel insecure outside the echo chamber.

12.9.2022 1:36pm - I'm proud of the Los Angeles community who understood the calling of history and leveraged the 75 anniversary of Resolution 181 to remind the world that Israel is here by historical right, not by favor, nor by might. @IsraelMFA should share this video w/ other communities.

12.9.2022 12:53pm - Tom Tugend, War Veteran and Lifelong Jewish Journalist, has Died at 97 via @jewishjournal
12.9.2022 12:54pm - A true friend, a lover of his people, a fighter for truth and a true mencsh. Tom, I love you. Your memory is already a blessing to us all.

12.9.2022 12:03pm - It's reassuring and humbling to know that, four years after publication, there are still researchers who are discovering the beauty of #Bookofwhy for the first time. This means that AI is not doomed for stagnation; reawakening still lies ahead of us. Enjoy!

12.9.2022 5:28am - The history of Arab's rejectionism, its depth and its totality, are well documented in this article by E. Karsh: It contains some interesting facts that are new to me, like the quotes by Nasser and Abu Iyad.

12.9.2022 4:54am - (Replying to @Aaron37913713 @Claire_V0ltaire and @EinatWilf) I've found it especially interesting for a different reason: How can such a simple and obvious truth escape the eyes of almost ALL Western observers, commentators, and peace makers, including the current administration?

12.9.2022 4:45am - Thanks for having me.

12.8.2022 10:44pm - (Replying to @Raamana_ @GaryMarcus and @ylecun) I would need to learn the rules first.

12.8.2022 10:18pm - It's great they decided to listen to students! Next, I hope, they will listen to faculty!

12.8.2022 10:06pm - Upset by neighbors who vow to exist, International pressure they pledge to resist, Would Mr. Putin please join to assist?

12.8.2022 9:13pm - I lost tract; which is worse, official or unofficial?

12.8.2022 2:52am - (Replying to @Claire_V0ltaire and @TankoZionism) He is dead wrong. The Ethno-genetic claim of Jewish endogeneity did not start in 2022. It was recognized already by Balfour, in 1917 , see

12.8.2022 2:20am - Linda Sarsour, our fake feminist (see is turning a fake Jew-lover. It doesn't occur to her that today's most marginalized community are Zionist Jews, yes, the same good Jews whom she adores and whom she tried to exclude from the feminist movement.

12.8.2022 1:03am - (Replying to @richard_landes) GERSHON BASKIN does not bring even one shred of evidence that Palestinians are now ready to be partners for peace. My litmus test has been, and still is, for Abbas to say the words "Equally Indigenous" (in Arabic, of course).

12.8.2022 12:36am - What Qatar Learned from Hitler’s Olympics via @jewishjournal
12.8.2022 12:36am - As much as we hate to invoke the name "Hitler" in new contexts, the similarity between Qatar's Soccer Games and the Berlin's Olympic Games of 1936 did not escape us as we watched the former. This article provides a comprehensive account of the history of deception using sport.

12.7.2022 11:37pm - "Image and Reality" - The greatest counterfactual reversal of our era.

12.7.2022 11:27pm - Why has "drawing assumptions" become so threatening to scientists lately? Because when assumptions are CAUSAL, they are comprehensible. No one would criticize you for assuming Dirichlet priors; who cares? But try to assume "Y does not affect X" and 10 experts would object.

12.7.2022 11:12pm - (Replying to @yudapearl and @AvivaKlompas) He doesn't even sound Jewish!

12.7.2022 11:04pm - (Replying to @AleksanderMolak and @PHuenermund) This is an excellent paper. I can tell just by seeing its awareness of the Ladder of Causation. I do not agree however with the assumption that our works on Fusion requires data on ALL variables; the Fusion Formula dictates what data is required.

12.7.2022 10:07pm - Herzl Warned Us via @jewishjournal
12.7.2022 10:07pm - Moving from the utopian to the practical, Ben Gurion actully followed Herzl steps and wrote (Nov. 1930): "In the same way that we want the Jewish people to be master of its own affairs,... so, too, we must seek for the Arabs." He would have been branded a "Traitor" by Ben-Gvir.

12.7.2022 9:23pm - I honestly do not understand @FranceskAlbs problem. If she is really concerned about her reputation as a fair-minded human-rights observer, all she has to do is to tell us whether Tel Aviv is stolen land or what kind of land it is!

12.7.2022 4:42pm - (Replying to @Sam_J_Street) This diagram may be helpful for the novice, but it is deceptive, in that it skips an important component in Bayesian analysis: Subjective knowledge. That's why I find the Billiard Table example (from Bayes' original paper) to be a better introduction to the issues involved.

12.7.2022 4:13pm - (Replying to @PHuenermund) I have the feeling that economists are not aware of one result in Given two arbitrary subsets Z and W, when would adjustment for Z be (asymptotically) the same as adjustment for W, namely, E_z P(y|x,z)=E_wP(y|x,w) We derived criteria for this equality.

12.7.2022 3:16pm - (Replying to @_ambodi and @DataSciFact) So many people write that "causal identification is useful" that it's becoming hard to tell who is serious about it and who is just echoing hearsay.

12.7.2022 3:12pm - (Replying to @GaryMarcus @johnsontoddr4 and @TonyD993) Where is the text taken from? Statistics (1962) or Deep Learning (2022) ?

12.7.2022 6:48am - (Replying to @PHuenermund) Why only "potentially suffer"? I've learned that, in observational studies, every method whatsoever requires untested assumptions to deal with unobservables. Is this not a golden rule in economics 101?

12.7.2022 6:19am - (Replying to @chroma_infinity) I am not familiar. Can it be described in Tweetable space, using input-output language?

12.7.2022 2:14am - (Replying to @ya3ili @FranceskAlbs and 2 others) What reputation? @FranceskAlbs has any?

12.7.2022 2:07am - (Replying to @AvivaKlompas) Is this Jesus speaking behind the mask? He doesn't sound Jesus-like to me.

12.7.2022 12:35am - (Replying to @DataSciFact) What brings @DataSciFact to resurrect an ancient quote of Gelman about selection bias? A phenomenon that is now well understood, both statistically and causally; see Has anything new been discovered about it lately?

12.6.2022 11:59pm - (1/2) (Replying to @soboleffspaces) Your "taxonomy of resistance" is interesting, and will be helpful to me as I start writing my comments on Nick's "Pearl before economists: the book of why and empirical economics" However, I think you downplay the critical role of ego's and cults
12.7.2022 12:05am - (2/2) (Replying to @yudapearl and @soboleffspaces) in the process of academic stagnation. I might write about it some day, when I no longer have urgent questions to ask about the science of cause and effect.

12.6.2022 11:21pm - A politician's stature is measured by the character of those whom he/she is afraid of alienating.

12.6.2022 8:17am - (Replying to @Claire_V0ltaire) Those who lack history cling to bracelets, while their neighbors celebrate Hanukah.

12.6.2022 7:12am - (Replying to @Slytherin_Quxxn and @Claire_V0ltaire) Of course you cannot make peace with the Jackals in your Zoo. But you are not talking to them, you are empowering our children, IDF soldiers, who carry an olive branch wrapped around a sword on their hats, and who need to see their neighbors to understand why.

12.6.2022 4:46am - Did you expect #JStreet22 to bridge Muslim and Jews when they are busy bashing Israel? You have to be proud of your heritage before you can respect others'.

12.6.2022 4:27am - Love it, @noatishby ! "This time around, we're not going away, ... Come for Shabbat..." Better yet, come for an Israeli Chanukah.

12.6.2022 3:42am - It used to be: "What do you mean you are a Jew? You don't look Jewish." No more. Today it is: What do you mean you are a Jew? You don't sing 'Hatikvah'". A new 'Jew' is emerging.

12.6.2022 1:06am - (Replying to @fredforthemets and @Claire_V0ltaire) I doubt "making life miserable for the Jews in the Diaspora" is the number one goal of the Palestinians. More likely: making life miserable so as to diminish Jewish support of Israel, erode International support of Israel, weaken Israel's unity and, await the next opportunity

12.6.2022 12:58am - (Replying to @Claire_V0ltaire) Well put, and well documented in Adi Schwartz & @EinatWilf book "The War of Return"

12.6.2022 12:49am - (Replying to @Slytherin_Quxxn and @Claire_V0ltaire) I love this Zoo, but I believe replacing the finger with a flower (or olive branch) would make it go viral. Alternative: a book and a microscope. Great!!

12.5.2022 11:34pm - It's Monday, again -- time to examine causal inference papers published last week: Readers concerned with complexity and scaling up issues, will find "On the Complexity of Counterfactual Reasoning" (by Han etal) to be of value.

12.5.2022 5:24am - Totally agree. The UN marking of May 15 as "Nakba Day" shifts attention from Palestinians grievances over Israel's policies back to the CORE of the conflict: Palestinian eliminationism, both in 1948 and in 2022.

12.5.2022 4:43am - (Replying to @soboleffspaces) Yet mediation is rung-3 while standardization is rung-2.

12.5.2022 4:36am - Too bad I'm not in shape lately. You know what? Who cares, I'll take a chance "Come see me, Kanye!, I may not be able to smash your ugly face, but I bet I can expose your ugly racism"

12.5.2022 3:41am - (Replying to @Foreman1David) Not surprised to see racist organizations decorate themselves under pseudo-scientific titles such as "Center for Economic Research and Social Change". I wonder how economic researchers feel about it.

12.5.2022 3:23am - Hey diddle diddle, The cat and the fiddle, The cow jumped over the moon; "And they killed a small portion of their own people...." Nursery rhyme? No! Sickly mentality of Israel's neighbors.

12.5.2022 3:06am - Likewise, this is laughable from those who refuse to acknowledge their role in blocking European Jewry from escaping the Nazis in 1939, only to call for "a war of extermination and momentous massacre" against their survivors 8 years later (Azzam Pasha, Oct. 11. 1947)

12.5.2022 2:50am - Congratulations to Hakeem Jefferies on his election as House Democratic Leader. Hakeem, to refresh readers memory, was on of the first to discredit @amnesty 's report against Israel, the country I grew up in.

12.5.2022 2:22am - I know some high schools in California that view it as a "cultural difference" that we Westerners should be able to tolerate.

12.5.2022 2:11am - (Replying to @Yair_Rosenberg) It should start at the Holocaust Museums which should allocate floor space to Jewish life and re-birth (i.e., Israel), not only to Jewish victimhood.

12.5.2022 2:04am - Incredible! "Hatikvah" in Bahrain. Now I am willing to bet this peace is permanent. Moreover, when they play it in Ramallah we would know that the Arab-Israeli conflict is over; all the rest, borders, settlements, capitals etc. are secondary.

12.5.2022 1:17am - UNGA call for Israeli-Palestinian peace parley in Moscow
12.5.2022 1:17am - Those who choose a place for a peace parley wish to tell us how serious they are about the word "peace" . They just did, and the place is Moscow.

12.4.2022 3:21pm - Regardless of any criticism you may have of AIPAC, this description of @jstreetdotorg is so accurate, that I could not resist retweeting it. Not a single recognition of Israel’s achievements or value. Not a single embrace of the Israeli people. I would refuse their invitation.

12.4.2022 6:23am - (Replying to @JehadAbusalim) Got news for you, @JehadAbusalim , the inevitable eruption you are so eager to see erupting has erupted already, several times in my memory, and has failed miserably. You know why? Because you are not hiding from Israelis what you intend to do to them once the eruption succeeds.

12.4.2022 2:38am - Brother. As I was strolling down the thread of your followers, I could swear it's been written by Jews, especially the stuff about "they hate us because we managed to survive, overcome their hate and contribute to society". We share this experience, and can learn from each other.

12.4.2022 2:13am - Antisemitism mainstreams as midterms strengthen the ‘Squad’ and moderates. Reports paint grim picture of campus life as celebrities normalize antisemitism and spur debate.
12.4.2022 2:13am - Comprehensive account of campus life during November 2022, by Alex Joffe, for Scholars for Peace in the Middle East. From the US election results, to the Berkeley Law School scandal.

12.4.2022 1:23am - (Replying to @erikbiz and @bobdylan) Holds or not, he should listen to the calling of history, visit Israel, and perform "neighborhood bully" in Park Hayarkon or the Kottel. He won't get another chance to make an inspirational move for his people.

12.4.2022 1:10am - I've just decided to follow @bobdylan , and to patiently await another gem like "The Neighborhood Bully" to sparkle out from his Twitter account. It's worth waiting a century.

12.4.2022 12:47am - I could not rest before finding out where/how they train diplomats to see terror-attacks as "scuffles". I found it! @TWennesland got a Master of Divinity degree in Norway. Wow! When I visited Oslo (2012) people seemed kind and basically rational. What is it, then?

12.4.2022 12:09AM - What a comforting revelation! The United Nations has a Special Coordinator for the Middle East Peace Process; hard to swallow! His name is @TWennesland who is "horrified" by a video (I watched it!) of a knife-carrying Milfeh stubbing an Israeli officer, then shot. "Peace Process"

12.3.2022 11:49pm - (Replying to @AvivaKlompas and @elonmusk) What does he say to his mother when he comes home? "Mom, I had a rough day in the office". Does he think she doesn't know? Do these people have mothers?

12.3.2022 9:59pm - (Replying to @AvivaKlompas and @m7mdkurd) Did you know that the first newspaper in Western Asia was printed in Safed (Galilee), in Hebrew, in 1570. While the first Arabic newspaper in the holly land was printed in Haifa, in 1908. Sometimes dates tell stories.

12.3.2022 10:16pm - (Replying to @yudapearl @AvivaKlompas and @m7mdkurd) Correction. The Safed printing of 1570 was the First Book in Western Asia, not the first newspaper. The first Hebrew newspaper in the holy land was Ha-Levanon, Jerusalem, 1863. What else does @m7mdkurd need to know? That he doesn't know a thing about the history of holy land?

12.3.2022 9:42pm - (Replying to @blakeflayton) The big boys told us the sound was coming from another kind of activity.

12.3.2022 4:34pm - (Replying to @formaformarum @sean_welsh77 and @ErsatzDoctor) Formalization may not do the execution for you, true, but it may help you avoid impossible executions, e.g., that you can't square a circle. Just think how many billions $'s society would have saved had DL gurus been aware that you can't get "understanding" from data alone.

12.3.2022 3:02pm - (Replying to @george_revutsky @NahumLucky and 3 others) You made my day! As I argue here Zionophobia is our only fighting word!

12.3.2022 2:50pm - (Replying to @RobertAWiseman @davidhazony and 3 others) The most comprehensive proof is in Bareinboim etal "On Pearl's Hierarchy..." But Causality chapter 3 has several demonstrations Note that to prove impossibility, all we need is one example of two conflicting answers from same data

12.3.2022 2:07pm - (Replying to @CriticalAI @chrmanning and @GaryMarcus) But while DL is "evolving" we have a mathematical proof that you cannot "evolve" from one rung of the ladder to another unless you have experience of specific type. Who shall we trust, evolution or mathematics?

12.3.2022 1:10pm - (Replying to @NahumLucky @davidhazony and 2 others) The question is not if they are or aren't antisemites. The question is whether it is a better strategy to call them "antisemites" or Zionophobes. The latter title distinguishes them clearly from classical AS, points to their specific disease and calls for specific antibodies.

12.3.2022 8:17am - (Replying to @davidhazony @SAFECUNY and @PeterBeinart) If you label all anti-Zionists "antisemites", you would be suspect of "paranoia", as if you haven't heard of enlightened exceptions like Chomsky and Beinart. If you label anti-Zionists "Zionophobic racists", it would be clear you have taken Chomsky and @PeterBeinart into account.

12.3.2022 7:54am - (Replying to @davidhazony and @SAFECUNY) Anti-Zionism divorced of antisemitism surely exist. Take writers like @PeterBeinart or N. Chomsky, for whom any form of Jewish sovereignty is "subjugation of others" hence contrary to Jewish values. They say they "love" Jews, as long as they are not free. Some Rabbis follow them.

12.2.2022 10:49pm - Abba Eban's quote was made when Homo Sapiens believed the earth is round. Today, the majority of UN members, and certainly the Arab states, believe the earth if in fact flat, so the 30 minority votes obtained yesterday amount to a new scientific revolution. Immense!

12.2.2022 6:47pm - (Replying to @trumanfrancis @soboleffspaces and 6 others) SCM, the bases of causal inference, has a variable called u, standing for "unit", which accounts for the idiosyncratic features of each individual in the population. Depending on the research question asked and the data available, some of those features can be estimated.

12.2.2022 6:40pm - I would even argue that the tendency of treating anti-Zionism as evil ONLY when it smelled like antisemitism, not on its own racist character, has contributed to the "stand-in" phenomenon you are decrying today.

12.2.2022 5:31pm - Kanye Suspended from Twitter After Posting Swastika via @jewishjournal

12.2.2022 5:22pm - Kanye = Gospel artist! Someone's crying, lord, kumbaya Oh lord, kumbaya

12.2.2022 4:34pm - (Replying to @davidhazony and @SAFECUNY) Not so fast, @davidhazony . Suppose "anti-Zionism" was NOT a stand-in for "anti-Jews", would that make it more acceptable? Less racist? Less eliminationist? Less genocidal?

12.2.2022 4:18pm - I'm not surprised at all to see anti-Zionists rushing to side with Kanye. They know very well that Israel is the culmination of Jewish history, and what could be more painful to Jews than mutilating their history? Perhaps even more painful than admiring Hitler's inventions.

12.2.2022 2:47pm - Had a stormy night, wrestling the angel with Jacob, up and down the ladder. Woke up more optimistic than ever, knowing that you and I (and Shahar) will not let this struggle continue old fashion way, but move it assertively and strategically into enemy territory. Anachnu Po.

12.2.2022 7:36am - Yes, @PeterBeinart , the Jewish nation has many enemies, and they come in all shades and colors. This hate-stricken nation will never forget the amount of support you have given to some of its most vicious predators.

12.2.2022 6:04am - (1/2) Readers wrote to me, and I agree, that @harari_yuval 's argument of connecting consciousness with "suffering" is not convincing. He merely noticed that, in humans, the two go together but, in its core, consciousness is a form of awareness, relating an agent to its physical
12.2.2022 6:04am - (2/2) or social environment. This renders consciousness a mechanizable cognitive function, which may or may not be accompanied by emotions of sorts. Bottom line, conscious robots will be given the right to vote, not by "suffering," but by convincing us it would benefit society.

12.1.2022 6:20pm - Happy to report that this pbs story about the Daniel Pearl Magnet School has had a happy ending. The Journalism teacher, Adriana Chavira, has been fully re-instated to her job.

12.1.2022 5:37pm - (Replying to @Yann_Le_Du and @goodside) Where is that from?

12.1.2022 2:32pm - (Replying to @eD__PR) "Suffering" is not hard to emulate on a machine using, say, undesirable signals. Is it "real suffering"? What would make it "real"?, appearance? Some people question whether chicken suffer like us, or a 3 day old fetus, or human beings different from us in appearance.

12.1.2022 2:05pm - (Replying to @jazchaz and @TimesofIsrael) I am willing to accept 'real politic,' Arab pressure, mob hysteria, and other ugly factors of life. What I can't bring myself to understand is how those delegates fly home on vacation and face their children: "Dad, did you read the resolution, before voting?"

12.1.2022 6:01am - UN General Assembly votes in favor of commemorating Palestinian ‘Nakba’ via @timesofisrael

12.1.2022 6:01am - Our answer to this UN vote: Highlights from the ceremony: Story of a Vote - Welcome (CG Newman) Mike Burstein reading Resolution 181 Pearl final remarks

12.1.2022 5:32am - (Replying to @MsWonderHeather) Did you know that Palestinian mentality behooves them to malign Israel 24/7, non stop, non-facts, no respect, no decency, malign, malign, malign?

12.1.2022 5:14am - What makes Micronesia and Denmark different than Yemen and China? Democracy? Human rights? Moral courage? A sense of decency? Respect for history? Resiliency to Arab pressure? Why do some countries have it and others don't?

12.1.2022 4:15am - (1/5) CONSCIOUSNESS: I stumbled on this video by@harari_yuval on consciousness (Con.), which I highly recommend to our philosophically minded readers. First, Harari makes an interesting connection between Con and Suffering, "Con. is the capacity to suffer"
12.1.2022 4:15am - (2/5) "Suffering" on the other hand is not merely feeling pain but the capacity to "reject painful reality". This, clearly, invokes counterfactual notion of "rejection", namely, the sensation that currently felt pain is not inevitable, but could have been avoided/alleviated.
12.1.2022 4:15am - (3/5) At the end of the video, Harari states that AI will never exhibit Con. because Con. has nothing to do with intelligence; it is not a "problem solving" endeavor but a rejection of reality. As evidence, Harari claims that past progress in computers and AI have resulted in zero
12.1.2022 4:15am - (4/5) advance in Con. He then predicts that the future too will see super-intelligent systemd operating with zero Con. In #Bookofwhy I disagree with both claims. Con. is defined as having a blue print of one's software and, in view of progress in multi-agent systems and the
12.1.2022 4:15am - (5/5) algorithmization of counterfactuals, there has been major progress toward Con. systems. Projecting to the future, I see no reason we will not see fully Con. systems by 2050, exhibiting social intelligence, communicating like humans about trust, responsibility and suffering.

11.30.2022 9:02pm - (Replying to @AnoopRKulkarni @KordingLab and 2 others) The reason PO researchers are reluctant to drawing assumptions is that it is hard to defend assumptions written in their language of PO. Try to defend this one: Y_x || X |Z for some set Z. It is much easier to say: "we will assume SUTVA throughout" and everything follows.

11.30.2022 4:19pm - (Replying to @KordingLab @AnoopRKulkarni and 2 others) Culture or no culture, you will become impoverished in no time. If you try to put down the rules of the contest, you will find out why. Eg. How would we describe the domain to the contestants? Natural language? Why don't we start with the firing squad of #Bookofwhy?

11.30.2022 3:36pm - (Replying to @neuro_data @KordingLab and 3 others) It is more than just familiarity with notation. Imbens and Rubin are highly familiar with the PO notation. If it takes them 3 pages to define "confounding", and if they can't tell when 4 PO statements are testable, then there is something else here -- representational power.

11.30.2022 2:34pm - (Replying to @SivanTal and @EinatWilf) It's beyond me how one can claim "connection to reality" and not account for the "bunker to bunker" determination of Israelis to preserve their sovereignty, meaning (1) self-defense and (2) control borders.

11.30.2022 3:27pm - (Replying to @SivanTal and @EinatWilf) I thought we were talking "connection to reality," feasibility etc. But you attach the term "exclusiveness" to the one side who says "let's co-exist" and, like Haj Amin, you call Bnei-Brak (my home town) "stolen land". Sorry, I must mute you. No patience for racists.

11.30.2022 2:24pm - (Replying to @KordingLab @AnoopRKulkarni and 2 others) Now I am really curious. Their skill then is to take a domain expert, discuss her understanding of where confounding might exist (in the REAL world), translate it to a formal model, then answer questions of interest. Exactly, which part in the process they excel and I would fail?

11.30.2022 2:14pm - (Replying to @ASingh21112 and @sbkaufman) I would go even further. If you read Imbens and Rubin (2015) 3-page definition of "confounding" you come to realize that the causal revolution is necessary to the understanding of confounding.

11.30.2022 12:59pm - (Replying to @KordingLab @AnoopRKulkarni and 2 others) I know that the QE crowd enjoy what they are doing. My question: What can I learn from them that will enrich my work? I am sincere in asking this question. I want to tell my students: So far, we learned how to do ABC, but we don't know how to do XYZ. This is where QE comes in...

11.30.2022 12:50pm - (Replying to @KordingLab @AnoopRKulkarni and 2 others) Logical equivalence does not entail representational equivalence. Try to prove that the sum of the angles in a triangle=2PI in Polar coordinate. Here I presented Imbens with 4 PO expressions and asked if they hold, if they are testable etc...Try it!

11.30.2022 9:27am - (Replying to @KordingLab @AnoopRKulkarni and 2 others) It is unfair to treat me as one who refuses to learn from others when I am asking you SERIOUSLY to tell me what I can learn. Is it only "Art of thinking"? Or "art of thinking" in a representation that makes all thinking hard. I am truly asking how to improve my thinking.

11.30.2022 9:19am - (Replying to @KordingLab @AnoopRKulkarni and 2 others) A debate about frameworks is not red herring if in one framework you can do think that you cannot in the other. E.g., find testable implication of your assumptions, E.g., see if your assumptions are redundant, E.g., see if your assumptions are consistent. E.g... E.g. ...

11.30.2022 7:45am - (Replying to @KordingLab @AnoopRKulkarni and 2 others) Can't parse any of your 4 sentences. One at a time please. Start with "potential confounders"; what are they?

11.30.2022 7:42am - (Replying to @SivanTal and @EinatWilf) The reality is that benevolent crocodiles are still crocodiles for the winging swallow; we need to consider the latter, not only how to feed the former.

11.30.2022 3:47am - (Replying to @SivanTal) That's a crocodile's view of "inclusion": X is "inclusive", if X aims to swallow Y. A Palestinian state INSTEAD of Israel, not next to Israel. No, thanks, said the swallow, winging swiftly through the sky; I'd rather be called "exclusive" than go for your "inclusion". @EinatWilf

11.30.2022 3:16am - When you see students inviting their peers on college campuses to join in celebrating a historical event, and giving thanks for their countries' support back in 1947, you can't but appreciate the basic human need for celebration and the social role of collective memories.

11.30.2022 2:44am - I think your students are ready to tackle most problems of causal inference. But I take issue with "missing arrow is unconditional independence," I believe it should be "conditional independence given a backdoor separation set"

11.30.2022 2:39am - Thank you @GeraldNGOM and @NGOmonitor for sifting through the @hrw report and exposing its factual flaws and shady motives.

11.30.2022 1:43am - In other words, we Arabs have no choice but to try to exterminate you. It's a matter of historic pride - which absolves us from responsibility, free will, moral principles and other Western notions. Coffee or tea? Sconce perhaps?

11.30.2022 1:25am - That's another reason I recommend @SSI_Movement for "end of year" giving. I haven't seen other organizations commemorating Nov. 29 - perhaps the most significant event in Jewish history - not @HillelIntl , not @ADL , nor any "Center for Jewish Studies". Any?

11.30.2022 12:44am - Thank you for posting my final remarks at the memorable event yesterday, at the Museum of Tolerance. I like my quote "Israel was not created at the UN - It was already there!"

11.30.2022 12:36am - (Replying to @SivanTal) You can call it Shangri-La, but words connotes ideas. To me, "Israel" connotes inclusion, e.g., "Two states for two peoples," while "Palestine" connotes exclusion, e.g., "From the River to the Sea". You choose.

11.29.2022 5:24pm - I could swear the number of "Giving Tuesday" emails I got this year is a 100-fold increase over that of last year. What's happening? My criterion is simple: An organization that seeks my support must be present on UCLA campus, and show interest in what I've learned past 60 years.

11.29.2022 5:14pm - (Replying to @juliocardenascr) Statisticians are tormented between (1) Strong conviction that causality is the driving force behind many of their thoughts, and (2) The absence of a mathematical language to capture those thoughts. The PO language only convinces them that causality must remain a mystery.

11.29.2022 4:59pm - (Replying to @SivanTal and @Ojdadana) Please do not teach me about the Kielce Pogrom. My mother is from Kielce, and my aunt was there during and after the Pogrom. I have had many conversation with her about her experience.

11.29.2022 12:32pm - A timely reminder of my tweet, on this "Giving Tuesday" when many of us are asking: which organization is most effective in fighting antisemitism on campus. I am contributing to SSI (Students Supporting Israel) @SSI_Movement

11.29.2022 12:22pm - This Black American diplomat is one of the least-known architects of the state of Israel
11.29.2022 12:22pm - Today, November 29, as we celebrate the 75th anniversary of the "forgotten miracle", we should remember the contributions of a black American, Ralph Bunch, to the re-birth of Israel, especially the cease-fire agreement of 1949.

11.29.2022 3:42am - I am examining again the abstracts in this workshop on the philosophy of statistics and find it odd that speakers avoid discussing the role of causality in statistics. This, in my opinion, is statistics' greatest "casualty" in the 21st century.

11.29.2022 12:03am - (1/2) We have received a new weekly list of papers on Causal Inference: From last week's list, I've read and liked "One counterfactual does not make an explanation"
11.29.2022 12:03am - (2/2) which describes experiments on what people regard as "good explanation" when considering counterfactuals.

11.28.2022 9:30pm - It was an honor to participate in this commemorative event today. And to all friends and readers who could not make it, here is a link to the video shown (10 min): which proves Weizmann's point: “No state is served to a nation on a silver platter.” (1947)

11.27.2022 10:20pm - On the counterfactual interpretation of "Giving Thanks", in the context of tomorrow's commemorative event Does it signal doubts in one's right to exist?

11.27.2022 2:17am - (Replying to @AleksanderMolak) Your tweet re-kindles my belief in the usefulness of Twitter. Some day Econometrica will open its pages to open discussions of these foundational issues, and perhaps others:

11.26.2022 5:53pm - Apropos, I find it hard to understand how economists can reject a framework that embraces, illuminates and generalizes the one they prefer. Teenage rebellion? #Econometric @causalin @PHuenermund @VC31415

11.26.2022 4:08pm - Glad the book is back on the shelf, last time I was at Ben Gurion airport they didn't have it. Frankly, it's the best "do it yourself" instruction manual for Kanye West and others who are trying to be the "true Jews"

11.26.2022 3:51pm - I'll be delighted to comment on Nick's review of #Bookofwhy as soon as I climb the paywall. In the meantime please re-read my comments on Imbens, and my monolog/debate with Heckman and Rubin:

11.26.2022 1:35pm - Please note the the event takes place on Monday, November 28 - not 29 The flyer is correct:

11.26.2022 12:03pm - As a student of cause and effect (#Bookofwhy), and a lifetime observer of the conflict I could not agree more with the analysis of @EinatWilf . The evidence is overwhelming and the logic is unassailable.

11.26.2022 1:19am - (Replying to @prem_k) In 1-2 years, we are going to have platforms that allow a physician to ask any question (in Nat. Lang.), insert any data she has, and get the answer in the form of annotated figures or formulas. We just need to transfer 0.0001% of the resources now going to data-centric ML.

11.26.2022 1:11am - Kielce, Poland, July 4th, 1946 My aunt was there, to comfort the survivors. President Truman heard of this massacre and ordered his generals not to impede with the movement of Jewish refugees on their way to Italian ports. Destination: UK Blockaded Israel.

11.26.2022 12:51am - My salute of the week. From the ashes of despair to a life of hope and meaning.

11.25.2022 10:07pm - (1/3) The story about my science teacher, Dr. Feuchtwanger, is not a legend, but honest truth. He also taught Daniel Kahneman who was 1yr above me in high school. For seventy years I knew that he was right and that some logic must show him right. What
11.25.2022 10:07pm - (2/3) the #Bookofwhy shows us that, from a policy maker viewpoint, the vaccine saves lives on average. Yesterday, however, it downed on me that we can also answer an individual patient's question "Will the vaccine help me?", using P(Bennefit) as defined here
11.25.2022 10:07pm - (3/3) I am excited to write it down soon, so as to tell Dr. Feuchtwanger: "Eureka, we finally have the logic you promised us 70 yrs ago. We can now tell each patient how likely it is for her to be harmed (or saved) by the vaccine." I wish he could be there to marvel with us.

11.25.2022 5:02pm - If you are asking about the #Bookofwhy, the answer is definitely Yes. I've already commended the narrator for doing a better job than I thought possible. It sounds like an Agatha Christie thriller.

11.25.2022 4:44pm - (Replying to @maths_gee) Curious. Where/how did you get this inner belief?

11.25.2022 4:29pm - Readers of #Bookofwhy asked why I used the smallpox story in Fig. 1.7, not Covid-19. Ans. I have a soft spot for smallpox, b/c it was my first hint of the dark side of causation (see 75 years later, we found the solution here

11.25.2022 1:43pm - Recall what Herzl said in the First Zionist Congress (1897): "Zionism is a homecoming to the Jewish fold even before it becomes a homecoming to the Jewish land." See @newzionists

11.25.2022 9:39am - This is indeed what we do! In the Jewish traditions we lament the destruction of Jerusalem by narrating, again and again, how beautiful a city Jerusalem was before its destruction. It inspires hope and rallies people for action and change, b/c "you deserve better"

11.25.2022 1:31am - (Replying to @naivebaesian) It's a good book. Chapter 4, on counterfactuals, is relevant to our discussion on "thanksgiving".See corrected version:

11.25.2022 1:25am - It's refreshing to see so many "professors" competing to be the "true Jews." Thinking: Perhaps we do have something valuable to offer mankind? And perhaps doing it under false identity helps us do it well.

11.25.2022 12:58am - Can the savage murder of a young (18) man galvanize a united front against savagery?

11.25.2022 12:24am - The LA Times, again, did not get it right (purposely, I surmise). The complaint is NOT against "refusing to invite Zionist speakers" but against writing it into their Bylaws. A restaurant can refuse service to any disagreeable customer, but it can't hang a sign "no blacks served"

11.24.2022 9:28pm - (Replying to @ColocateDisrupt) To us, AI researchers, the Substance of the Thing lies in our ability to emulate the Thing on a digital machine. So don't be angry.

11.24.2022 8:54pm - (Replying to @Jay5w) You just used the word "better", which invokes a comparison, albeit implicitly.

11.24.2022 8:36pm - (Replying to @ColocateDisrupt) Far easier, indeed. I used to prey to that Being every morning. But to those of us who are also interested in the cognitive processes involved in "thanking" and the computational resources needed for such processes, the theory of counterfactuals offers some insights.

11.24.2022 8:28pm - (1/ ) As we celebrate Thanksgiving, I am reminded that "thanking" is purely counterfactual concept; you don't thank anyone unless you can imagine an alternative world, significantly worse than ours. While such worlds are sometimes hard to imagine, the mere sound of the word "thanks"
11.24.2022 8:28pm - (2/2) evokes the existence of those worlds and puts a smile on your face: Yes, things could be so much worse. Now I understand why we celebrate Thanksgiving Day -- we celebrate the capacity of our software to invoke counterfactual logic and smile in the face of adversity.

11.24.2022 5:18pm - Speaking counterfactually about "Thanksgiving", we should not forget the "Jewish thanksgiving day", and the incredible thanksgiving project that the city of Netanya has launched: just to offer its thanks to the Goddess of history.

11.24.2022 5:03pm - The logical absurdity of anti-coexistence philosophers demands a humorous exposure: I choke you to death. You scream: 'I can't breath!' I say: You see, I told you, co-existence doesn't work, now you must cease to exist! Causality 101. @harari_yuval

11.24.2022 4:24pm - (Replying to @YKampf @EinatWilf and @YUVAL_YUVAL) As a student of causality, I would first examine whether the perceived weakening of Israel's democracy is not a temporary byproduct of rising Palestinian eliminationism. Or, as the poet said: "Give us one day of normalcy, and we'll show what democracy looks like." @harari_yuval

11.24.2022 3:30pm - UC Berkeley student groups' refusal to invite Zionist speakers draws civil rights complaint
11.24.2022 3:30pm - The Berkeley lawsuit comes to the LA Times, and Dean Chemerinsky continues to talk "religion" and "viewpoints," while students are screaming: "identity"! Is it tone deafness? Selective hearing? A new style of leadership? @Ostrov_A , @JGreenblattADL

11.24.2022 3:10pm - Correcting my error: It's @harari_yuval that is profoundly wrong, not @yuval_yuval which does not exist. And I join @EinatWilf in urging @harari_yuval to analyze the prevailing attitude in the Palestinian society towards the dream we once called "co-existence".

11.24.2022 2:49pm - The jackal, so the legend goes, is the first to howl when its victim dies.

11.24.2022 2:41pm - (Replying to @EinatWilf) @Yuval_yuval is somewhat right when we superficially consider naked facts on the ground; he is profoundly and scientifically wrong in making it sound like it is Israeli's choice. The 3-classes are simply: the safe, the dangerous, and the redeemably dangerous.

11.24.2022 2:21pm - The counterfactual "Thanksgiving" is both an affirmation that "things could have been worse" and an attitude to "view the ordinary as a miracle." Do we want our robots to thank us for the ordinary? Too bothersome. But we certainly won't like them to take things for granted.

11.23.2022 10:50pm - Extremely delighted to see a contest on SCM implementation in #python. Can we have more details, e.g., what the research question was, or the book's Table of Content .

11.23.2022 10:42pm - Why we filed a civil rights claim against UC Berkeley Law School | Opinion
11.23.2022 10:42pm - Good question: Why did you file a Civil Rights Claim against UC Berkeley Law? Isn't it a model campus with a platform for every thinkable "just cause"?? (With perhaps one negligible exception?)

11.23.2022 9:49pm - Mother Theresa has struggled all her life to deserve the level of Sainthood and purity that Palestinians have achieved with ordinary bombs, stuffed with nails, painted "Justice in Palestine."

11.23.2022 3:01pm - It seems that Ambassador @DTzantchevdid did not get my Tweet. "Only people @DTzantchev follows or mentioned can reply." Can someone convey my message to him (below)? Else he would never know "what the Palestinians really want". Thanks

11.23.2022 2:47am - He was only 16, and left us to ask the hard questions: When has our mind accepted the banalization of evil as a feasible process? Can we tell his parents that human mind is wired that way? HaShem Yinkom Damo,

11.23.2022 1:45am - I am checking the condemnations - you're right, the word "Palestinian" is not mentioned next to the word "terror". It's the victory of terrorism tp watch Western media play the game of symmetry, pretending to see everything but the perpetrators,

11.23.2022 1:26am - I wish I could be there with you, to celebrate today's Sigd. "Blessed Him, who gathers even the most abandoned parts of our people." I used to pray it every morning, thanks for rekindling the words in my mind, and for making them relevant to our day and age. Baruch!

11.22.2022 11:34pm - The esteemed and horrified Ambassador should take 30 minutes to examine how the European Union contributes (unwittingly, of course) to this terror. It's only 30 minutes:

11.22.2022 11:20pm - (Replying to @JonathanSKatz @SAFECUNY and @CUNY_Prof) The world likes Jews who appreciate the protection they are getting, not those who insist on making such protection unnecessary.

11.22.2022 2:30pm - (Replying to @hail2pitt01) These guys truly believe that Mehdi Hasan and other Al-Jazeera graduates should succeed in giving them an air-tight civilized cover. After all, some Western organizations have begun calling Al Jazeera cronies "journalists".

11.22.2022 1:37pm - Qatar must be having second thoughts on the wisdom of hosting the FIFA games, allowing the whole world discover what a rotten sewage Al-Jazeera has been trying to conceal all these years under its "news outlet" title and its British accented anchors. I've seen Doha 2012; stenchy.

11.21.2022 11:45pm - Friends ask: Why is it hard for Westerners to understand what's going on in the Israeli-Palestinian conflict? My answer: Because for Westerners, "understanding" entails SOME LOGIC while, in this conflict, some players defies ALL LOGIC.

11.21.2022 9:25pm - This Title IV Complaint against @BerkeleyLaw 's bylaws, aside from protecting Jewish students from discrimination, will have another impact on campus life, perhaps more profound, to render anti-Zionism incompatible with societal values.

11.21.2022 7:27pm - Scary to realize they already have more than 7 followers.

11.21.2022 7:24pm - Glad they placed your "La Guerra del Retorno" next to "Palestina," to maintain some commensensical balance.

11.21.2022 6:51pm - Would the "real Jews" step up and replace them!

11.21.2022 6:22pm - (Replying to @DavidLobron) Yes, all you need to do is ask Google Scholar to send you an alarm for specific citations. I use Causality (2009) and @Bookofwhy (2018).

11.21.2022 5:59pm - (1/2) If in LA next Monday, join me in commemorating the 75th anniversary of the historical UN Vote of November 29, 1947. The event will take place November 28 (not 29), 11:30 am, at the Los Angeles Museum of Tolerance. See flyer: The program will include
11.21.2022 5:59pm - a re-enactment of the vote, videos and testimonies from the period, school choirs and more. I will provide A Witness's Reflections. For background, see The Forgotten Miracle: Nov. 29, 1947

11.21.2022 1:32pm - We are ready again to scan this week's yield of Causal Inference papers: I can spot already a few titles which should invite our curiosity, and I will come back to them during the week, as soon as deadlines and alligators are pacified.

11.20.2022 5:51pm - Mehdi Hasan's logic eclipses only his mentality. He just discovered a couple of Qatari mosques that do NOT teach hate. Even Al Jazeera program "Sharia On Line" is no loner hosted by Al Qaradawi -- Lay off! Qatar is no longer the hate capital of the world, there are others!

11.20.2022 3:10am - A clash of two "rights" does not make one of them "wrong". See how the notion of "equally indigenous" is defined in the post-colonial era:

11.20.2022 1:33am - We now received a link to the original @CWRUPresKaler statement of Nov. 9 and, as expected, President Kaler is being attacked by @CAIRNational and other Zionophobes. He deserves a broad community support at

11.20.2022 12:49am - (Replying to @jeremy1174 and @EinatWilf) You just defined the essence of the conflict in no ambiguous terms: "The Palestinians are happy to coexist with Israelis - just not in Palestine, which they consider solely theirs, all of it, to the last inch, to the last man!" Coexistence Palestinian style and their tragedy.

11.20.2022 12:35am - (1/2) @CWRUPresKaler , President of Case Reserve, belongs to a handful of academic leaders who know the effective language with which to curb BDS hate. "A vote for this [BDS] resolution," he wrote on Nov. 11, "is clearly a vote against Israel and an aggression toward the
11.20.2022 12:35am - (2/2) Jewish members our community.” What sets this statement apart from so many others is: (1) It does not propose a "philosophical study of antisemitism," and (2) It explicitly takes "a vote against Israel" to be a moral depravity, incompatible with community values. Hats off!!

11.19.2022 11:45pm - (1/2) Readers asking to see formal treatments of the Firing Squad and Oxygen-Match stories, can find them here (Figs. 7, 9) and here The reason I take them as litmus tests of "understanding" is: (1) Every 10-year old would vouch
11.19.2022 11:45pm - (2/2) for the plausibility of the answers derived, and (2) I am not aware of machines capable of generating these answers, except those programmed with the counterfactual logic shown. Curious, can someone run LLM on the Firing Squad story? @sejnowski @GaryMarcus

11.19.2022 7:23pm - (1/3) We are in receipt of a link to an illuminating symposia of Anthroponeny: About -1:52 into the video, we can find Terry Sejnowski's talk which asks whether ML, Neural Nets, and Large Language Models can understand "intelligence", "explainability",
11.19.2022 7:23pm - (2/3) and "consciousness". What readers on this educational channel would find weird is the absence of formal definitions, and mini-Touring- tests of "understanding" as provided for example in #Bookofwhy and exemplified by the firing- squad or the Oxygen-Match stories. I remain
11.19.2022 7:23pm - (3/3) unconvinced that ill-understood large models can provide us with a better laboratory for testing computational theories of cognition than well-understood toy problems. @sejnowski @GaryMarcus @ylecun @TrendsCognSci

11.19.2022 7:22pm - (Replying to @jeremy1174 and @EinatWilf) If you do not Al Jazeera, where did you get this maligned understanding of the "Zionist plan." My understanding of the plan is "Co-existence" with only ONE condition: "Existence". This entails no "buckling under" ecept dismantling the current Palestinian plan of "Sole-existence".

11.19.2022 3:15pm - (1/2) A groundbreaking and well documented report on campus assaults on Jewish identity: How do I know it hits the right issues? Pure statistics: The word "anti-Zionism" appears 4 times more often than the word "antisemitism"; a never-failing litmus test.
11.19.2022 3:15pm - This 4:1 ratio is at least reversed in all previous reports by campus watchers (i.e., #ADL, AJC, Hillel, etc). The ratio approaches ZERO in ALL reports issued by University administrations or EDI offices (including UCLA). Hail to statistics! One thing it's good for.

11.19.2022 2:40pm - For our German readers: A "must read" for everyone who is really interested in a solution to to the Israeli-Palestinian conflict.

11.19.2022 1:25pm - Well put! Plus note the strange correlation between those who oppose IHRA and those who deny Israel's right to exist. By endorsing the former, Stewart emboldens the latter.

11.19.2022 12:39pm - There is nothing like waking up Saturday morning to news about PEACE - the first Shi'ite Muslim country to have an Embassy in Israel. You feel like singing something to all the doomsday prophets, but the song is yet to be written.

11.19.2022 1:11pm - (Replying to @haig) Show me one "peace agreement" in the world that was not signed out of some practical "incentives" to the sound of warnings: "don't be naïve!". On Nov. 19, 1977, 45 years ago today, President Sadat landed in Israel to the sound of similar warnings. Yet it's worked, & More to come!

11.19.2022 12:57pm - Will President Abbas ever rise to Sadat's height and address the Israeli Knesset with a message of peace? In 2008 I was so naïve as to write the speech for him, a speech that would convince Israelis that he means what he says.

11.19.2022 2:58am - (Replying to @ShreeParadkar) Your piece implies that @TorontoStar 's readers are too feeble minded to distinguish “Denying the Jewish people their right to self-determination" from criticizing policies of Israeli government. Invariably, those who oppose IRHA are also guilty of the former. Not a coincidence!

11.18.2022 6:59pm - (Replying to @PHuenermund) Who left?

11.18.2022 3:55pm - Honored to join and support Moment Magazine's gala, noting that Moment's Daniel Pearl Investigative Journalism Initiative is launching new projects. dpiji story "An Inconvenient Genocide" won two first place journalistic awards.

11.18.2022 3:48pm - (Replying to @jeremy1174 and @EinatWilf) You can evaluate the mentality of Al Jazeera through your words and beliefs, according to which blood-thirsty IDF is shooting children for sport. And they have succeeded to convince even people of your intelligence that this is journalism.

11.18.2022 9:03am - Police said “four shots were fired from a loaded weapon” into the exterior of the home next to the city’s Old Synagogue but that no one was injured.

11.18.2022 8:55am - (Replying to @jeremy1174 and @EinatWilf) Often false and sometimes merely distorted. You would never see an Israeli child on Al-Jazeera -- where have all the mothers gone?

11.17.2022 9:19pm - (Replying to @soboleffspaces) Amazing, isn't it? Now imagine my child-like glee upon discovering it in 1999 (See Theorem 4 here: I felt like I solved Abraham's problem "What if there were 50 righteous people in the city?"

11.17.2022 4:03pm - (Replying to @soboleffspaces) Moreover, under monotonicity PN is identifiable, and so are PNS and PS.

11.17.2022 4:01pm - (Replying to @jeremy1174) No. If you sift long enough through Al-Jazeera's texts, you are likely to find a fact or two that are untainted by hate and deceit, but who has the time. This is what I wrote about Al-Jazeera's tactics in 2008 - they haven't changed since. @EinatWilf

11.17.2022 3:48pm - (Replying to @Mahamoudmusa1) I have "The War of Return" with me on Kindle. Can you point me to ONE fact or ONE figure that you believe is fictional? Warning: I know my facts and figures, and I am also an eyewitness to some.7

11.17.2022 3:44pm - (Replying to @gorby_gorbachev @foodsevelt and 3 others) I am missing the mathematician's point. Is it worth un-missing?

11.17.2022 3:05pm - I believe your formula is valid only under monotonicity. How do I know? We know that PN is rung-3, while the formula you gave is rung-2. This is only possible under some Rung-3 assumptions. Monotonicity is such an assumption.

11.17.2022 1:50pm - How do we get Jon Stewart to read "The War of Return"? Ans. Don't call him "antisemite", call him an "innocent Al-Jazeera reader."

11.17.2022 6:33am - (Replying to @lucasbergkamp) Invariably so. Non-causal problems are hardly perceived as "problems".

11.16.2022 11:23pm - (Replying to @AvivaKlompas) Worse yet; rumor has it that Israelis are still breathing oxygen to sustain their lives and, thus prolong the colonialization of Palestine, especially in the colonies of Tel Aviv and Ramat Gan. Nasty "Oxy-Washing"! "Weaponize Nitrogen"!

11.16.2022 10:20pm - STAN GREER: A Major Court Case Could Decide Just How Much Power Union Bosses Have Over Employees. Here’s How via @dailycaller
11.16.2022 10:20pm - An interesting court case affecting academia and beyond. I think the 6 CUNY professors have a point: If a given Union is spending its membership dues to malign symbols of identity of its members, the impacted members should have the option to obtain another union representation.

11.16.2022 9:29pm - (1/2) The word "parsimonious" was given to me by a leader in the SEM industry, when I asked: If SEM is but a statistical model of the data, why not estimate just any parametric model of the distribution? He answered: "SEM is both parsimonious and meaningful". The fact that both
11.16.2022 9:29pm - (2/2) qualities emanate from the causal nature of SEM escaped him, as it escape several chapters in Hoyt's Handbook Meaningfulness, in particular, is the transparency of causal models, often discussed on our channel, and fiercely resisted by PO folks.

11.16.2022 4:30pm - This handbook (and my chapter in it) reminds me of my decades-long struggles with SEM practitioners concerning the meaning of SEM. Watch how it started: I suspect such nasty examiners are still ruling SEM education.

11.16.2022 12:31pm - (Replying to @pmddomingos @KordingLab and 8 others) We don't even need to start with the overly complex infant's mind, because it's very easy to learn from data (ameba +evolution). I have examined this Tabula Rasa (model-free) paradigm here, and I am not too enthusiastic about its promises.

11.16.2022 12:14pm - We are in receipt of a free copy of "Handbook of Structural Equation Modeling" (2nd Ed) I have a chapter (3) there, which is trying to convince SEM researchers that, contrary to tradition, SEM's are not parsimonious statistical models but causal models.

11.16.2022 12:05pm - (Replying to @yudapearl @KordingLab and 9 others) How I love this answer!

11.15.2022 10:28pm - Reminds me how the media condemned "The Cycle of Violence" when my son was murdered. @TonyBlinken

11.15.2022 8:39pm - (1/2) We are informed of a new workshop in the philosophy of statistics: Searching the abstracts, I've found one reference to Causation, by Jon Williamson, who makes an astounding discovery: "Causal inference is not statistical inference". Needless to say
11.15.2022 8:39pm - (2/2) I was disappointed to find Williamson classifying graphical causal models as "purely statistical method" and arguing that "this is a mistake." As an alternative, he proposes "Evidential Pluralism," which sounds like SCM to me. What they want to know?

11.15.2022 7:28pm - (Replying to @Claire_V0ltaire) When Muhammad says "According to many reports" she is drawing on Yasser Arafat's technique of saying "Everyone knows that.." (e.g., to Clinton in Camp David: "Everyone knows the Jews never had a Temple in Jerusalem"). It works when your lie doesn't sound too credible.

11.15.2022 7:08pm - Jewish-power theories may have hidden virtues. Recently unearthed records of interviews with British officials reveal that it was fears of "international Jewish power" that jolted the British Government to issue the Balfour Declaration; the rest is history

11.15.2022 3:44am - 1947 U.N. Partition Resolution Memorial to be dedicated Nov. 29 via @JNS_org
11.15.2022 3:44am - While LA is planning to commemorate the 75th anniversary of this historical moment in song, dance and re-enacting the vote, the city of Netanya has erected a monument, to be seen by 200,000 people each day. A tribute to history!

11.15.2022 3:13am - I am sure readers are eager to view this week's harvest of new causal Inference papers. Here they are : I haven't had a chance to search for breakthroughs, perhaps our readers can alert us of any.

11.15.2022 2:10am - An honest NYT account by @blakeflayton on being a progressive Zionist in 2022. The reason Americans fall for the lies they are being fed is that they are rational; too rational to explain the irrationality of a lie-producing machine.

11.14.2022 10:08pm - Congratulations to Naftali Weinberger for receiving the Philosophy of Science and Race Prize, and for introducing Causal Analysis to the analysis of Racial Discrimination. An author tagged @DAGophile should get an extra prize just for the courage.

11.14.2022 7:23pm - (Replying to @pmddomingos @KordingLab and 2 others) The answer is, almost no benefit. Causal theories to date have been viewed as orthogonal to inductive biases; the latter dealing with fitting functions (eg, how to avoid overfitting), while the former dealing with the question of what functions need fitting.

11.14.2022 1:46pm - (Replying to @pmddomingos @RasulElon and 2 others) Agree. Now, causal theories have developed tools for dealing with this incompleteness (ie, missing actions). I am not familiar with such tools deployed in the "inductive bias theories." Are there?

11.14.2022 1:37pm - (Replying to @pmddomingos @KordingLab and 2 others) Causal theories are a type of inductive bias only in the very weak sense that both are needed on top of the data. But to say the XYZ is a type of ZYX is only helpful if XYZ can benefit from the tools developed for ZYX. Here the two diverge, with almost zero intersection.

11.14.2022 11:52pm - (1/2) Speaking of "infinite data" one must distinguish "infinite sample on a given subset of variables" from "All subsets of variables". Both are unavailable yet representable. The former by a distribution function, the latter by a causal model in which assumptions are made about
11.14.2022 11:52pm - (2/2) variables not present in a training set. E.g., Observational studies can be viewed as leveraging assumptions about action variables not available in the data. Another example is "transportability" where experimental data in one environment are missing.

11.14.2022 5:44am - (Replying to @RabbiWolpe) That's why I much prefer the more powerful and less misconstrued accusation: Zionophobia. I haven't found many followers among Jewish leaders who either use the antisemitic card, or warn against overusing it.

11.13.2022 11:17pm - Readers wondering why I prefer toy examples to abstract discussions would be convinced, I hope, by this discussion on what "inductive bias" is, and whether an "inductively biased" machine can properly answer counterfactual questions in the firing-squad story of #Bookofwhy p.42.

11.13.2022 11:03pm - (Replying to @pmddomingos @KordingLab and 8 others) You can see now why I prefer toy examples to fancy words. Students of causal inference know that ANY DATA can equally fit another firing-squad story, in which Captain listens to Marksman-1 and Court listens to Captain's order. Here, counterfactuals get different truth values.

11.13.2022 10:38pm - Excellent story for the next edition of #Bookofwhy.

11.13.2022 12:35pm - It took almost a century for statistics to snap out of "Causation is just a species of correlation" (Pearson) to "Correlation is not causation". I hope it takes less for Machine Learning folks to snap out of the "inductive bias" mantra and examine some causal inference tools.

11.13.2022 7:14pm - (Replying to @ChristianBSmar1) What holds back many ML folks is this mantra of "Causation is just a species of 'inductive bias'", so, there is no need to worry, it would pop out automatically from more of the same, more data, and more talk about power of inductive bias to turn lead into gold.

11.13.2022 5:15pm - (Replying to @jwbelmon) What's #Fediverse or #Mastodon? I hope they are not buried in "inductive bias".

11.13.2022 5:13pm - (Replying to @pmddomingos @KordingLab and 8 others) #Bookofwhy p. 42 "The data will consist of two kinds of events: either all five variables are true, or all of them are false... There is no way that this kind of data can help predict the results of persuading marksman A not to shoot." I may come back to this thread in 2-3 days.

11.13.2022 4:54pm - (Replying to @pmddomingos @KordingLab and 8 others) If you throw causal theories into the "inductive bias" salad, you obscure it from seeing the light of day, and you make it harder for ML folks to ever distinguish what works from what doesn't. I advise readers to stay away this obstruction.

11.13.2022 3:39pm - (Replying to @KordingLab @pmddomingos and 8 others) Yes. Submit the Firing Squad story to an "inductive bias" machine and ask it some questions, like "would the prisoner still be dead had soldier-1 not shot?" etc. etc.

11.13.2022 3:34pm - (Replying to @KordingLab @pmddomingos and 8 others) If I understand your question properly you are asking whether we should understand what wea are doing. Ans. Yes.

11.13.2022 3:22pm - (Replying to @pmddomingos @KordingLab and 8 others) The apparent speed with which you go very far with inductive biases is a result of traveling in a flat terrain of Rung-1 curve fitting, far from the bumpy terrains of Rung-1&2 tasks, where inductive biases are of zero help, unless you call everything that works "inductive bias".

11.13.2022 12:58pm - (Replying to @PhilDawid @pmddomingos and 5 others) I agree with everything you say except the last, which I would reverse to read: DT is not the most suitable language for formulating and conducing CI, since both its basic assumptions and its methods of analysis are better represented in SCM (or CBN).

11.13.2022 12:48pm - (Replying to @KordingLab @scac1041 and 8 others) You can call the rules of arithmetic and the whole scientific method "inductive bias" -- no objection, as long as you don't expect significant scientific problems to be solved by the broad brushstroke methods of "inductive bias" used so far in ML.

11.13.2022 11:52pm - (Replying to @PhilDawid @pmddomingos and 5 others) I agree with everything you say except the last, which I would reverse to read: "causal models are needed to justify the incomprehensible "weak ignorability" judgments.

11.13.2022 11:50am - Summarizing what I've learned in the past two days about "inductive bias". The "prior information" needed is, of course, assumptions about the causal relationships involved. These assumptions are problem-specific and, hence, could not be encoded as "inductive bias"

11.13.2022 11:38am - (Replying to @ylecun @KevinIndrebo and 10 others) In addition to Inductive bias, and its "no-free-lunch" theorems, which lift us from finite samples to distributions, we also need prior information to properly translate our research questions into the language of distributions - a non-trivial task, not addressed by induction.

11.13.2022 11:08am - To our Spanish readers: This is your best introduction to the Arab-Israeli conflict -- well researched, well documented and, most importantly, well reasoned.

11.13.2022 4:37am - Finally, a crisp characterization of anti-Zionists, free from antisemitic ambiguities. Valid even if you are fighting antisemitism tooth and nail, or visit a Holocaust museum twice a week.

11.13.2022 1:42am - (Replying to @pmddomingos @ylecun and 11 others) Data "can" tell us everything when available. The role of causality is to encode assumptions that would make up for data unavailable, eg, data from actions and experiments. I would hate calling such assumptions "inductive bias" lest folks would think I'm a "catch-all evader".

11.13.2022 12:46am - In commemoration of the 2nd Yahrzeit of Rabbi Lord Jonathan Sacks, I am re-tweeting my reflections on the life and legacy of this great man: "Stay the course, and be true to yourself" he told me.

11.12.2022 3:23pm - Do you concur with @ylecun 's statement: "asymptotically, with infinite data, there is no need for priors." ?? If so, does it also apply to "inductive bias"? I hope you say Yes, because that would clarify the source of my confusion.

11.12.2022 12:34pm - (Replying to @AlexTensor @KordingLab and 8 others) Even "initial assumptions" is too vague. When we say "assumptions" we need to specify: assumptions on what? On the structure of the NN? On the distribution of the data? On the world behind the data? The word "knowledge" behooves you to be clear on these distinctions.

11.12.2022 12:29pm - (Replying to @ylecun @GaryMarcus and 11 others) "So, asymptotically, with infinite data, there is no need for priors." This sentence may hold the key to the "inductive bias vs. knowledge" debate. The latter is needed even asymptotically, e.g., no amount of data can tell us that the barometer does not cause tomorrow's weather.

11.12.2022 5:49am - (Replying to @gottfriedmath and @attilacsordas) Well put! Pre-Dawinian theories of purposeful evolution show our preference toward causal explanations over selection mechanisms, which even Reichenbach missed.

11.12.2022 5:42am- (Replying to @PhilDawid) Believe me, I feel much more comfortable discussing Abraham, whom I know from childhood, than discussing inductive bias, a new arrival on my block.

11.12.2022 5:24am- I love it. Especially the history of Israel in 7.5 seconds.

11.12.2022 4:49am - (Replying to @Ostrov_A) Hard to decide between you and Nancy Pelosi. She said: "Zionism is the most noble liberation movement in the 20th century." I can't think of any counterexample to either statement.

11.12.2022 4:24am- (Replying to @benj_kerstein) Abraham might say: No, all I wanted is to check consistency between God's action and the moral order which He himself devised.

11.12.2022 4:14am - (Replying to @RonKenett) Puzzling: "And God saw every thing that he had made, and, behold, it was very good." Even all-knowing God could not foresee the result of his action, else he wouldn't be surprised with the outcome. Lesson to us: No matter how smart, check your results.

11.12.2022 4:02am - (Replying to @PhilDawid) Abraham did not ask "what if I knew that..." He asked "what if there were". Beside, God has made a decision already. So, Abraham knows that, factually, the number is below threshold. Asking for decision-changing conditions is thinking about conditions that contradict known facts.

11.12.2022 2:20am - (Replying to @KordingLab @scac1041 and 8 others) These are some of the reasons I prefer "Knowledge" over "bias", which often goes to "gimmicks" or "hacks", like preferring one structure of NN over another, not knowing whats behind the preference except "it works". Compare it to "knowing" say that aspirin tends to cure headache.

11.11.2022 10:39pm - (Replying to @AvivaKlompas and @SAFECUNY) Progress, but falling short of explaining the moral deformity of BDS, as Martha Polack did (President of Cornell):“[BDS] often conflates the policies of the Israeli government with the very right of Israel to exist as a nation, which I find particularly troublesome.”

11.11.2022 10:25pm - (Replying to @KordingLab @scac1041 and 8 others) This is the first clear definition I received of what "inductive bias" is - a concept that has always eluded me, and that I prefer not to use, now that I know it's just "prior knowledge". Why? I can taxonomize "knowledge" by its origin, type and representation. Not "bias".

11.11.2022 7:21pm - (Replying to @KordingLab @scac1041 and 8 others) It requires information, yes, but calling any required information "inductive bias" is not serving any purpose, unless it tells us what's special about this information, where it can be found etc.

11.11.2022 7:08pm - (Replying to @KordingLab @scac1041 and 8 others) We know that observations+pertub are enough for rung-2 tasks, but I have never been able to see how things like "inductive biases (like smoothness), taken totally from the world of prediction, can move us an inch higher.

11.11.2022 7:01pm - (Replying to @MatthewEGunter @cleverclue and @johncutlefish) I do not characterized my work on causality as Bayesian. (I actually wrote a paper "Why I am only half Bayesian" I consider the two to be orthogonal, nor can I understand how causality could be viewed as constraints over NN; the latter serves the former.

11.11.2022 6:50pm - (Replying to @KordingLab @scac1041 and 8 others) Sure. When I "listen to" X, I do f(X) regardless of whether X is perturbed externally, or changes spontaneously or stays constant.

11.11.2022 4:28pm - (Replying to @scac1041 @PhilDawid and 8 others) I do not regard a causal DAG as a primitive. The primitives are the "listens to" relationships with the help of which we construct the DAG. These provide the "deeper foundation" needed.

11.11.2022 4:25pm - (Replying to @soboleffspaces and @PhilDawid) It would be additive if Abraham were to ask: What if we ADD 50 righteous person to the city. But Abraham did not talk about addition, only about "if there were".

11.11.2022 4:22pm - (Replying to @PhilDawid) Abraham believed, I presume, that God could count and therefore God knew the true number of righteous people in the city, call it X. So, what's the point of asking "What if X=50". Obviously, X was not equal 50. Hence counterfactual.

11.11.2022 4:10pm - (Replying to @PhilDawid) Abraham did not ask "what if I knew that..." He asked "what if there were". Beside, God has made a decision already. So, Abraham knows that, factually, the number is below threshold. Asking for decision-changing conditions is thinking about conditions that contradict known facts.

11.11.2022 4:10pm - (Replying to @attilacsordas) The #Bookofwhy summarizes Reichenbach "Principle of Common Cause" and its faults, on page 199. If you think we missed anything of it, or did not fully appreciates its significance, I would love to re-discuss it. Today, having a language of causation, we can do it meaningfully.

11.11.2022 4:03pm - (Replying to @mundt_martin and @GaryMarcus) I am always supportive of bridging sub-fields but, humbly, the description I read does not help me believe that continual learning can "enable knowledge transfer... when distributional shifts are experienced". Is it demonstrable on a toy example, say fig.1

11.11.2022 3:31am - It's the first counterfactual "what if" in the Bible; Abraham the scientist is trying to unearth the laws of universal justice; it's a language even Gods understand, as told in the #Bookofwhy.

11.11.2022 2:44am - (1/2) For those who read my "Miracles of November" and were wondering how the 75th anniversary to the UN Vote, Nov. 29, 1947, will be celebrated, here is the information: A community event is planned for November 28, 11:30 am at the Museum of Tolerance, LA.
11.11.2022 2:44am - (2/2) Israel's Consul General has invited the diplomatic corps in LA to attend, see flyer: The program will include a re-enactment of the vote, videos from the period, school choirs and more. Mark your calendar and join us in a joyful thanks-giving fest.

11.10.2022 1:07am - Another paper deserving our discussion is Weinberger's "Path-Specific Discrimination" which illuminates debates about the meaning of "fairness" or "discrimination" in the context of causal theories of mediation, and path-specific effects.

11.9.2022 9:25pm - (Replying to @herdiants) Unconvincing. Population-specific interventions are routine in SCM, requiring no change in the meaning of counterfactuals.

11.9.2022 6:57pm - Congratulations to professor Jeffrey Herf, upon receiving the 2022 Bernard Lewis Prize for his book "Israel's Moment: International Support for and Opposition to Establishing the Jewish State, 1945-1949" (Cambridge University Press, 2022)

11.9.2022 3:58pm - (Replying to @JKugelgen) Seeing the prisoner dead, I wish to conclude: "Had soldier-1 not shot, the prisoner would still be dead". Regardless of whether soldier-1 turned disobedient or his rifle got jammed. I do not believe the subjunctive mode is appropriate for evidential (backtracking) conditionals.

11.9.2022 6:52am - Israel unearths ivory comb from 1700 BCE inscribed with plea to stop lice
11.9.2022 6:52am - Why do such discoveries make Israelis jump from joy while evoking no reaction from their neighbors, who claim to be the true descendants of the Canaanites? Ans.: The former stood there with Joshua when he said: “Sun, stand still at Gibeon, and moon, in the Valley of Aijalon.”

11.9.2022 6:13am - (Replying to @ruescasd) Figuratively saying "breaking the laws of nature" really means "changing state of the world". See Firing squad example in #Bookofway, eg, "had soldier-1 refrained from shooting".

11.9.2022 5:45am - The Jewish Studies Professors Who Traffic in Antisemitism via @jewishjournal
11.9.2022 5:45am - Liberal-minded readers who are as disturbed as I am by the right-swing outcome of the Israeli elections, should find this paper useful in countering the recent dooms-day interpretations offered by Jewish Studies professors who never reconciled to the idea of a democratic Israel.

11.9.2022 4:53am - (1/3) It's time again to examine this week's production of causal inference publications: Among the many inviting titles (5 1/2 pages), my eyes zoom on "Backtracking Counterfactuals" which formalizes an alternative mode of
11.9.2022 4:53am - (2/3) counterfactual reasoning, called "backtracking", within the SCM framework. Evidently, there has been some dissatisfaction in the philosophical literature about the First Law, and alternatives are now being proposed to the 3-step method of computing counterfactuals. Instead
11.9.2022 4:53am - (3/3) of breaking the laws of nature, we modify the initial conditions. I am still trying to convince myself that this is a reasonable alternative in some situations, but I'm missing a compelling example. Perhaps some philosophy-minded readers could help us here??

11.8.2022 1:41am - Can't wait to see this tablet in a museum, commemorating the successful completion of King Kezkiya tunnel, having been reading this story many time in schools and re-living it while visiting Jerusalem 2012. Oddly, I can hardly read a word, and I thought I spoke Hebrew.

11.7.2022 11:00pm - (Replying to @nathanb_science) I am missing the effect of What?

11.7.2022 10:34pm - I purposely selected this semi-amateurish performance of "Eli Eli" which, unlike the usual concert-hall performances, reminds me of the many evenings we (as teens in Israel) used to sit around the camp fire, tell stories, fall in love, and sing our hearts out: "Eli Eli".

11.7.2022 9:41pm - 78 years ago, on November 7th, 1944, Channah Szenes was executed. You are probably familiar with some of her poems, but the one below, Eli Eli, became more than a poem; it became a calling for generations:

11.7.2022 9:03pm - (Replying to @JohnHMorris @jhurwitz and 3 others) I am inclined to agree with this statement, with emphasis on "today". This does not mean that AI will not bridge this difference tomorrow.

11.7.2022 8:40pm - (Replying to @cjpberry) The unprincipled claims we are hearing from data-fitting enterprises (eg, ML, DL) remind me very much of pre-Socratic debates in anient Athens -- "it seems to work, hence..." Socrates method urges debaters to start with agreed principles, then proceed to the conclusions.

11.7.2022 8:14pm - (Replying to @fisagor) I'll vouch for that!

11.7.2022 3:42pm - WATCH: Iranian girls fight back, refuse to believe teacher's lies about 'Zionists' | World Israel News
11.7.2022 3:42pm - I wish our students here, at UCLA, would gain the courage to scream at their history professors: "Your arguments are false, stop blaming Zionism for evils you are promoting." Thank you, Iranian girls, for educating our timid American students.

11.7.2022 1:27am - We are spending all our resources fighting a hate that the haters claim is hard to define. If we spent a tiny fracture to fight anti-Zionism, which every 10 year old can define, and which ferments at least 50% of the hostilities, we would save a fortune, and make some progress.

11.6.2022 12:59pm - (Replying to @diomavro) The don't count "de-confounders" as "expert witnesses"

11.6.2022 4:36pm - (Replying to @yudapearl and @diomavro) Misspelling correction. It should read: They (i.e., the courts) don't count "de-confounders" as "expert witnesses"

11.6.2022 4:34pm - Thanks for reposting, doubly relevant today, with thoughtless "Jewish scholars" urging the UN NOT to accept the IHRA definition of antisemitism, and misguided Jewish professors telling the @latimes that: Israel is no longer what our parents imagined it could be. Read Avot 1:11

11.6.2022 1:49pm - Video clip from Rabin's memorial in Jerusalem. President Herzog 15:20 min, PM Lapid 28:40 These are Israel's answers to all the scavengers who have gathered to bemoan her democracy, devour her legitimacy, or diminish her spirit of perpetual rebirth.

11.6.2022 1:09pm - To Itzhak Rabin, murdered Nov. 4th, 1995. Every person has a name which his sins gave him, and which his longing gave him. Every person has a name which his enemies gave him and his love gave him. Every person has a name which the sea gave him, and which his death gave him.

11.6.2022 6:08am - On Nov. 4th, 1995, Yithak Rabin was assassinated. To mark this day, I've chosen to share the song "Every Person has a Name", sung by Hava Alberstein Every Person has a Name that God gave him and which his father and mother gave him

11.6.2022 3:04am - (Replying to @Michael_Wgd) Sadly, 128 so called "Jewish studies scholars" submitted a statement to the UN, imploring it NOT to adopt the IHRA WDA. The "Jewish studies Zionist network" was formed to prevent such hair-splitting to ruin the reputation of "Jewish studies" proper. See

11.6.2022 1:15am - (Replying to @Michael_Wgd @sunwins888 and @WatsonSpeaking) I can't figure out how any Jew can be "fake" when Jewishness is a state of mind. Either you have this state, and you identify with the collective memories of the Jewish people, or you don't. What is there to fake?

11.5.2022 11:17pm - (1/2) Recent discussions on IPW with different sets of covariates call my attention to a 2014 paper on "confounding equivalence" which, sadly, has not received the attention it deserves. (Only 46 google citations - shame!). It tells us when one covariate
11.5.2022 11:17pm - (2/2) set is guaranteed to yield the same bias reduction as another. It also extends the backdoor criterion to post-treatment covariates. It's one of those forgotten papers I'm truly proud of (thanks of course to my coauthor, Azaria Paz).

11.5.2022 11:11pm - (Replying to @soboleffspaces) Rubin's proclamations have had no negative effect on my students, but they have set back statistics and econometric education for at least 30 years. I have colleagues at UCLA who are still teaching that way.

11.5.2022 8:56pm - (Replying to @soboleffspaces) This shock can be seen as one of the worst educational consequences of the Potential Outcome framework. By refusing to deploy cognitively-compatible models of reality, it blinds people from detecting fallacious statements about reality, let alone making them.

11.5.2022 3:40pm - Pelosi Cites Israeli Poem in 1st Public Appearance Since Attack on Husband
11.5.2022 3:40pm - My Israeli friends are singing it after the election: “I have no other country. Although my land is burning, my veins, my soul, with an aching body and with a hungry heart, here is my home,” “I will not be silent. For my country has changed her face. I will not give up on her.”

11.5.2022 12:51pm - (Replying to @AsraNomani and @Columbia) We had a professor at the Technion, who used to scold us: "You behave like Asiatic!". It never occurred to me that, some day, I would be defending the honor of my Asian heritage, together with my Muslim sisters from India.

11.5.2022 12:36pm - (Replying to @soboleffspaces @RaulMachadoG and @spphubc) It's stronger! Inverse-treatment-probability weighting is equivalent to stratification and averaging (the formula produced by do-operator) if stratification variables are the same in both, regardless of whether they satisfy the identifiability criterion.

11.5.2022 12:23pm - (Replying to @policytensor and @_alice_evans) I couldn't find out what "sort of thing" you refer to, nor who "They" are. Someone must have deleted the Tweet.

11.4.2022 8:23pm - The problem with taking proactive security measures is that, eventually, you will be blamed for being the aggressor. Israel has been under such dilemma for 75 years.

11.3.2022 5:56pm - A brilliant lecture presented last week at UCLA, by professor Walter Russell Mead which should guide and educate us, as we are preparing to celebrate the 75 anniversary of the historic UN Vote of Nov. 29 1947, @DovWaxman , @Ostrov_A

11.3.2022 1:14am - (Replying to @LekhtNaya) Well put. And the worst violator of this principle is @jstreetdotorg with their offensive slogan "pro-Israel, pro-peace", as if being pro-Israel does not automatically entail being pro-peace. As if one must go to great effort to find peace-loving Israelis. What an insult.

11.2.2022 11:59pm - (Replying to @RaulMachadoG @soboleffspaces and @spphubc) The paper shows that many causal models, not merely backdoor types, have computationally efficient "inverse probability type" estimators.

11.2.2022 11:44pm - Having checked #Columbia Professor @JLeeSoc 's recent papers, I am not sure she is aware of modern definitions of the phrase: "may have also benefited from their racial status" in the causal inference literature. @eliasbareinboim works on causal fairness are relevant here.

11.2.2022 11:35pm - Give me a platform like the UN and I'll find you 300+ Jewish "scholars" who would rush to reject any definition, of any concept, from motherhood and apple-pie -- this is what the Talmud teaches us to do; hair-splitting at any cost.

11.2.2022 10:57pm - Confirmed. I've watched Mehdi Hasan the past 5 years. He is no lesser bigot than David Duke or Itamar Ben-Gvir. He is definitely more dangerous, given his saintly platform on @MSNBC .

11.2.2022 10:49pm - The logic is simple. If Israel is indeed as evil as @amnesty describes her, then given that most Jews support Israel makes Jews either hopelessly morons or dangerously evil, both deserve harassment by the lunis.

11.2.2022 2:35pm - The Balfour Declaration of 1917, which committed Britain to support the creation of a Jewish homeland, came about in large part due to behind-the-scenes efforts by three women, writes historian Natalie Livingstone. via @WSJ
11.2.2022 2:35pm - I've always felt there must have been a woman's hand behind the Balfour Declaration of Nov. 2, 1917, Weizmann could not have pulled it all by himself. This essay names 3 such women. Let's remember them and what they have done as we celebrate 105 years to that historical moment.

11.2.2022 1:47pm - (Replying to @cto_maverick) Another reason: "everybody is fascist" said @leibowitzadak , so he must be right.

11.2.2022 12:46pm - In my little corner, I haven't seen #ML and #DL claiming X; all I see are claims about X being achievable, eventually, given enough #ML and more #DL. These claims may have an element of wisdom, once we rethink #ML and completely revamp #DL:

11.2.2022 12:11pm - Finally! Perfect timing! Today, Nov. 2, 2022, exactly 105 years after the Balfour Declaration of 1917,, Israel's neighbors found a reason for rejecting co-existence. Makes sense.

11.2.2022 12:01pm - True, but the main point of the Balfour Declaration was that place of residence in the PAST should be weighed against historical attachment at the PRESENT. The Arabs had more of the former, the Jews had more of the latter, therefore, SHARE.

11.2.2022 12:58am - (1/2) Readers who question the causal connection between the Balfour Declaration (Nov. 2, 105 years ago) and the rescue of 400K Jews from Hitler's Europe, should note that most of these Jews were able to come only after a state-like infra structure was laid in Palestine
11.2.2022 12:58am - (2/2) by the Zionist pioneers who, inspired by the Declaration, immigrated in the years 1917-1927. The land was in total ruins at the end of WWI. The inspirational impact of the Balfour Declaration far exceeded His Majesty's anticipation, as described here

11.1.2022 10:37pm - It also means that those who do not believe in "both do" must wait till they do.

11.1.2022 10:21pm - Los Angeles City Council Passes IHRA Definition of Antisemitism via @jewishjournal
11.1.2022 10:21pm - Though I rarely use the charge of "antisemitism" against Zionophobes, today's vote of the LA City Council labels many of my UCLA colleagues "antisemites" (primarily for denying the Jewish right to self determination), thus giving students the nerve to call them out in class.

11.1.2022 9:11pm - Every time I see this derivation I fall in love again with mathematics. I wonder how ML folks can go through life without napkins. #Bookofwhy

11.1.2022 8:53pm - (Replying to @HumzahAlkindi) I don't know if you are speaking from personal experience or hearsay. The "living together" fable is paved with bloody reality. (My wife has experienced the latter).

11.1.2022 8:49pm - (Replying to @_ChrisAlbertyn @IbrahimZina and 9 others) Oh God! This looks like my handwriting! You caught me in poetic mood, thanks.

11.1.2022 8:47pm - (Replying to @soboleffspaces and @spphubc) You are underestimating the grip and duration of academic loyalty oaths; we hear of many defectors from countries and religions -- very few from academic paradigms, where allegiance is institutionalized.

11.1.2022 1:52pm - She should have been suspended 2 years ago, on a much stronger and more provable charge: anti-Zionist form of racism.

11.1.2022 1:29am - (1/2) I literally owe my life to the Balfour Declaration. It encouraged 400K Jews to escape inferno Europe and immigrate to the holy land between 1917 and 1939. Among them was my father who, by virtue of being a British subject (1924), was able to go back to Poland, marry my
11.1.2022 1:29am - (2/2) mother, and bring her over to Israel (1935). Her family was stranded in Europe and perished in the Holocaust, together with 6 million other victims of the British White Paper. I therefore find it offensive for The Guardian to apologize for supporting Balfour in 1947.

11.1.2022 1:34am - Sorry, I meant "for supporting Balfour in 1917".

10.31.2022 11:43pm - (Replying to @soboleffspaces and @spphubc) If you are asking "Why", I have a soft spot for the question. My answer is: (1) Because Rosenbaum and Rubin did not know about backdoor in 1983 and (2) Because they signed their disciples on a life-long loyalty oath, so that backdoor will remain a mystery for generations.

10.31.2022 10:44pm - (Replying to @Claire_V0ltaire) They will not be tired, as long as we refrain from calling them "Zionophobes". Only then they would wake up to ask: "What have we done wrong? or stupid? or totally thoughtless?" It's our responsibility to give them a mirror.

10.31.2022 10:27pm - (Replying to @artistexyz) I have long been forgiving to people who are allergic to DAGs; it is after all a new language that some find spooky. What I cannot understand is blindness to nonparametric analysis, to transparency and testability.

10.31.2022 10:21pm - Opinion | Opinion | Is Israel really a settler colonial state?
10.31.2022 10:21pm - I have written at length about the absurdity of associating Israel with "settler colonialism," but this paper by Zasloff etal puts the absurdity in its proper historical-ideological perspective. Main take: "Colony" in the 19th century meant a "refuge". @Ostrov_A @michaeldickson

10.31.2022 9:01pm - (Replying to @titudeadjust @Popehat and 3 others) Can't tell what we agree on.

10.31.2022 8:56pm - (1/4) More CI papers have been added to our harvest at Deserving our attention is Imben's CAUSALITY IN ECONOMETRICS: CHOICE VS CHANCE, a printed version of his Nobel lecture, that should always be read with the corrections listed in
10.31.2022 8:56pm - (2/4) My comments pertain to the last section, titled Causality and Computer Science, in which he essentially tells economics students: Oh, BTW, there has been some work done outside of economics, but it has merits only in computer science, where there are "large number of
10.31.2022 8:56pm - (3/4) distinct components", not in economics, where "credible causal inference is particularly challenging". Not so quick! Credible inference (as emphasized in requires: Transparency, Power and Testability, three features lacking in the PO framework
10.31.2022 8:56pm - (4/4) to which Imbens is so totally wedded. To ignore these three deficiencies of the PO framework is to summon economic students to another decade of isolation from mainstream causal inference.

10.31.2022 6:06pm - Happy to share a beautiful concert at Kibbutz Maagan Michael, Israel, celebrating the Daniel Pearl World Music Day, by Ensenble Consart (Leipzig, Germany) The song performed is "Shalom Aleichem," a traditional Friday night song written in the 16th Century

10.31.2022 3:36pm - Another Monday's harvest of papers dealing with causal inference Catching our attention is Woodward's "There is no such thing as statistical explanation", though based on quick reading, Woodward is yet to internalized the First Law

10.31.2022 10:24am - (Replying to @artistexyz @stephensenn and @eliasbareinboim) Enticed by your post, I've read the "Lord Paradox" again, both in the #Bookofwhy and in Agree, it's beautiful and clear (specially when each dining hall has only one diet.) I'll stay out of the new debate -- no point spoiling clarity for soiling sake.

10.31.2022 9:55am - (1/4) This coming week brings us an inspiring event, the 105 anniversary of the Balfour Declaration, Nov. 2, 1917 -- the first international recognition of the Jewish people right to a homeland. I have written about its historical significance in 2017,,
10.31.2022 9:55am - (2/4) and my thoughts have not changed much, despite The Guardian recent editorial, in which it lists its support or the Balfour Declaration among the paper's "worst errors of judgment". [How cute!] Speaking of monumental events,
10.31.2022 9:55am - (3/4) we are just a few weeks ahead of November 29, the day the UN Voted for the partition of Palestine, exactly 75 years ago. I have written about this forgotten miracle here: and, believe it or not, this year the LA community will celebrate the event
10.31.2022 9:55am - (4/4) at the Museum of Tolerance. Mark it on your calendar, please, Sunday Novemeber 29, and wait for final details and program. Happy Halloween and a meaningful November.

10.29.2022 4:17am - (Replying to @AnnGreenberg and @noatishby) But that has not prevented campus climate from becoming what it is, and has not equipped students with the weapons to tell their bullies: "You sound like a Zionophobic racist! Don't tell me you are one!" Bullies smell weakness.

10.31.2022 2:47am - (Replying to @WSB_OfficialMod @HarvelaHarvey and @noatishby) The 1967 border can only be respected when Arabs recognize the right of Israel to exist IN ANY BORDER. As long as Palestinians, to the last person, are committed to obliterating Israel given the opportunity, it's merely a cease-fire line. For facts:

10.31.2022 1:41am - (Replying to @Ostrov_A) Our sages said (Avot, 1:11): "Scholars, be careful with your words, because your enemies will twist them against you ... and your disciples will die." @PeterBeinart prides himself on studying the Talmud. Alas, his disciples will not be there for him, when Mehdi's friends prevail.

10.31.2022 12:51am - (Replying to @GeraldNGOM @theintercept and 3 others) @jrosyfield meant it to be a parody of how the "apartheid hyenas" of @KenRoth and @amnesty are emptying the word from content. Tomorrow's parody: Israelis sell land to Israelis, Palestinians can't - apartheid! Tuesday's: Palestinians are paid to kill, Israelis r not - apartheid!

10.31.2022 12:04am - (Replying to @guntax59 @user14201 and 13 others) This is where a formal theory of causality comes in; it tells us that dry data is useless w/o the ropes behind the data; what the players' aims are, and what would happen if one player stops shooting.

10.30.2022 11:46pm - "Zionophobia" changes the conversation entirely from *AZ: Zionism is inherently anti-Palestinian *NT: It's a mistake, I'm not ... to *NT: Anti-Zionism is inherently racism; it denies a homeland to a persecuted people. Why are you spreading racism at Berkeley? *AZ: I'm not...

10.30.2022 7:14pm - (Replying to @Minnhawk1 @shubham19881 and 11 others) Religion played an important role for both Pakistanis and Palestinians (1947). The former sidestepped the extremists and compromised on partition, the latter were led by a religious fanatic (The Mufti) who fabricated the "Al Aqsa in danger" fable, and foiled the partition option

10.30.2022 6:30pm - (Replying to @VladTheSlav5 @ChesterMorton96 and 11 others) "He was a fighter" Palestinian logic formalized. 1. X attacks Y. 2. Y defends itself (mandatory service, reserve service, even women and seniors.) 3. X declares men and women legitimate targets (children too -they grow to be soldiers) 4. Y: Enough! 5. X: War crimes! Apartheid!

10.30.2022 7:48am - (1/2) Comment # 1 on @EinatWilf video The rationale given by the Arabs (in 1920's) for prioritizing their war against a Jewish State over all other considerations was not fears of dispossession by Jewish immigrants, but a profound ideological
10.30.2022 7:48am - (2/2) resistance to the idea that some Arab farmers would find themselves in a minority, at the mercy of a Jewish-majority state. This was expressed explicitly in the Arab newspaper Al-Carmel (1929) and quoted in Jabotinsky's Medina Ivrit (1937). See:

10.30.2022 6:53am - (Replying to @Crazy_fox65 @backforhowlong and 12 others) Me too, with both eyes.

10.30.2022 6:21am - (Replying to @ShaharAzani @Israel and 9 others) Warped ears in the West have long been programmed to translate Palestinian acts of terror into heroic acts of freedom fighting. How ready Palestinians are as "partners for peace" is better betrayed by their intellectuals, now openly dancing: "From the river to the sea".

10.30.2022 5:53am - (Replying to @soboleffspaces @RaulMachadoG and @AlfredoMorabia) This would be an insult to Epi literature, for not reading Yule (1900), who also advocated stratification in order to block potential confounders. I'm sure many epidemiologists used stratification in the period 1900-1973 to remove the bad influence of potential confounders.

10.30.2022 5:41am - (Replying to @anndvision) The premises in SCM are those fragments of knowledge that you consult when you decide that no arrow should go from X to Y. Namely, Y listens to other variables, not to X. "Listening" is a "unit-level" (rung-3) concept. Interventional premises are averages:

10.29.2022 4:37pm - (Replying to @KordingLab and @metrics52) I don't see where the generalization comes in. Is it that we have SEVERAL problems, P_1, P_2...? or is it that we have both P(y|do(x)) and P(y|x) ?? Do we need more than one DAG to explain what your method is doing? Recall, one DAG can represent many problems and many realities.

10.29.2022 11:44am - (Replying to @a_sarentorbic @bsgallagher and @GaryMarcus) Such characterization makes me suspicious that you haven't read #Bookofwhy.

10.29.2022 10:48am - (Replying to @GaryMarcus) A cause-effect question. Do we appreciate Kant because we are here, or should we read him to get further ahead of where we are?

10.29.2022 4:17am - (Replying to @AnnGreenberg and @noatishby) But that has not prevented campus climate from becoming what it is, and has not equipped students with the weapons to tell their bullies: "You sound like a Zionophobic racist! Don't tell me you are one!" Bullies smell weakness.

10.29.2022 3:28am - Something is missing. These ignorant students who are regurgitating populist slogans at @noatishby do not even begin to suspect that they are engaged in less than super-righteous activity, let alone a racist movement with genocidal goals. The word "Zionophobia" is badly missing.

10.29.2022 2:55am - OTD. On October 29, 1969, 53 years ago, the internet was born at UCLA. I was hired only 3 weeks earlier, and I didn't quite understand what the people on the floor below me were celebrating. Now I understand.

10.29.2022 2:21am - (Replying to @AlfredoMorabia and @soboleffspaces) So, can I assume that my account of causation in epidemiology, narrated here:, is not missing a major conceptual breakthrough, although I skipped many prominent names. (I may still have a chance to add, if convinced otherwise.)

10.29.2022 1:42am - @AliAbunimah , the founder of Electronic Intifada sides with Putin. Not surprising. The Russians and the Palestinians are the only two peoples who decided that their neighbors do not have the right to exist. Both fight for ONE-STATE solutions, and deem co-existence as injustice.

10.29.2022 12:25am - The only explanation I have why @EinatWilf 's video does not garner >100K "likes" is that Elon Musk is now meddling with Twitter's moderation algorithm and, evidently, has biased it against simple, inconvenient truths. Please retweet before it's cancelled.

10.28.2022 5:18pm - (Replying to @AlfredoMorabia @soboleffspaces and @ken_rothman) I believe I've read this article at the time, but it confirmed the perspective I had while writing Causality, in 2000. Do you think I've missed important conceptual milestones in my recent review:, ???

10.28.2022 4:36pm - (Replying to @totteh) Hilarious! If it were not serious. This means that readers on our channel would have to start faking it and talk like PO parishioners, eg, "treatment assignments" "target trials" "death to Rung-3" etc. I feel manipulated already.

10.28.2022 4:23pm - (Replying to @DKedmey) Please do nominate me to Twitter's "content moderation council" but, warning, my first recommendation: Zionophobic content is moderated same as "child pornography"; perhaps even harsher, same as Islamophobic content.

10.28.2022 8:09am - American friends often ask me: What is it about Israel/Palestine that you, Israelis, think you know and that we don't. I cannot think of a better answer than this video clip of @EinatWilf Please don't miss a word, even if you think you know and we don't.

10.28.2022 6:36am - Dear @FranceskAlbs Those who deny Israel's right to Tel-Aviv, are hardly in position to lecture us how she can defend herself in Ramallah.

10.28.2022 5:49am - Check out this article:
10.28.2022 5:49am - Anyone knows how this might affect our educational channel? Would we still be able to call out non-scientific approaches to data science? Would Ye return to his rants? Would the secret outreach algorithm be politicized?

10.28.2022 2:47am - @kanyewest has given bad name to antisemitism through his incoherent and blunt racism. Much more dangerous are the soft spoken "apartheid hyenas" at the UN, whom many people, mostly the gullible and uneducated, tend to parrot.

10.28.2022 2:20am - (Replying to @soboleffspaces and @AlfredoMorabia) My question to those who read Miettinen: Did he state conditions for the set Z to make the adjustment formula correct, or merely wrote down the formula by intuition. I haven't read him, assuming (perhaps prematurely) that, lacking notation, he did the latter only. Did I err?

10.27.2022 10:26pm - (Replying to @soboleffspaces and @AlfredoMorabia) My question remains, if you have no notation for causal effects, no do(x), no Y_x, we are back in Yule's days; adjustment seems desirable, but what is an appropriate adjustment set? What was Mietinen answer? I'm not contesting, just curious. Is an answer feasible w/o notation?

10.27.2022 8:18pm - Curious. How did Miettinen handle stratification without causal notation? I have credited the scientification of epidemiology to Greenland and Robins (1986),, seeing that Miettinen tried to define confounding using probabilities. @AlfredoMorabia

10.27.2022 7:53pm - We need a stronger word, perhaps "Jew-cleansing" to describe "Jew washers" like Mehdi Hasan, who get terribly indignant when people point them to this glaring hypocrisy.

10.27.2022 7:25pm - (1/2) (Replying to @AleksanderMolak and @rasbt) Most of my causal knowledge came from language. My father taught me why we need to water the orange trees, and how often. My Dr. told me to exercise or else. There is no reason therefore that LLM will not obtain causal knowledge from text sprinkled with causal utterances.
10.27.2022 7:34pm - (2/2) (Replying to @yudapearl @AleksanderMolak and @rasbt) However, I would not call it "implicit", because in my mind I have a clear arrow going between X=water and Y=tree-growth. I therefore conjecture that we extract causal knowledge from language by filling in the missing parts of a preprogrammed template (eg, DAG) = curiosity.

10.27.2022 4:37pm - (Replying to @itamarcaspi @DuduLagziel and @giladmp) I would also recommend the paper: "Understanding Simpson's Paradox" which goes beyond #Bookofwhy, and resolves the "paradox" from both decision making and cognitive perspectives.

10.27.2022 1:42pm - (Replying to @xiaodai10838475) The word "biased" is a compliment to a Zionophobe. 1:42 PM · Oct 27, 2022 ·Twitter Web App

10.27.2022 4:39am - No Mehdi, it's got nothing to do with your being or not being an antisemite. No, you are a carrier of a much uglier affliction: Zionophobia - a moral deformity that cannot be blamed on alcoholic parents or abusive upbringing; its purely calculative, and knowingly genocidal.

10.27.2022 3:02am - Glad this Tweet was re-liked, for it has become timely again in light or the new @ylecun - @GaryMarcus debate. I'm not sure though whether this definition of "Deep Understanding" would meet universal acceptance in "Deep Learning" land.

10.27.2022 1:53am - I don't know about you, but I get emotional watching high profile meetings such as this: Just watching two great nations rising above the cynics and re-asserting their unique greatness makes my heart melt.

10.27.2022 1:21am - (Replying to @arieljalali) I am not taking position here on whether hate mongering or child pornography should or shouldn't be allowed on Twitter and other social media channels. I am merely requesting that Mehdi's Tweets on anti-Semitism should be given equal treatment.

10.27.2022 12:03am - (Replying to @blakeflayton) Once you baptize yourself with the holy water of anti-Zionism, people might begin to take you seriously, even when you have nothing to say.

10.26.2022 11:49pm - (Replying to @PHuenermund) I've read that sharks are scared of people!

10.26.2022 11:47pm - (Replying to @jhurwitz @soboleffspaces and @stu_frost) Wow, Nov. 7 is just around the corner!

10.26.2022 11:45pm - Is this a proof that "diplomacy works"? Or that IDF' nightly raids into Nablus Kasba are beginning to show results? Or that the Palestinian Authority is going underground?

10.26.2022 11:38pm - Many of our readers would recognize, among the 11 victims, the name of Joyce Fienberg, the wife of Professor Steve Fienberg, a leading statistician at CMU, (Co-author of "Discrete Multivariate Analysis, 1975, and "Statistics and the Law", 1980). YZB

10.26.2022 11:27pm - On a scale from 1 to 10, what would be more unbearable, for Twitter to reinstate Ye West's antisemitic account or for Twitter to let Mehdi continue his anti-antisemitic lecturing?

10.26.2022 8:11pm - (Replying to @Eradicator_NYC and @mehdirhasan) I would forgive him if he was paid. Some organisms truly believe in what they are taught in Al Jazeera school of journalism.

10.26.2022 6:19pm - It hard to believe @mehdirhasan is still tweeting "on the right"..."on the right" as late as Oct. 25, as if he truly believes his audience have not noticed his hypocrisy, and his grand contribution to anti-Jewish hostilities in this country. Or, is he trying to deflect guilt?

10.26.2022 6:02pm - They rarely attack unapologetic Zionists, namely students or faculty who do not proclaim themselves "unapologetic Zionists", but act like ones.

10.26.2022 5:53pm - (Replying to @SMehrizi @ylecun and @GaryMarcus) Scientists should be forgiven for having big egos, it goes hand in hand with being scientists. They are expected, however, to look counter theories and evidence supporting those counter theories.

10.26.2022 5:47pm - (Replying to @gottfriedmath @nntalebbot and @nntaleb) That prediction is not enough was known centuries ago. What I dared do was to define "understanding" as the ability to answer questions from all three rungs of the Ladder.

10.26.2022 5:42pm - (Replying to @think___y) I begin to think like you, even at the expense of my inborn modesty.

10.26.2022 4:24pm - This piece is amazing in several ways. (1) That @jdforward published it. (2) That already in 2012 I warned the community about the epidemic we see today: Zionophobia. (3) That my vocabulary in 2012 was punchier than it is today. (4) The community was truly surprised in 2012.

10.26.2022 3:49am - Shalom Tel Aviv. I see an "Adidas" sign on the right; the store has probably tripled its sales after cancelation of the contract with West. The buildings and the pigeons are smiling. Shalom.

10.25.2022 11:04pm - (Replying to @erikbiz) All we have to do is watch her glittering gilded outfits, how she "loves" her Jewish friends, how sweet she is to her animals and how she hates carbon emission -- her Zionophobic poison is craftily gilded and sneakingly penetrating.

10.25.2022 10:41pm - Oddly, Bella has managed to escape universal condemnation. Why? Partly because she "loves" (stateless) Jews, and partly because the public does not deem her Zionophobia as dangerous as West's antisemitism. They are profoundly wrong; her gilded poison penetrates much deeper.

10.25.2022 8:22pm - It took them time to notice where the guy is heading, non-stop. It remains to see if @adidas values of "diversity and inclusion" cover Zionists like me, not too sure about their "Semitic" DNA, but sure like hell about their historical identity. Would @adidas shun anti-Zionists?

10.25.2022 7:10pm - (Replying to @FJnyc @michaeldickson and 3 others) "Blocked by @MSNBC 's Mehdi Hassan" has become a badge of honor for rational thinking peace activists. I encourage journalists and bloggers to apply for this honor. I'd be looking for it on the resume of any job applicant. Why? You really need to threaten his racism to get it.

10.25.2022 6:52pm - (Replying to @adamsafron and @DaphneKoller) It is a pleasure indeed to listen to Daphne Koller going over the history of turbo-codes and Belief Propagation algorithms, and reliving those days in the 1990's when the connections among ideas were discovered.

10.25.2022 6:33pm - (Replying to @einsley23) For interventional tasks (rung 2) we do not need the twin network, we just use do-calculus. Is learning do-calculus what you find to be hard?

10.25.2022 9:47am - (Replying to @ylecun and @TiernanRayTech) The antithesis of Tabula Rasa would be a system that requires "domain specific" knowledge whenever a new domain is encountered. DL prides itself on saving users this labor, as described here: Is "new architecture" another word for "domain knowledge"?

10.25.2022 9:32am - Just finished my presentation at the @AIBigData22 congress (Barcelona), grateful for posting my main slide "The Seven Wisdoms of Causal Inference" and for the insightful questions at the QA period.

10.25.2022 2:10am - If I were a Palestinian, I'm not sure I'd welcome the support of someone with Kanye's reputation. Would you?

10.25.2022 1:26am - (1/2) I see another debate between @ylecun and @GaryMarcus on zdnet. Gary raises again the question of where causal inference resides in @ylecun 's philosophy of learning. I just want to add that I am not complaining @ylecun "does not mention my work" (I have enough citations) but
10.25.2022 1:26am - (2/2) that he doesn't call the attention of his mega-crowd of followers to (1) the importance of causal reasoning and (2) what they can learn from the causal reasoning exercise onto how to snap out of the Tabula Rasa trap in which Deep Learning is finding itself. @TiernanRayTech

10.25.2022 12:01am - (Replying to @artistexyz @einsley23 and 4 others) Why would they do a nasty thing like that? All I am asking is to understand how one would construct a FFRCISTG, so we can understand what the "g-formula" is telling us.

10.24.2022 11:55pm - This past week has brought us another flood of papers in causal reasoning Too many to follow, but my eyes fell on which demonstrates "The Insufficiency of Statistics for Detecting Racial Discrimination by Police" using Simpson's P.

10.24.2022 11:16pm - (Replying to @artistexyz @einsley23 and 4 others) Nice to see a DAG with (Y(0), Y(1)) properly inserted. See also Fig. 11.7 in Causality (2009)

10.24.2022 11:08pm - (Replying to @JamesA1102 @SarahKSilverman and @JewishJournal) Wrong. Of course there are no "Jew-Free Zones" at Berkeley, it was described figuratively to sound a warranted alarm. Saying that what happened at Berkeley should not trigger a coast-to-coast alarm (as it did) is short-sighted and would eventually bring about such Zones.

10.24.2022 8:01pm - (Replying to @EpiEllie @soboleffspaces and 2 others) The issue is not which is a "nicer visual tool," but what scientific information is conveys by DAGs vs. FFRCISTGs. We know what questions a scientist must ask herself in constructing a DAG; I don't know the corresponding questions for FFRCISTGs. The meaning of "g" rests on it.

10.24.2022 5:23pm - As promised, here is the interview I gave to "Observational Studies", with the Reference section fully repaired. Enjoy, and feel free to share your comments.

10.24.2022 4:08pm - (Replying to @wv012 @mvanhout71 and @ESMDcan123) What problem was discussed in Duncan? What it solved?

10.24.2022 3:48pm - (Replying to @Claire_V0ltaire and @ashleyblair21) Let's continue; And calling out anti-Zionism is where it gets toughest, noblest and most needed.

10.24.2022 8:08am - (Replying to @Ole_Koldalone @toon_mikwee and 2 others) You are insulting Palestinians by assuming that only their neighbors are capable of taking positive corrective actions towards a lasting peace.

10.24.2022 12:54am - Speaking of Einstein's Museum, I hope they display his 1947 letter to Nehru (then PM of India), pleading with him to vote YES on the UN Partition Plan on Nov. 29, 1947, in the name of "the August scale of justice". See The Forgotten Miracle,

10.24.2022 12:37am - (Replying to @Ostrov_A and @KenRoth) In particular @KenRoth , something you do not deem important but we do: they promised to rescind the Palestinian National Charter which calls for the destruction of Israel. It was never officially rescinded, their youngsters know it, and act accordingly.

10.24.2022 12:20am - ‘Greatest Scientist of our Time’: Israel to Spend $18 Million to Build Albert Einstein Museum in Jerusalem
10.24.2022 12:20am - Having prayed all day for some good news, here is one: An Albert Einstein Museum in Jerusalem, at the Hebrew University, where we will be able to see Einstein's handwritten scribbles (including the E=mc^2 formula). I'm so proud to be associated with HUJ (Honorary doctorate 201?)

10.24.2022 12:07am - (Replying to @einsley23 @soboleffspaces and 3 others) Why do you need to see (Y(0), Y(1)) in the graph if you can derive what you want from the graph itself, without seeing them. Plus, if you really want to see them, you can, e.g., see Fig. 2b in

10.23.2022 11:48pm - I havn't seen ONE defense measure taken by Israel which was not deemed a "war crime" by her neighbors.

10.23.2022 11:24pm - (Replying to @Claire_V0ltaire) Let's not kid ourselves @Claire_V0ltaire ,THIS song is NEW. I havn't heard it, in this shrilling key, any time before. Old songs induce complacency. This song is NEW.

10.23.2022 9:45pm - (Replying to @policytensor and @bpetershome) OK, peace on earth, but please warn me next time that your "regression" comes from you.

10.23.2022 9:43pm - (Replying to @policytensor and @bpetershome) I remember doing tensor analysis in grad school and I bow to Einstein's theory of coordinate independence, but causal models do depend on choice of variables. Here I show how the direction of statistical time is reversed by choice of coordinates: (11.2)

10.23.2022 9:24pm - (Replying to @calimagna @noahholl and @FrankRHutter) I don't see why we need to guess about the gain that SCM gives them when we have given them a simple tool to diagnose and pinpoint the source of their gain: Run a toy problem, and see.

10.23.2022 5:41am - (Replying to @Ostrov_A @StandWithUs and @SAFECUNY) SHSHSHSHSH!!!! Don't tell the @nytimes !!!

10.23.2022 5:05am - A new film is out, "Why Ukraine", by my friend and mentor Bernard-Henri Levy @BHL . Don't miss! Oct. 27. At the UN.

10.23.2022 12:40am - (Replying to @mhd01 @AsraNomani and 2 others) Thanks. I hope I haven't used it innocently, as an endearing adjective, on anyone.

10.23.2022 12:32am - Great to get cultural content on Twitter once in a while - stay tuned.

10.22.2022 11:59pm - Recalling personal experience, Reza Aslan is not a person I would trust. Not surprised he is now working for the Iranian regime.

10.22.2022 11:52pm - Should be in the tool-kit of every instructor.

10.22.2022 10:17pm - (Replying to @erikbiz @SAFECUNY and 2 others) Agree, honest and simple. Which makes you think: Shat does it take for celebrities like @BarbraStreisand or @SaraSilverman to be honest and simple and come out of their silent-Zionist closets.

10.22.2022 7:21pm - (Replying to @FJnyc @jjz1600 and @arnoldroth) The Mongol raiders, so I am told, complained bitterly when the Great China Wall was erected: Apartheid! Apartheid! They shouted. "If this ain't apartheid what is?" Western analysts echoed, and echoed, and echoed.

10.22.2022 6:31pm - (Replying to @skdh) I've noticed a sense of comradeship in your tweets. They will not replace us.

10.22.2022 3:48pm - (Replying to @mvanhout71 and @wv012) "Human anatomy" in healthcare translates into "intrinsic properties of target population" in social science, invariant across backgrounds and aspirations of participants.

10.22.2022 2:13pm - (Replying to @mvanhout71 and @wv012) My Dutch is a bit rusty, but I hope you touched on these considerations:

10.22.2022 3:05am - Jewishness is based on "Amech Ami, Ve'Elohaich Elohay" ("Your people are my people and your God is my God" - Ruth to Naomi). Meaning: "Peoplehood first - Religion second".

10.22.2022 2:55am - (Replying to @AsraNomani @BOSCityCouncil and @Quillette) What an Orwellian inversion. #MahsaAmini must be turning in her grave to see the hijab, which to her (and to majority of Iranian women) meant a symbol of women subjugation celebrated as a symbol of free expression in a US city called Boston, home to MIT and Harvard.

10.22.2022 2:40am - (Replying to @HenMazzig) There is one logical flaw in your argument, @HenMazzig , the word "share" does not exist in Palestinian vocabulary. When it comes to Israel, "share" is replaced by "stolen", and "I choose" is now "I'm a victim".

10.22.2022 2:32am - (Replying to @Eve_Barlow and @ArielElyseGold) Inspired by @Eve_Barlow

10.22.2022 1:51am - (Replying to @AleksanderMolak and @rasbt) I am trying to understand what you mean by "implicit heuristic world models". If this is something that is learned by the agent, then it must be stored someplace, in some code. That code then is explicit. What makes it different from "explicit world model"???

10.22.2022 1:27am - I was about to swear this is a joke ridiculing our "apartheid parrots". It isn't! Seeing @ArielElyseGold , I reckon these parrots still exist, seriously squeaking and chirping: "Me too, Me too, Dying to be original, dying to be provocative, Me apartheid parrot too, Me too!"

10.22.2022 12:31am - (Replying to @CitalanEdward and @WealthInc247) It was my first exposure to the philosophy of science, dissolving my guilt for not understanding speculative philosophers like Kant and Hegel, and leading me to Carnap, Hampel , Goodman and Popper.

10.21.2022 9:53pm - (Replying to @tbednall and @WealthInc247) It changed my life too, forcing me to re-examine long-held assumptions from a totally new perspective.

10.21.2022 9:50pm - (Replying to @WealthInc247) Hans Reichenbach "The Rise of Scientific Philosophy".

10.21.2022 12:31am - (Replying to @CitalanEdward and @WealthInc247) It was my first exposure to the philosophy of science, dissolving my guilt for not understanding speculative philosophers like Kant and Hegel, and leading me to Carnap, Hampel , Goodman and Popper.

10.21.2022 8:58pm - (1/3) We used to have Rubin's disciples among our readers (eg, Yale economists), perhaps they can help us interpret the analogy. I assume he refers to the fact that SCMs appear to be redundant because they provide, not only ignorability conditions, but also their justifications.
10.21.2022 8:59pm - (2/3) In Rubin's framework, ignorability conditions are assumed not because they are deemed true, but because they are needed to support a claim; justification is left to post-analysis hand-waving. Graph-based justifications are so foreign to PO analysts, that Rubin once compared
10.21.2022 8:59pm - (3/3) collider-bias to "trying to balance a multidimensional cone on its tips," see I would thus be somewhat hesitant seeking hidden meaning in geometrical analogies produced in the PO camp. Anyone who understands what the triangle means - pls, we hear you

10.21.2022 10:33am - (Replying to @penguinvondoom @GaryMarcus and 2 others) Forgot what "cool RPG premise" stands for.

10.21.2022 3:16am - (Replying to @PHuenermund) Agree with the insights. I will soon post a cleaned up copy, with unmolested references.

10.21.2022 3:14am - (Replying to @AvivaKlompas) And when I point out that Zionophobes are suffering from a moral deformity, friends tell me "don't get down to their level".

10.21.2022 2:49am - Today, all Palestinian factions have declared a "Day of Rage" in memory of Super-Martyr Uday Tamimi (22), who murdered Noa Lazar 11 days ago. I am used to Days of "Rage", but I can't get used to Noa's eyes and smile, which are haunting me for 11 days:

10.21.2022 2:27am - (Replying to @jeremy1174) I thought you understood what Zionism means: Coexistence as equally indigenous peoples. The disrespectful word "entity" tells me that you don't.

10.21.2022 2:14am - (Replying to @jeremy1174) Nothing would please me more than seeing our Palestinian neighbors greet this Ethiopian boy as a comrade and close kin. That is precisely what my grandfather expected from our cousins.

10.21.2022 1:56am - The only thing missing from the video is whether this man listens to Al-Jazeera or Mehdi Hasan on @MSNBC

10.21.2022 1:48am - Family lore has it that, when my grandfather arrived at Jaffa port, in 1924, he likewise fell down and kissed the ground. This makes me a comrade and close kin of this Ethiopian boy, transcending all age and background differences.

10.21.2022 1:41am - (Replying to @RaulMachadoG @soboleffspaces and @ObservStudies) The criterion: "50% chance that the claim is true," sounds statistical until we realize that the "claim" itself is causal, and renders statistics totally helpless when it comes to legal responsibility.

10.20.2022 6:51pm - (Replying to @AviMayer) No need to get upset, it's just a matter of semantics. When I first heard (in 2000) that the land my grandfather bought in Bnai Brak (1924) was "stolen", I realized that "stolen" in Arabic means "cultivated". Thus, Hadera swamps, Tel-Aviv, Falafel, Humus, Debka, are all "stolen".

10.20.2022 5:41pm - Great question. I did some thinking about it. See Section 2.8 on "Statistical Time" here: Especially last paragraph. Also, Section 11.2 in, on "Reversing Statistical Time".

10.20.2022 2:24pm - (Replying to @AI_AM001 and @isaacdecastrog) To the best of my knowledge no one denies Iran the right to exist. Enjoy.

10.20.2022 4:57am - (Replying to @AleksanderMolak and @rasbt) That's why I prefer toy problems; we know the answer in advance and we can test our heuristics, plus we theorem of impossibility. For example, there cannot be an "implicit world model" if the user has not provided one. Causal information is a conserved quantity.

10.20.2022 4:15am - Journalism Award Stripped from Palestinian Journalist Over Antisemitic Facebook Posts
10.20.2022 4:15am - Truly Amazing! And I was almost going to give up on the integrity of the Journalistic community.

10.20.2022 3:18am - My take-aways from Rubin's and Heckman's interviews @observStudies match precisely what I anticipated, as expressed in my interview, even though it was completed 2 years ago, and we were not permitted to see each other's texts.

10.20.2022 1:08am - (Replying to @Claire_V0ltaire) We have also seen Jewish scholars expressing their irritation, some of them inadequately: @michaeldickson @Ostrov_A @EinatWilf @CotlerWunsh @SSI_Movement @BarbraStreisand @SSIUCLA @vardi @CAMERAonCampus @bariweiss @JGreenblattADL @deborahlipstadt

10.19.2022 11:12pm - (Replying to @HenMazzig) Would you object if I were to replace "antisemitism" with "anti-Zionism", and say out laud: Anti-Zionism isn’t okay, period. You see, Hitler gave a bad name to antisemitism, so people (including antisemites) would rush to join your statement; only the decent would join mine.

10.19.2022 10:18pm - (1/3) I beg to respectfully disagree with Waxman's 2nd interpretation of Trump's statement, the "Inversion of Dual Loyalty", according to which saying that American Jews SHOULD care about Israel amounts to making "antisemitic assumptions". Americans are mature enough to understand
10.19.2022 10:18pm - (2/3) that Jews care for and support Israel by their very nature and that Jews with spine support Israel openly and proudly, unhindered by fears of dual loyalty accusations that antisemites often voice against them. That of course doesn't excuse Trump of the 1st danger Waxman
10.19.2022 10:18pm - (3/3) correctly identified: trafficking with antisemitic stereotypes that his followers like to hear, which does encourage antisemitism, inexcusably so.

10.19.2022 3:06pm - (Replying to @KirkDBorne) I was not familiar with Barber's book, fairly impressive! And I truly love the quote from "Pearl believes that current AI is mostly just glorified curve fitting, and that adding causal intelligence to current AI will be necessary ...". Amen!

10.19.2022 2:25pm - (Replying to @AmandaKMontoya @b_snef and @rlmcelreath) So, UCLA Psyc dpt has a future after all - best news I got from campus in months. Hoping our paths cross.

10.19.2022 3:38am - Who said computer science is for nerds only?

10.19.2022 3:12am - (Replying to @Claire_V0ltaire) @AviMayer , you are the last person I know whom @mehdirhasan has not blocked yet. So, please answer his question for me: No! Mehdi, you should have disqualified yourself from talking anti-semitism, your persistent delegitimization of Israel, renders you blatantly incompatible.

10.19.2022 1:26am - We’re Jewish Berkeley Law Students, Excluded on Campus via @thedailybeast
10.19.2022 1:26am - From students mouths, this article describes Berkeley Law School as "State Sponsored Humiliation." No Jew is excluded, of course, but they must sit at the end of the bus. Dean Chemerinsky should acknowledge that: (1) "legal" doesn't make a "right", (2) "wrong" is his to undo.

10.19.2022 12:05am - I love this quote from Imbens and Rubin, because it reflects the attitude of the entire PO culture. It says: "Some people find graphs useful in their applications but, in our applications (read: everything on earth), we found them "unhelpful"." That's what PO worshipers repeat.

10.18.2022 11:48pm - (Replying to @soboleffspaces @prem_k and 7 others) Rubin began his career at Educational Testing Service. Imbens and Rubin book (2016) is titled "CI in Stat. Social, Behavioral & Biomedical Sciences" (read: everything on earth), where "we have not found this [graphical] approach to aid drawing of causal inferences" (p.22). Nada!!

10.18.2022 10:20pm - (Replying to @AmandaKMontoya @b_snef and @rlmcelreath) Great, we should meet one day. Do you teach causal mediation/moderation (e.g.. or stick with Barron/MacKinnon/Hayes ?

10.18.2022 10:09pm - (Replying to @HumanBeingIHope) Equating "Giving a Voice to the Palestinian People" with "Me and Hitler are one" makes a mockery of that voice, monsters of the Palestinians, and thoughtless of @thomsonreuters

10.18.2022 9:57pm - (Replying to @robwilliamsiii) I have been influenced by many philosophers of science from Locke and Hume to Suppes and Reichenbach. But I would not call myself a card-carrying philosopher, because that would behoove me to toe the line and refrain from professing that Suppes and Reichenbach were wrong.

10.18.2022 9:41pm - (Replying to @AmandaKMontoya @b_snef and @rlmcelreath) Please do. Are you at UCLA Psych? Normally they don't hire causal inference folks, unless they have had a change in tradition.

10.18.2022 8:39pm - (Replying to @NewsHour and @johnyangtv) Most people think that Mehdi Hasan and Linda Sarsour are using this sudden Jew-caring rhetoric to stain Trump with antisemitism. Wrong! Their aim is to stain Israel with Trumpism.

10.18.2022 6:23pm - I'll be happy to act as an information-dispatcher of causal inference related job openings, like the one you posted here: . My only requirement, the post must mention "causal inference" explicitly and credibly. (There are many impostors around).

10.18.2022 6:02pm - "Me and Hitler are one" [Reuter's honoree]; and I used to place trust in the journalistic community.

10.18.2022 2:24pm - Another blessed feature in your post: You resisted the temptation of including outdated books that claim to be causal, but are graph-tabooed, e.g., Imbens-Rubin (2016), Angrist-Pischke (2014), Gelman-Hill (2006). Why outdated? See:

10.18.2022 3:41am - (Replying to @prem_k @soboleffspaces and 7 others) My mind does not retains applications, but I am sure you can find many works in education in the lists of #CI works I post here weekly. Perhaps some readers can also direct you to such works.

10.18.2022 3:32am - (1/2) I like your blog post I can see how a new sailor, confused by the ever expanding terminology, would find use for your map, as well as the 6 causal books described, so as to sail straight to the target research question at hand or to a new buzz-word
10.18.2022 3:32am - (2/2) one is likely to encounter in the literature.

10.18.2022 3:15am - Great picture of Haifa. Imagine studying 4 years in this city - who can study Maxwell's Equations if you can look down the balcony and watch this bay?

10.18.2022 3:05am - Australia finds itself in a strange company. If any of my decisions was praised by Hamas and Islamic Jihad I would think twice if I am on the right side of history. Australians will.

10.18.2022 1:45am - (Replying to @HenMazzig) What they fear most is your disruption of their mythical narrative of how Muslims and Jews have been living together as good neighbors, for thousands of years, until - you guessed - until the State of Israel was established in 1948. (My wife was from Baghdad).

10.18.2022 1:36am - Most people think that Mehdi Hasan and Linda Sarsour are using this sudden Jew-caring rhetoric to stain Trump with antisemitism. Wrong! Their aim is to stain Israel with Trumpism. @Ostrov_A

10.17.2022 11:30pm - I read at least three of them -- really good!

10.17.2022 7:49pm - I must confess there were two significant differences from the time I was took this bus. (1) I did not have a helmet - unheard of. (2) We weren't allowed in the back door; the driver insisted on watching our gleaming faces as we marveled our power to stop a bus anywhere in town

10.17.2022 3:47pm - Another Monday - another flood of papers on causal inference for your consideration:

10.17.2022 3:28pm - (Replying to @HenMazzig) Stay the course, Hen, and may your story illuminate the world with a wider understanding of what Israel has accomplished that her predators wish to suppress. From strength to strength.

10.17.2022 2:56pm - (Replying to @Eve_Barlow @AvivaKlompas and @MehdiHasanShow) Me too! How about starting a "Mehdi Hasan Me Too" movement? Membership would then become a litmus test for things he can't tolerate --- minimal understanding of the Middle East.

10.17.2022 1:55pm - (Replying to @MissDiagnosis) Great! Let me know I am willing to advise them on how to resolve the Israeli-Palestinian conflict. The US State Dpt. does not heed my advice, perhaps the Saudis will.

10.17.2022 6:36am - (1/3) A third paper with attractive title is: Models, laws, and evidence – philosophy of causal research in macroeconomics, PhD dissertation at the University of Warsaw. It is refreshing to see a treatise on the philosophy of macroeconomics that is not
10.17.2022 6:36am - (2/3) a priori handcuffed by allegiance to the Potential Outcome framework, or to the quasi-experimental paradigm of Angrist and Imbens, but explores the foundations of structural equations. Disappointing is the chapter on "ceteris paribus" which fails to reach the First Law, and
10.17.2022 6:36am - (3/3) the one on "probabilistic causality" which, as I've noted several times", is an outdated philosophical dead end. @causalin #Bookofwhy @PHuenermund @VC31415 @eliasbareinboim

10.17.2022 5:25am - Another title that caught my attention is: On Explaining Confounding Bias Again, I am unable to understand the problem, now due to absence of a graph to specify what we assume, what is observed, what we suspect, and what kind of "explanation" we seek.

10.17.2022 4:00am - I would take very seriously what Dr. Qanta Ahmed is telling us about the transformation of #SaudiArabia. Qanta is not a PR agent; she is a thoughtful critical observer. Do we see any progress in Saudi-Israel relations?

10.17.2022 3:41am - (1/2) In examining the causal inference papers published last week, I was attracted to one titled DOMAIN GENERALIZATION - A CAUSAL PERSPECTIVE, which offers a survey of causal domain generalization models. Sadly, I could not quite figure out the scope of
10.17.2022 3:41am - (2/2) the ideas surveyed, and how they are related to the theory of transportability and data-fusion. While the paper invokes graphs, d-separation, do-operators and SCM's , it is totally oblivious of selection-diagrams and their complete implications. An enigma -...??? #ML

10.17.2022 12:38am - Watch two alligator's lamenting a recent rise in toddler-snatching on the river banks. "We all have to condemn it" says junior gator. "Yes, of course," says gator senior, tears in his eyes, huge tears in his eyes....

10.16.2022 11:43pm - Watch the Israel in which I grew up. So glad it is still there, and that the bus driver still stops when I wave.

10.16.2022 8:26pm - @eliasbareinboim noted an oversight in my interview with I said that Twin Networks are not really needed, because we can derive “conditional ignorability” from the DAG. This is true for Rung-2 tasks; Twin Networks are still needed for Rung-3 tasks.

10.16.2022 8:03pm - (Replying to @AnnGreenberg @Stanford and 2 others) There is also the distinction between "other" as "not like any of us", and "other" as "unique, special, interesting and valuable like any of us". The whole idea of celebrating "diversity" rests on the latter interpretation, which EDI officials deny Jews, let alone Zionists.

10.16.2022 5:54pm - (Replying to @AnnGreenberg @Stanford and 2 others) I must say that my son Danny never complained about anti-Jewish climate at Stanford. But things have changed since he graduated (1985), see Against this background, it's crucial to have inspired-inspired Jewish leaders deal with Stanford's EDI officers.

10.16.2022 3:16pm - (Replying to @JamesA1102 @SarahKSilverman and @JewishJournal) Of course it doesn't, it only has "Zionist-Free Zones", true. But seeing the distinction as a reason to relax causes much greater damage.

10.16.2022 2:55pm - "Task forces" and "Advisory committees" about Jewish High Holidays don't change campus climate. What does? Zionist representation on the students' board for "Equity, Diversity and Inclusion". @Stanford @HillelStanford @aykelman

10.16.2022 2:05pm - White House ‘deeply disappointed’ after Abbas tells Putin he doesn’t trust US via @timesofisrael
10.16.2022 2:05pm - Abbas: Putin should bring us peace. Biden: Deeply disappointed

10.16.2022 6:20am - (Replying to @blakeflayton) Some campuses need Israeli and Zionist speakers to come and cause some trouble, nothing but "trouble" will shake those campuses from their racist anti-Israel climate. It was 85 Zionist "trouble-makers" that shook up USC into semi-soberness More are needed.

10.16.2022 5:46am - "From the river to the see - Iranian women will be free" : A new chant for Berkeley Law School students. [Sorry, Zionists can join the chant on weekends only].

10.16.2022 5:34am - It's gratifying to know that @Twitter is considering requests to look into the content of Khamenei's tweets. He should be notified that "saber rattling" is what keeps his account active; he, Putin and Abbas needs to do more of it to keep the world safe.

10.16.2022 3:50am - (Replying to @DaveEvansPhD @MaryAnnMBates and @rglenner) I recently heard someone say: " How do we adapt evidence of what works from one part of my country to another? No problem, we have a complete formal theory for such questions!" Do you think he is exaggerating? See, and #Bookofwhy

10.15.2022 7:03pm - (Replying to @808RAC and @questionsin2014) You must have someone else in mind; the new Jew does not cry, nor whines: "antisemitism"; the new Jew calls you a "Zionophobic racist", and waits to see if you absorb the new title, enough to look at the mirror and internalize the ugly face you see there..

10.15.2022 6:47pm - (Replying to @ramone_jonah @SensiblePoliti9 and 2 others) You can critique Zionism as much as you wish, and critique justice and freedom and science and commonsense - free speech. But don't be surprised if in doing so you will be called a racist, of the first degree; racists too have a right to free speech.

10.15.2022 6:22pm - (Replying to @thingy665 @SensiblePoliti9 and 3 others) I must mute you, sorry. Anyone using the term "apartheid" on Israel confesses to be vulnerable to populist slogans, to having zero critical thinking, and to aiding the propagation of conspiratorial theories. I can't see myself aiding this propagation, sorry.

10.15.2022 6:09pm - (Replying to @808RAC and @questionsin2014) You want to test me on history? Fact and figures please, no versions, no opinions. I'm ready.

10.15.2022 6:02pm - (Replying to @yudapearl @thingy665 and 4 others) Lifting the Arab siege of Israel means co-existence, lifting Israel's siege of Bazza means the end of Israel.

10.15.2022 5:57pm - (Replying to @thingy665 @SensiblePoliti9 and 3 others) Israel has been under Arab siege for 75 years.

10.15.2022 5:49pm - (Replying to @thingy665 @SensiblePoliti9 and 3 others) The Brits tried very hard to understand both sides so that they could justify their Mandate. But they had to give it up. Why? Bevin explained why. Anyone saying "apartheid ethnostate", not pausing to ask "don't I sound regurgitating?" deserves to be told how it sounds.

10.15.2022 5:17pm - (Replying to @thingy665 @SensiblePoliti9 and 3 others) Palestinian actions, thoughts, dreams, talks, writings, identity, and mentality are aimed at one and only one objective -- the elimination of Israel, even at the expense of their own independence and their own freedom. (It's not my observation but Ernst Bevin, UK FM (1945-51))

10.15.2022 1:49pm - (Replying to @SensiblePoliti9 @thingy665 and 3 others) Not only Hamas. Follow every Palestinian "intellectual", from Nora Erakat to Sari Makdisi, you would not find a single one who is not preaching the dismantling of Israel -- stripping 7 million refugees from their freedom, dignity and right to self-defense.

10.15.2022 1:39pm - (Replying to @ramone_jonah @SarahKSilverman and @JewishJournal) I never conflate the two, because I believe anti-Zionism is an uglier form of racism than antisemitism.

10.15.2022 1:32pm - This link is indeed the best way to get a panoramic view of the project, and access to all interviews and commentaries.

10.15.2022 12:22pm - (Replying to @soboleffspaces) I am not even sure that, in this case, you can eat with your hands.

10.15.2022 12:18pm - (Replying to @markmelvin0123 @Brett297758541 and 2 others) I wasn't referring to Silverman's post. I was reacting the way every rational person should react to someone saying "the Zionist apartheid state of Israel" as if it had a grain of validity. I didn't realize you merely quoted it from someone else. Sorry.

10.15.2022 11:59am - (Replying to @FJnyc and @TimesofIsrael) It's a must read for all Reform Rabbis in America, especially the one's that do not recite the "Prayer for Israel" on ideological grounds, and thus deprive their parishioners of the strength and inspiration that Israel bestows upon world Jewry.

10.15.2022 11:48am - (Replying to @AnnGreenberg and @EricRWeinstein) The #Bookofwhy actually supports the slogan "Correlation is not causation", it highlights however what assumptions must be made if we wish to draw causal conclusions from correlation.

10.15.2022 5:18am - You will see of course areas of fierce disagreements among the 4 interviewees -- I hope you side with me. If not, by all means, bring up issues for discussion and I will try my best to elaborate and clarify. #DataScience #economics #Bookofwhy #Econometrics @causalinf

10.15.2022 5:00am - (Replying to @PatientStormDoc @RWJE_BA and 19 others) To simplify, suppose the "guessed threshold is "blood pressure > 150" . What could be "fake" here that would prevent us from modify it to "blood pressure > 160", and test its predictability of the output of interest.?

10.15.2022 4:41am - (Replying to @PatientStormDoc @RWJE_BA and 19 others) You have an event (i.e., a collection of thresholds on various measurements) that does not predict well the outcome of interest. Why not just search for another event that does predict well? I do not see the fakeness.

10.15.2022 2:28am - Evidently, there are two Chinese translations. One is a Taiwanese version, and the one below is "simplified Chinese," published more recently. Understandably, I am unable to recommend one or the other, and I hate to start a military conflict in the region.Enjoy.

10.15.2022 1:48am - The Journal Observational Studies has just published interviews with 4 causal inference contributors, Don Rubin, Jamie Robins, James Heckman, and humble me. Here's a link to my interview: The other 3 can be found via Obs. Studies. Ian Schrier is moderating

10.15.2022 1:25am - For readers waiting for the Chinese translation, here is a photo of the cover.

10.15.2022 1:02am - (Replying to @lawrence0321bk) Thanks, Laurence.

10.15.2022 12:58am - (Replying to @disruptiveleigh @SarahKSilverman and @JewishJournal) Almost so. Unfortunately, Palestinians do not have a system where they can vote on such questions, and Israelis must rely on what they hear from leaders, textbooks, TV anchors, clerics, etc. Thus far, 100% rejection of co-existence.

10.15.2022 12:52am - (Replying to @markmelvin0123 @SarahKSilverman and @JewishJournal) I apologize if I (indirectly) implied that people who use the term "the Zionist apartheid state of Israel" are racists or morons. You taught me something I wasn't quite aware of, that many use this phrase not realizing how fallacious and offensive it is. Thanks.

10.14.2022 11:34pm - Correcting my mistake, this is the Japanese translation, not Chinese. I know that the Latter has also appeared this year, but I can't find the picture. Please forgive. I can't wait for the Hebrew translation, which would complete my return to my Asian roots.

10.14.2022 11:11pm - (Replying to @soboleffspaces @lyubyvy and 4 others) My point was that every pair (V1, V2) in the graph can be causal, not only one pair X--->Y

10.14.2022 11:05pm - (Replying to @matsdeland @SarahKSilverman and @JewishJournal) You keep using "a different political view" not realizing that for many human beings, identity is more than a "political view" and that Zionism, for most Jews is a matter of identity. Corollary: anti-Zionism is no less despicable than Islamophobia or white supremacy.

10.14.2022 10:52pm - Thank you Chris. Indeed, today I've received a copy of the Chinese translation of #Bookofwhy. Amazing, they still write top-to-bottom, including equations, and the pages go right to left, like Hebrew! NYT: Breaking News - China has joined the Causal Revolution! Beware!

10.14.2022 10:04pm - (Replying to @matsdeland @SarahKSilverman and @JewishJournal) You will never catch me saying "the anti-semitic BDS movement", because I do not believe BDS deserves the title anti-semitic, it is much more dangerous, racist and openly genocidal. See why:

10.14.2022 9:55pm - (Replying to @markmelvin0123 @SarahKSilverman and @JewishJournal) Anyone who can Tweet the phrase "the Zionist apartheid state of Israel" and not pause to ask himself "don't I sound stupid, regurgitating populist slogans like a moron?" deserves to be told how that phrase sounds.

10.14.2022 6:55pm - (Replying to @LuisaBorrell) I prefer complete articles on Tweeted snippets.

10.14.2022 6:53pm - (Replying to @winemaven7181) Enough will be Enough if and only if faculty gain the courage to say "Enough is Enough", as they have done at USC

10.14.2022 6:49pm - When I see a new paper on generalizability, my first question is: Can it be done without selection diagrams? My past readings on this important problem taught me that selection diagrams are a must. Am I exaggerating?

10.14.2022 6:34pm - (Replying to @Andrew___Baker @artistexyz and 5 others) Joking or asking???

10.14.2022 6:30pm - (Replying to @PatientStormDoc @RWJE_BA and 19 others) What I find hard to understand is the concept of "fake measurements". After all, measurements are observable, so they are not fake in themselves; their interpretation may be fallacious but what you measure is solid.

10.14.2022 6:27pm - (Replying to @Andrew___Baker @artistexyz and 5 others) I am not sure if you are mocking or asking. Sorry for asking, we have had a few mockers here before.

10.14.2022 6:20pm - (Replying to @soboleffspaces @lyubyvy and 4 others) Agree, but I believe people tend to ignore bullet (2) because they can't complete bullet (1), i.e., they do not have a meaningful representation of their assumptions, so they can't act upon them. Watch how PO folks dance around ignorability assumptions that they can't understand.

10.14.2022 6:15pm - (Replying to @lyubyvy @soboleffspaces and 4 others) There is no issue if you do things correctly, understanding that propensity score does not reduce bias in itself, it is merely a convenient estimator . I blame the PO culture for creating persisting myths around PS, see and

10.14.2022 6:06pm - (Replying to @soboleffspaces @Andrew___Baker and 5 others) I would only generalize it by saying: "identifiable by conditioning or some other way". To include front doors and other do-calculus doors.

10.14.2022 6:02pm - (Replying to @UtkarshaniJ) I don't understand they question. What do you expect "benchmark" to provide you that theory can't?

10.14.2022 3:37pm - Readers noted that it's unfair to call Putin and Abu Mazen "dreamers", when the former is actively engaged in a military implementation of a "One-State" solution, and the latter merely pays his disturbed youngsters to keep the "One State" dream alive. I apologize.

10.14.2022 8:58am - (Replying to @danilobzdok and @JoshEwen) Not sure John Ioannidis read #Bookofwhy, so here is a conceptual introduction to "single subject prediction":

10.14.2022 8:46am - (Replying to @lyubyvy @artistexyz and 4 others) The Causal ML philosophy is (1) Decide what you want estimated and what assumptions you are willing to make, (2) Find out whether it is estimable and how, Finally (3) Estimate it by the best methods available, be it Tree, Forest or Jungle.

10.14.2022 8:19am - We call these Doctors: "Quasi-experimentalists"

10.14.2022 8:04am - Two of the world's most creative dreamers of a "One-State-Solution." What would they dream of next?

10.13.2022 5:35pm - For readers who are into AI and Big Data, I'm going to give a talk at your Congress, Oct. 25 See you there, through Zoom.

10.13.2022 4:52pm - (Replying to @kareem_carr @PWGTennant and @JohannesTextor) Not so much of a "push back" but lots of "we can do it the old fashion way" or "causal inference is a missing data problem" or "it's bad science". See interview with David Hand

10.13.2022 3:23pm - (Replying to @franuniovi and @kareem_carr) It seems that in the days of powerful software the trapping becomes more intense.

10.13.2022 3:06pm - Remarkably, British factories are still making this kind of brain.

10.13.2022 8:54am - (1/2) Dean Chemerinsky has taken a major step forward by stating, in first person, why the new Rule is offensive to him personally - "because it denies the existence of the state of Israel, the historical home of the Jewish people.” What he hasn't done is to treat this
10.13.2022 8:54am - personal offense as an attack on his identity as a Jew; a category deserving no less protection than race, religion or sexual orientation. Here we are not talking "legalities" but protection in the court of ethics and values, which he is in charge of overseeing. @bandlersbanter

10.13.2022 7:20am - The warpage of the human brain may soon exceed its creativity, which renders AI systems exceedingly timely and urgently necessary.

10.13.2022 2:17am - My emphasis would be on the word "equal" in the last sentence. It leads us to the conclusion: The way to fight antisemitism is through Zionism - the Jewish quest for equality.

10.13.2022 12:26am - I just subscribed to Blacklisted

10.13.2022 12:05am - (Replying to @mdrealresults and @Eve_Barlow) There is no question that the attacks are more vicious and mainstream than they have ever been. True. Yet the awakening to the new Zionism is likewise more energetic, assertive and enthusiastic than it has ever been, perhaps since the Six Day War. Example:

10.12.2022 11:05pm - Strange, my friends are going out of their way to tell me that they can't stand the stupidity of anti-Zionists, and how the recent attacks on Zionism only make more people understand how noble a movement it is. Shall we exchange friends?

10.12.2022 6:32pm - (Replying to @artistexyz @Susan_Athey and 3 others) This is the standard strategy of Rubin's camp economists: assume away identification and estimate through random forests. What is special about Uber?

10.12.2022 3:12pm - (Replying to @soboleffspaces) The absence of a certain substance in itself is just an event, but saying "the absence of substance S1 leads to reduction of substance S2" is a description of a causal relationship that exists in (model of) reality; i.e., existence of a non-trivial function in the SCM.

10.12.2022 8:24am - Leading New York Reform rabbi launches push against anti-Zionism in the movement via @timesofisrael
10.12.2022 8:24am - As expected, recent events are causing a healthy shake-up of American Jewry, followed by a sober realization that only Israel, peoplehood and Zionism can save it from a content-less, spine-less death. Watch the video, and you will see a reform movement redeeming itself. Miracle!

10.12.2022 5:05am - (1/3) To find the primitives of knowledge representation I believe we need to examine the vocabulary with which scientists communicate, because effective communication behooves the speaker to target directly the primitives of the listener, there is no time for translation. To this
10.12.2022 5:05am - (2/3) end, I would refer readers to the Appendix of this paper, which contains a text of ordinary scientific discourse, and demonstrates that bodily functions are described in terms of “absence” or “presence” of certain agents, not by the means through
10.12.2022 5:05am - (3/3) which those agents are controlled or manipulated. I take it to mean that statements about interventions (such as weak ignorability) derive from more fundamental primitives, such as “enabling,” “disabling,” “promoting,” “leading to,” or, more generally, "listening to."

10.12.2022 4:00am - (Replying to @Change4_Better1) Yes, if Western countries signal to the Palestinians that it's time to move on and that future support in their cause is contingent upon addressing the elephant head on, I believe there is some chance for positive movement, though it would take a generation.

10.12.2022 1:23am - (Replying to @Eve_Barlow) Three Jewish leaders addressed the events at the Berkeley Law School (including the Dean) and, in retrospect, I wish they didn't. See why:

10.12.2022 12:56am - @BarbraStreisand asked the same question, but my answer was: And what if it doesn't cross? then what? Does it make anti-Zionism more acceptable? Less racist? Less eliminationist? Less genocidal? Aren't we legitimizing anti-Zionism when we calibrate its evils by its crossings?

10.12.2022 12:33am - (Replying to @kennedytcooper) It is not your Jewishness that Zionists want to erase (this you do quite diligently on your own), but your self-righteous arrogance, thinking that once you raise the banner of anti-Zionism, you cannot be accused of racism. Mistake! Zionophobes are racists of the first degree.

10.11.2022 11:50pm - Zionophobes, like most predators, constantly think of a new way to devour their preys, thinking it's original. No, it isn't. The solution of @Kennedytcooper has been tried by the Brits in 1947, and ended up a total flop - Britain had to quit the holy land.

10.11.2022 11:28pm - We can't afford to silence the voice of @HillelNeuer at the UN Human Rights Council, the only voice of reason in that Lion Den.

10.11.2022 9:52pm - (Replying to @PatientStormDoc @statsepi and 18 others) I tried hard to understand the source of the error in pathologic consensus, but, being totally unfamiliar with sleep apnea syndromes, I could not see the pattern. If you can find someone capable of translating the story to graphical language, I bet remedial tools will be found.

10.11.2022 9:22pm - (Replying to @DluLevine and @JewishJournal) In 2022, I find myself reluctant to expect "people of good conscience decrying this blatant antisemitism". The re-establish of Israel made Jews like me less reliant on the decrying of others, and more reliant on our own actions and resourcefulness. Think agency, not victimhood.

10.11.2022 8:56pm - (Replying to @AMCHAInitiative and @SAFECUNY) When a University President does not spell out the nature of the "alleged offense," and further trumpets a years-old commitment to fighting "antisemitism" (an easily-deflected charge) I begin to suspect his/her commitment to remedial action.

10.11.2022 7:09pm - I often wonder what our cousins lack more; is it an moral lens to see us as humans, or a good lesson in cause-effect logic. We have failed on promoting the former, perhaps we could do a better job with the latter, I have once taught class on distinguishing cause from effect.

10.11.2022 4:26pm - My mother used to tell me that Polish high schools, in the 1930's, made sure that Jewish students were not excluded from any classroom, they just had to sit in the back rows of the class, or stand up when all seats were taken. But, as Berkeley says, 2022, no exclusion whatsoever.

10.11.2022 3:46pm - (Replying to @vineettiruvadi @YiMaTweets and 3 others) I used to teach nonlinear/optimal control in the 1970's and, as much as I try, I cannot see how that powerful theory can help me solve the firing-squad problem in #Bookofwhy p. 39, and answer: "Had soldier-A refrained from shooting the prisoner would still be dead."

10.11.2022 3:31pm - (Replying to @PhilDawid @KordingLab and 6 others) Dressed in interventional notation, weak ignorability is the basis of "Causal Bayes Networks" (see Def. 1.3.1. (i)-(iii), so it certainly has place. Yes, our difference amounts to my considering interventional knowledge to emerge from deeper foundation.

10.11.2022 3:00pm - (Replying to @YoujunHu @YiMaTweets and 3 others) Unless what you are doing is digging deeper and deeper into the rabbit hole.

10.11.2022 2:55pm - (Replying to @HumzahAlkindi) Your position is a dangerous one, for it assumes symmetry where symmetry does not exist. We can't forget Golda Meir's words: If Russia stops shooting there will be peace, if Ukraine stops, there will be no Ukraine.

10.11.2022 2:41pm - I like the title "causal data science", though it is redundant - there can be no "data science" that is non-causal. Next year, the title should read: "scientific data science" to be distinguished from "data-centric data science".

10.11.2022 2:36pm - (Replying to @stephensenn @eliasbareinboim and 4 others) Causal inference does not exclude considerations of variances, as some people assume. They merely add to such considerations the analysis of bias, which pre-causal analysts tend to neglect.

10.11.2022 11:59am - Berkeley’s Jewish-Free Zones Are Worse Than You Think via @jewishjournal
10.11.2022 11:59am - The essence of Berkeley Law action is State sponsored humiliation: Though the first nine rows of the bus are barred to Zionists, Jews should be thankful for getting to sit in the back.

10.11.2022 4:24am - (Replying to @just_whatever and @mehdirhasan) Glad you positioned him for what he is - a Zionophobic racist of the first degree.

10.11.2022 2:56am - @mehdirhasan blocks me, so I can't read what he tweeted on "alleged" and "antisemitism" but, if this is the same Mehdi that I knew, he is many time more dangerous in his Zionophobic coat than some of the worst antisemites that hit the air. We do not need his crocodile sympathies.

10.11.2022 12:13am - The James Ray's among us, @GoodbyePolitics , are beginning to grasp that countries that deny their neighbors' right to exist may end up facing similar fates and similar judgments.

10.10.2022 11:22pm - Sister Candace, Sister @RealCandaceO ! If you honestly want to say the word 'Jewish' "without people getting upset," say it together with these 150 authors as they examine what the words "I am Jewish" mean to them. @Ostrov_A @michaeldickson @Claire_V0ltaire

10.10.2022 9:49pm - I'll stick to my claim, that the issue is not whether these despicable individuals are Antisemites according to one definition or another, but whether they are "racists," a quality that does not require hordes of lawyers/philosophers to be defined and identified.

10.10.2022 7:49pm - (Replying to @rickwsint) Just to clarify (not that it takes away from your argument), the author, J C Pearl is not related to @yudapearl

10.10.2022 7:04pm - (Replying to @KordingLab @PhilDawid and 7 others) This question of how we store our causal knowledge has negligently been ignored by centuries of philosophers and statisticians, while psychologists, as we can see here, are talking "causal DAGs", not experiments.

10.10.2022 3:20pm - Our weekly harvest of new publications in causal inference is swelling A few titles, inviting special attention, will be discussed in the next few days.

10.10.2022 3:02pm - (1/2) (Replying to @PhilDawid @pmddomingos and 6 others) Not so fast. Weak ignorability becomes (semi) "comprehensible" only after rephrased in the language of "interventions" "intended treatments" etc., not as stated in the language of "potential outcomes". It remains incomprehensible when "interventions" cannot be envisioned.
10.10.2022 3:13pm - (2/2) (Replying to @yudapearl @PhilDawid and 7 others) Furthermore, weak ignorability rephrased in interventional terms is "comprehensible" only if one theorizes that humans store causal knowledge in the form of hypothetical experiments; I refure this theory, in favor of causal DAGs, from which results of experiments are deduced.

10.10.2022 2:44pm - This link should work even in thunder storm:

10.10.2022 11:34am - (1/2) Please join me and my family in celebrating Danny's 59th birthday. Many communities celebrate it through musical concerts, interfaith gatherings and press-freedom panels, still others recite his last words: “My father is Jewish, my mother is Jewish — I am Jewish. Back in
10.10.2022 11:34am - in the town of Bnei Brak, there is a street named after my great-grandfather, Chaim Pearl, who was one of the founders of the town.” I've offered my personal reflections here:; 150 other interpretations are collected in the book "I am Jewish"

10.10.2022 3:55am - This paper, aside from its relation to Bell's Theorem, should be of interest to economists. It shows that, contrary to conventional wisdom (eg, Inbens+Angrist 1994), Instrumental Variables do have testable implications, labeled "Instrumental Inequality".

10.10.2022 2:08am - (Replying to @skdh) The recent interest in Bell's Theorem reminds me that, in 1995, I derived a simple inequality that generalized Bell's Theorem: (Thm 1). I am posting here just in case a physicist finds it useful. In causal inference it offers a test for instruments.

10.10.2022 12:53am - I would also recommend reading my comment on @DeatonAngus and @NanCartwright , in which I complain about scientists appealing to abstract notions of "triangulation" while ignoring algorithmic triangulation.

10.10.2022 12:36am - (Replying to @RonKenett @eliasbareinboim and 5 others) Transportability theory takes "interactions" to be omni-present, hence it can't be less general that theories that take "interactions" to demand special considerations.

10.9.2022 11:42pm - (1/2) This is one of the best written paragraphs on generalizability, starting with: "Science is about generalization, and generalization requires transportability. Conclusions that are obtained in a laboratory setting are transported and applied elsewhere, in an environment that
10.9.2022 11:42pm - (2/2) differs in many aspects from that of the laboratory. We all understand that if the target environment is arbitrary, or drastically different from the study environment nothing can be learned and scientific progress will come to a standstill. ..." Worth RT-ing weekly!

10.9.2022 4:02am - As violence and tragedies intensify in the holy land, lonely voices for sanity may get a chance to rise above it all and to be heard. The honest confession of this young Palestinian is one such voice, listen:

10.9.2022 2:09am - We've heard of similar experiences from many economists on this twitter platform, mostly from researchers who were disappointed with how causality is treated in economic textbooks and how isolated the field is from neighboring disciplines. Join us.

10.9.2022 1:08am - With a broken heart, I send condolences to the family and friends of SGT Noa Lazar who was killed by a terrorist last night, while defending my family in Israel. May her memory be a blessing.

10.8.2022 7:29pm - (Replying to @ThomSeaton @JamesEBruce and @Brahmski) Happy 5783 to all Am Israel. It is scary to think that I was born only 7,696 years after the Big Bang, and I can hardly recall major differences from our life today. Oh, yes, we cooked on Kerosene oil.

10.8.2022 7:18pm - (Replying to @PhilDawid @pmddomingos and 6 others) I took another look at your dimensions and, while I agree with your critics of PR (potential response/outcome) I do not agree with your lumping SCM together with PR. SCM does not suffer from PR's main weakness: demanding ignorability assumptions that no mortal can comprehend.

10.8.2022 5:47pm - (1/2) (Replying to @CardinalNichols) Your Holiness, You may be a great theologian, but I have invested a few good years studying causes and effects and, on that basis, I must contradict your prediction and state that UK moving its embassy to Jerusalem would hasten, not impede a 2-states-based peace; a
10.8.2022 5:55pm - (Replying to @yudapearl and @CardinalNichols) a necessary ingredient of such peace is mutual respect of, not denial of each side's connection to Jerusalem. Your arguments support denial, not respect. Hoping theology does not preclude reconsiderations. Respectfully, A British subject at birth,

10.8.2022 4:47pm - I welcome @michaeldickson 's followship, and happy to RT @RozRothstein 's article. The one thing I would add to it: Jewish ties to the holy land are not based on Biblical fables (as mocked by Zionophobe Roger Waters) but on collective memories that are deeply etched in the
10.8.2022 4:47pm - minds and souls of living Jews, memories that continue to govern and guide the intellectual activities of breathing Jews, their language, their humor and their creativity. Lacking such ties prevents our cousins from understanding the futility of their rejectionism.

10.8.2022 3:45pm - (Replying to @UnderwaterBepis) Welcome to causality land. So, you want to start someplace else, not assuming a DAG. This is possible; you can start with DAG-equivalent assumptions, or you can start with data and try to discover a DAG, etc. What DAGs tell you is where you cannot start, and what to look for.

10.8.2022 8:41am - (Replying to @ChristinABuelow) I have never suffered from this problem, perhaps because my to determination to naively pursue what appears to be "the most important problem in existence" has always produced publishable results. I even got tenure with this naïve approach.

10.8.2022 8:20am - (Replying to @ChristinABuelow) I am grateful to you for telling me that #Bookofwhy is helping readers better understand ourselves, and the world in which we live. I hope the book continues to thrill you, as you proceed backward, from Chapt. 10 to Chapt. 1. And don't let the harsh truth stop you.

10.8.2022 7:41am - (Replying to @KouMurayama @artistexyz and @georg_henning) I hope this paper convinced you. What's the point in saying "causal inference" is a "missing data" problem, if we do not have a calculus for the latter?

10.8.2022 7:34am - (Replying to @Ostrov_A @KombatKlutch and 13 others) I've only used it twice; once on a statistician that kept saying "you are wrong" with no content. And now, on a Zionophobe that started making up stories "when I went to tel-aviv".

10.8.2022 7:26am - (Replying to @spikypotatoface @ShaharAzani and 13 others) I think you are underestimating the gullibility of good people, especially in the West. This is how we got the "apartheid" culture to propagate on college campuses by innocent well-meaning students. I simply don't have the time to ridicule every stupid allegation on this screen.

10.8.2022 7:07am - (Replying to @KombatKlutch @BecomeOblivion and 13 others) Apologizing to other readers on this thread, but I must mute Mr. @KombatKlutch . [My 2nd mute in the past 4 years]. When malicious lies are manufactured in a coat of "when I went to tel aviv" they become palatable and I can't allow this platform to assist their spreading. Sorry.

10.7.2022 6:40pm - (Replying to @PhilDawid @pmddomingos and 6 others) Glad we are converging. The only issue left is to agree along what dimension we should judge one language more "suitable" than another. I have suggested several such dimensions in, but am willing to examine others.

10.7.2022 6:32pm - (Replying to @Bijun_med) Merry sailing. And share with us the joys of learning.

10.7.2022 9:26am - (1/2) Interesting, but in our corner of the woods we have perfect consensus about what generalization means and how to test it. Moreover, our understanding of its scope and limit is in some sense complete. Perhaps the NLP discord can be alleviated by
10.7.2022 9:26am - (2/2) examining the approach we took to the problem of generalization and the method of analysis we applied towards its solution. Some of it, I am fairly sure, may be transferable to NLP. @eliasbareinboim @GaryMarcus

10.7.2022 6:31pm - (Replying to @pmddomingos) Not enough, sure. But if we agree that "generalization" is necessary for AI then it makes sense to go to a peaceful corner of the wood where generalization was solved, and see if we can borrow a few tricks.

10.7.2022 8:36am - I join UN Watch in calling on Secretary Antony Blinken @SecBlinken to oppose the appointment of Francesca Albanese as a special rapporteur on Palestine. @HillelNeuer @Ostrov_A

10.7.2022 3:07am - I can't lift my eyes from the facial expression of the rooster on the left. His urge to project virtue and sing louder than other chicken in the barn could be admirable, if it weren't for the tragic end that awaits the thoughtless.

10.6.2022 8:24pm - (Replying to @KordingLab @PhilDawid and 6 others) ?? decipher please ??

10.6.2022 7:29pm - We are waiting anxiously for Al-Jazeera @AJEnglish to describe Ahmed as "settler colonialist," and for Berkeley's Queer Caucus to accuse Israel of weaponizing gender tolerance to hide its unholy conception. See background:

10.6.2022 7:06pm - (Replying to @PhilDawid @pmddomingos and 6 others) I may have misinterpreted what @PhilDawid truly meant to propose, but I hope we agree on the golden rule: To predict the effect of actions we must start with either experiments or causal judgments. Calling the latter "Statistical DT" does not negate its causal content.

10.6.2022 6:53pm - (Replying to @KordingLab @tiagopeixoto and 6 others) I am an odd ball then, b/c I learned Alef-Bet from my mother and how to grow citrus trees from my father, and electromagnetic theory from Maxwell; I couldn't wait for evolution to punish me for not knowing these survival tricks, see

10.6.2022 6:06am - (Replying to @biorankings and @UCBerkeley) Legally, I presume it can. But morally, banning an ethnic group and humiliating the ethnic group by enshrining the ban into Bylaws may not be compatible with what the Bylaws of "Equity Diversity and Inclusion" call for.

10.6.2022 5:10am - A nice gesture by Professor Jeff Lax, to speak to the nine @UCBerkeley law student groups who vowed (in their byelaws) never to listen to him and his like. I'm more inclined to favor the approach of "rescind or else", proposed here: by 150 organizations.

10.6.2022 2:06am - (Replying to @pmddomingos @AlexTensor and 6 others) A semi-classic reference by JP and @eliasbareinboim which shows mathematically (as well as through concrete problems) what kind of information is absolutely needed for generalizing experimental findings:

10.6.2022 12:45am - (Replying to @yudapearl @pmddomingos and 6 others) I must take some time off this interesting thread. Might come back when I understand "inductive bias" in engineering terms, i.e., input-output.

10.6.2022 12:37am - (Replying to @pmddomingos @AlexTensor and 5 others) I would love to see one such assumption which can be articulated without appealing to causal vocabulary. Recall, any such assumption is provided by a human programmer/researcher, so it must be comprehensible to 90% of readers on this Twitter channel.

10.6.2022 12:17am - (Replying to @pmddomingos @KordingLab and 6 others) We agree on the last point: AI will be solved by combining empiricist and rationalist ideas. What the causal revolution has added to this abstract agreement is a framework demonstrating a concrete symbiosis of the two, armed with division of labor and performance guarantees.

10.6.2022 12:09am - (Replying to @KordingLab @pmddomingos and 5 others) Can you describe the input-output properties of the system (judgments, data, conclusions, etc.) BTW, I forgot the existence of my paper, thanks for the reminder. It's worth reading.

10.5.2022 11:57pm - (Replying to @pmddomingos @AlexTensor and 5 others) I have at least proven that (1) to predict effects of action we must have either experiments or causal judgments (2) To process counterfactuals (eg identify causes of effects) we must have functional relations (eg SCM), and (3): Given (1), a complete algorithm exists for Rung-2.

10.5.2022 11:40pm - (Replying to @KordingLab @tiagopeixoto and 6 others) Simulated evolution provides indeed a theoretical possibility of climbing the Ladder's rungs, but it's wisdom and practicality are questioned here: along 4 dimensions.

10.5.2022 11:32pm - (Replying to @pmddomingos @AlexTensor and 5 others) Back to the original question: is causal inference necessary for AI? There are (seemingly) other options. Example: We take causal graphs and call them Neural Networks, we take causal inference engines and call them "inductive bias", finally, we tweet: "Victory, everything is NN".

10.5.2022 11:11pm - In view of some misunderstandings of the boundary bet Rung-1 and Rung-2 of the Ladder of Knowledge, I am RT-ing this thread, hoping to make the boundary crisp and unassailable to all readers: To climb Rung-2, data must be enriched with experiments or judgmental causal information

10.5.2022 11:11pm - (2/2) augmented with "action indexing" variables that mirror the causal information conveyed by the starting graph. "No causes in - no causes out" (Nancy Cartwright, 199?) Notice that I keep referring to "causal information" as a conserved quantity; its representation may change.

10.5.2022 10:37pm - (1/2) (Replying to @pmddomingos @AlexTensor and 5 others) We may differ in what counts as "resolving" Simpson's paradox. For me, "resolving" means explaining how for heaven's sake it's possible for two opposing decisions to emerge from two identical datasets that differ only in the causal structures of their generators. It's called:
10.5.2022 10:47pm - (2/2) The Second Simpson's Paradox in I need to understand the reason for your resistance to the causal resolution of SP, which now provides a general calculus for resolving any problem instance that may be encountered.

10.5.2022 9:01pm - (1/2) (Replying to @pmddomingos @GaryMarcus and 5 others) It is a mistake to interpret Dawid's video as saying that causal inference is not necessary for predicting effects of actions. I've written an editorial comment on Dawid's arguments -- his framework begins with a causal graph, and ends with a graph
10.5.2022 9:11pm - (2/2) (Replying to @yudapearl @pmddomingos and 6 others) augmented with "action indexing" variables that mirror the causal information conveyed by the starting graph. "No causes in - no causes out" (Nancy Cartwright, 199?) Notice that I keep referring to "causal information" as a conserved quantity; its representation may change.

10.5.2022 8:43pm - (Replying to @AlexTensor @pmddomingos and 5 others) Agree. You can't solve Simpson's paradox without causal information, and certainly not with "logic" alone. The linked video calls it "context" - a pre-scientific term used by statisticians to explain why same data dictate two different actions depending on the causal structure.

10.5.2022 8:34pm - Many reciprocal thanks to our Muslim cousins who are reaching out to us through interfaith activities, people-to-people, building a peaceful world for our children and grandchildren.

10.5.2022 8:19pm - Hello All. I hope everyone who was fasting had an easy and meaningful fast. My three fast-breaking stars have been joined by many others, sparkling in the sky to celebrate this unique phenomenon of Jews atoning for errors they may have committed; a peculiar form of free will.

10.4.2022 6:06pm - Yom Kippur holiday is descending upon us. Wishing all readers easy fasting, and a day of deep reflection over those imperfections in our software that we can improve upon. Gmar Chatima Tovah

10.4.2022 5:50pm - (Replying to @Coffefe123 @RoniLitman1 and 16 others) The pre-1967 borders were merely cease-fire lines of 1949, with one side refusing to sign a peace-agreement. That cease-fire was violated by one side, thus renewing the old war and erasing the lines. At this point: recognition first, borders second.

10.4.2022 5:42pm - (Replying to @pmddomingos @StephenPiment and 5 others) If this were the case, we would not have 6 generations of philosophers and statisticians arguing confusedly over Simpson's paradox Interpreting "good for" to mean "more likely" concludes that a miracle drug does exist, vividly supported by data.

10.4.2022 3:38pm - (Replying to @pmddomingos @StephenPiment and 5 others) Not so simple. "good for men" means "action-1 produces good results knowing that Joe is a Man" which requires reasoning about actions and about knowledge. See "Sure Thing Principle"

10.4.2022 3:14pm - (Replying to @pmddomingos @ylecun and 4 others) Trouble is, logic does not gives me: "Had soldier-1 not shot, prisoner would still be dead". We need a logic of "Had" and "Would". I don't find it in logic books, can I find it in "inductive bias" books?

10.4.2022 3:03pm - Is the Tide at Berkeley Beginning to Turn? via @jewishjournal
10.4.2022 3:03pm - What Berkeley's Dean begins to understand is that, in addition to enforcing Rules, it is his duty to set norms of civil discourse. This he can do only by publicly stating the school's principles and values. In particular, that barring Zionists is banning a people. @DavidSuissaJJ

10.4.2022 1:02pm - (Replying to @pmddomingos @ylecun and 4 others) So far I have only seen the words "inductive bias" used to mean "everything that ML folks want to do" which is not very informative. When I see "inductive bias" handle one toy problem (eg, "firing squad" -#Bookofwhy) I'll become "inductive biasist, till then it's "everything".

10.4.2022 12:10pm - (Replying to @pmddomingos @StephenPiment and 5 others) When do we say that something is "a necessary primitive"? When we can't find an alternative. I have not found an alternative way of building a machine that would slap me on the wrist if I find a drug that's "good for men, good for women, bad for a person". No other logic does.

10.4.2022 2:51am - (Replying to @yudapearl @StephenPiment and 6 others) Terrible typo, it should have read: I try to minimize "arguments" and maximize explorations into what is doable and what is not.

10.4.2022 2:41am - (Replying to @tdietterich @Grady_Booch and 5 others) Yuval Harari (Sapiens) connects it to the development of counterfactual thinking and "market for promises". You won't convince a monkey to give you a banana for a promise of infinity banana stream in monkeys' afterlife. People go for it!

10.4.2022 2:08am - Exclusive: Pro-Israel Groups Release a Statement on Berkeley Controversy via @jewishjournal
10.4.2022 2:08am - Breakthrough in the Berkeley scandal. Zionist organizations are fighting back, proudly, assertively and purposefully: "Rescind or consequences!" Who is missing from the list: @ADL , @HillelIntl and, to my great disappointment - faculty groups. Unforgivable!

10.4.2022 1:35am - I lost a comrade in that war, I can still see him smiling. 2,500 such friends. I'll remember them all in Yom Kippur.

10.4.2022 1:25am - This is what happens when @Reuter hires journalists with @AJEnglish training.

10.4.2022 1:20am - As we stand in Yom Kippur prayer tomorrow night, let's remember this service and the fact that the majority of these people soon found themselves on rotting steam ships, only to be re-interned by the British navy, trying to stop them from reaching Palestine. They prevailed.

10.4.2022 12:58am - Thank you Rabbi Wolpe for years of leadership to the community, and for your inspiring words of wisdom in sermons, prayers, and public gatherings. Please continue to guide us from your unofficial chamber.

10.4.2022 12:47am - (Replying to @AlexTensor @pmddomingos and 3 others) I was taught that God main function up there is to answer the question: "Why bad things happen to good people?"

10.4.2022 12:44am - (Replying to @AleksanderMolak @ylecun and 4 others) We know that we are good at some things, and we know that our thoughts are governed by causal logic. Conclusion: there may be some causal relationship between the two.

10.4.2022 12:41am - (Replying to @Abel_TorresM @pmddomingos and 6 others) My goal in causal ai is to have machines conduct human-like conversation which, as we know, is loaded with causal utterances. The formalism is just an attempt to interpret those utterances.

10.3.2022 10:52pm - (Replying to @quasiireastclly @galt_dead and 3 others) There is no such thing as "more than just defending X" as long as Y calls publicly, relentlessly and demonstrably for wiping out X, regardless of borders or circumstances. This is where "right to exist" comes in, it implies permanent existence and "end of claims".

10.3.2022 9:17pm - (Replying to @StephenPiment @pmddomingos and 5 others) I try not to minimize "arguments" and maximize exploration into what is doable and what is not.

10.3.2022 9:14pm - (Replying to @pmddomingos @ylecun and 4 others) I wouldn't know how to start without it, except perhaps fitting data to some complex curves. If anyone can show us how to do any of these seven tasks with "inductive bias" - I'll quit causal inference and become "inductive-biasist".

10.3.2022 9:03pm - (Replying to @ylecun @aminkarbasi and 4 others) The goal of AI is not to talk microphysics but to capture the illusions that make communication with end users feasible and human-like.

10.3.2022 8:17pm - (Replying to @quasiireastclly @galt_dead and 3 others) Is it really so hard to see the relevance? If X has the right to exist and Y is about to undo that existence. Is it really unreasonable for X to try and prevent Y from accomplishing his plan, even if it means entering Y's territory to spoil upcoming attacks?

10.3.2022 7:29pm - (Replying to @Comuniity_ @WickedViper23 and 4 others) The rule that "history is written by the Victors" has one exception: the Palestinians lost in all their attempts to liquidate Israel yet, strangely, they rewrote history so well that even some Western historians believe they were (and are) the innocent, helpless underdogs.

10.3.2022 6:43pm - (Replying to @ShawhinT @Basir7900 and @TDataScience) If you have a link to a pdf version it would be great.

10.3.2022 6:07pm - (1/2) (Replying to @dylanarmbruste3 @mackinprof and @MattMadore576) Yes, I know this "framework". The #Bookofwhy discusses it in a few pages, concluding: "Every one of these criteria has demonstrable exceptions, although collectively they have a compelling commonsense value and even wisdom. " I'm surprised to see a revival of this
10.3.2022 6:12pm - (2/2) (Replying to @yudapearl @dylanarmbruste3 and 2 others) pre-scientific "heuristics" in our era of scientific causality. I hope the authors understand science to tell us why/when heuristics is needed.

10.3.2022 5:06pm - Sharing a video I just received from the LA Press Club, with Rob Eshman presenting the Daniel Pearl Award for Courage and Integrity in Journalism to Markus Yam, of the LA Times.

10.3.2022 3:48pm - If Berkeley's Zionophobic tantrum succeeds in arousing more such wishes, more Jewish pride, and more Zionist pushback, then perhaps it might have positive overall effect, in addition to exposing the total incompetence of University administrators.

10.3.2022 12:40pm - The past week saw the following published articles in causal inference: My attention was drawn to which introduces causal inference, head up, into traditional statistical curriculum - a healthy step forward.
10.3.2022 1:36pm - (Replying to @asknbid) I do not understand the question.

10.3.2022 5:03am - Khamenei: Riots in Iran were planned by Israel. Some of us have suspected it all along, but were silenced of the Jewish-controlled media, and the Jewish Lobby. Finally, @hrw broke the silence, heroically exposed Israel's crimes and made top nominee for Iran's highest HR Prize.

10.3.2022 4:37am - (Replying to @Ostrov_A and @hrw) Most organization retract careless allegations when shown to be baseless; not @hrw .

10.2.2022 11:58pm - (Replying to @thinkyparts and @EinatWilf) I do not believe the DNA is religious. Proof: Christian and secular Palestinians are as adamant about Israel's demise as their Muslim compatriots. DNA may sounds too immutable, agree, I prefer "mother milk," but it takes 8 additional characters.

10.2.2022 11:40pm - (1/2) I wish @BarbraStreisand would respond, and reflect upon the process that made Jewish leaderships regard anti-Zionism to be benign if it does not "bleed into broad anti-Semitism," and to warrant no fight nor alarm before we see signs of "bleeding". What could make one think
10.2.2022 11:40pm - that some benevolent form of anti-Zionism exists which would not, sooner or later, "bleed into broad anti-Semitism" or which could do no harm until bleeding signs begin to discomfort otherwise comfortable Jews. I beg to disagree, @BarbraStreisand @bandlersbanter @Ostrov_A

10.2.2022 10:32pm - (Replying to @thinkyparts) The list of conditions that have existed for fostering such good ooze is longer than the diaspora. (See: Schwartz and @EinatWilf book: The War of Return). The problem is not absence of conditions but presence of a DNA that defines the life of one nation by the death of another.

10.2.2022 9:54pm - (1/3) Some readers responded by asking that Zionists try to regain the high moral ground they once held. It's a fair request. But what higher moral ground can you get than standing up and screaming: Show me one Palestinian leader, teacher, cleric, journalist or TV-anchor willing
10.2.2022 9:55pm - (2/3) to accept Israel into a permanent peaceful co-existence! (Paraphrasing Lapid at the UN, last week). Now, here comes the greatest surprise of your life: Most Americans would say: "I'm sure there are a few Palestinians willing to accept Israel." Surprise: No, there aren't. Sure
10.2.2022 9:55pm - (3/3) you can find many declaring: two states tomorrow. But none would tell their children that Israel is here permanently, by right, a homeland to an equally indigenous people. What parents teach children is the litmus test of whether they can/will honor a final peace agreement.

10.2.2022 4:35pm - (Replying to @parallelmonoga1 @WickedViper23 and 3 others) Separating between those who (loudly) call for your destruction and those who don't is not "appartheid", it's commonsense, even if you decide to treat the former differently from the latter.

10.2.2022 3:49pm - (Replying to @parallelmonoga1 @WickedViper23 and 3 others) The phrase "Everyone knows it" reveals immediately where one gets information from. This happened to be the battle cry of Palestinian propagandists when they run out of arguments and seek help from their echo-chamber of gullible supporters. It started with Arafat, 2000.

10.2.2022 2:58pm - (Replying to @WickedViper23 @parallelmonoga1 and 3 others) I don't know how a country can "take control" of its neighbor's narrative, except of course by documenting and elaborating its own narrative. This is indeed the purpose of re-enacting the "Forgotten Miracle of Nov. 29, 1947":

10.2.2022 2:40pm - (Replying to @parallelmonoga1 @WickedViper23 and 3 others) Good question: "Where I got my education on all this". I happened to be born there and to thoroughly explore, facts and figures, how it started and what it is today. Oh, you forgot to add "apartheid" to the list of populist slogans that gullible observers love to fall for.

10.2.2022 1:56pm - We have had several discussions on Stanford's "Foundation Model", with multiple requests for someone to define/explain what it is. Today I received a newsletter about it, still no explanation, but a mild admission that no one knows yet: @StanfordHAI

10.2.2022 1:17pm - (Replying to @yudapearl @parallelmonoga1 and 4 others) And it seems to me that Palestinians "have not been looking for peace at all," not because they don't need too, but because they view "peace" as a defeat, and Western Zionophobes keep telling them they can achieve victory by continuing to reject peace.

10.2.2022 1:17pm - (Replying to @yudapearl @parallelmonoga1 and 4 others) And it seems to me that Palestinians "have not been looking for peace at all," not because they don't need too, but because they view "peace" as a defeat, and Western Zionophobes keep telling them they can achieve victory by continuing to reject peace.

10.2.2022 1:00pm - (Replying to @parallelmonoga1 @WickedViper23 and 3 others) One of the best description I heard of Palestinian society: "Colonialist mindset, denying indigenous people any rights." Contrast it with the Zionist credo: "Equally indigenous," incarnated in the UN partition vote of Nov. 29, 1947, which Arabs saw (and still see) as a defeat.

10.1.2022 6:03pm - College students! Here is your answer to campus Zionophobes. Mark your calendars and join @IACmishelanu , @SSI_Movement and other student organizations in a nationwide celebration of "The Forgotten Miracle of Nov. 29, 1947"

10.1.2022 5:46pm - (Replying to @nowthatiamiwas and @mynameisjro) Sarter's words would be 12 times more relevant to present day Zionophobes, give or take a few juicy observations of their mentality.

10.1.2022 4:30pm - (Replying to @DaneOrShine @nowthatiamiwas and @mynameisjro) The terms "apartheid tactics toward the Gaza strip" unveil a total disconnect with "the crux of the issue". Language is important. Bye.

10.1.2022 3:17pm - (Replying to @WickedViper23 @parallelmonoga1 and @BarbraStreisand) A minor addendum. I'll listen to complaints when Israel does dumb thing, but only if the complainer considers the unique circumstances under which the dumb thing was done, including having a neighbor sworn to her destruction, relentlessly, in song, verse, actions and DNA.

10.1.2022 9:49am - A rare, almost forgotten document, resurrected from oblivion! Thanks. I remember Gill Gate speaking at UCLA and, invariably, when asked about the future of AI, he would praise the marvels of Bayesian Networks. Too bad he hasn't stayed around to see the marvels of Causal Inference

10.1.2022 9:08am - (Replying to @erikbiz and @BarbraStreisand) I don't believe @BarbraStreisand would think for a moment that there is a shred of legitimacy in opposing the existence of Israel. She is just following the outdated Jewish tradition of preferring to fight antisemitism over protecting Zionism, presuming it carries more weight.

10.1.2022 8:28am - (Replying to @DaneOrShine and @mynameisjro) The moment you utter the word "both" (i.e., homeland of both people) you are a Zionist, and you won't be able to speak at Berkeley. Nothing could be more objectionable to Zionophobes than the Zionist credo: Two states for two peoples, equally legitimate and equally indigenous.

10.1.2022 8:12am - (Replying to @erikbiz and @BarbraStreisand) @BarbraStreisand did not advocate stripping Israel from its sovereignty. She merely tried to protect non-involved selfish Jews from violence aimed at involved Zionist Jews. It's a futile attempt because Israel predators view the former as likely supporters of the latter.

10.1.2022 7:52am - (Replying to @DaneOrShine and @mynameisjro) By listening to what Palestinian intellectuals say, write, teach and preach. Which of the anti-Zionist aims you doubt: to uproot one people, the Jewish people, from its homeland, to take away its ability to defend itself in sovereignty, and to delegitimize its historical identity

10.1.2022 7:29am - (Replying to @DaneOrShine and @mynameisjro) Consider reading

10.1.2022 6:43am - (Replying to @mynameisjro) Feeling strange but happy to agree with a Zionophobe. "Anti-Zionism is inherently not antisemitism". Of course, it is an uglier and more dangerous form of racism.

10.1.2022 5:42am - (1/2) @BarbraStreisand responded to the Berkeley scandal with "When does anti-Zionism bleed into broad anti-Semitism?." Her tweet has evoked many supportive replies among her followers. I beg to disagree with the wholesale focus on the "bleeding", instead of on the moral deformity
10.1.2022 5:42am - of anti-Zionism in itself. This habit of ours to calibrate the evil of anti-Zionism by the degree to which it bleeds into broad anti-Semitism is partly responsible for the Berkeley tactics, where students are hoping to escape the bad name of antiSemitism.

10.1.2022 5:17am - (Replying to @AvivaKlompas) It's like telling Christians their belief in resurrection is disqualifying.

10.1.2022 4:34am - (Replying to @BarbraStreisand) Beg to disagree, @BarbraStreisand . Suppose it DOES NOT "Bleed into broad anti-Semitism", then what? Does it make anti-Zionism more acceptable? Less racist? Less eliminationist? Less genocidal? Berkeley's Zionophobes are taking advantage of our habit to focus on the BLEEDING.

10.1.2022 4:11am - (Replying to @DavidHirsh) This is what I meant in saying: These organisms grow up without a mirror, in a supportive environment that tells them "you are the champions of goodness", and not one mirror to reflect their ugliness. This is also where a "Zionophobic" mirror can help.

10.1.2022 3:23am - For readers in competitive moods, UAI is offering several competitions among machine learning and reasoning systems: May the best man/woman/machine win! "Hail, Science, those who are about to loose salute you" !

9.30.2022 10:22pm - (Replying to @blakeflayton) There is a huge difference between Israel and the war in Ukraine. Putin claims ownership on only four regions, while Palestinians go "from the river to the sea" - the whole cake. Westerners are generally critical of Putin, but find the Palestinian claim cute, if not attractive.

9.30.2022 7:10pm - (Replying to @rustbenisrael and @SDrrizzy) I am well aware of the stories behind The Vote. The real hero was not Truman but his former business partner Jacobson and dozens other Jews worldwide, including of course the Yishuv leaders in Israel who vowed: "We will have a state even if they vote NO."

9.30.2022 6:32pm - I still owe you a piece I wrote in 2008 "The forgotten Miracle: Nov. 29, 1947" Describing some of the stories behind the miracle and why it behooves us remember them.

9.30.2022 6:18pm - (Replying to @AlexTensor @ylecun and 2 others) I am referring to "domain knowledge" also known as "world model". How is it provided, how is it represented, what must the provider think about when providing it, etc etc.

9.30.2022 2:55pm - Here are the color slides: and here is the accompanying video: From Deep Learning to Deep Understanding Happy sailing!

9.30.2022 1:53pm - For readers wishing to see what a November 29 celebration looks like, here is what we did in 2012: But even a more modest event, re-enacting the Vote, reading the UN Resolution, and giving thanks to friendly communities would make for a memorable event.

9.30.2022 1:38pm - I've defined "understanding" here: It is a formal definition, once you tie it with the Ladder of Causation, where the capabilities of each layer is formally defined.

9.30.2022 6:44am - (Replying to @AlexTensor @ylecun and 2 others) This still does not tell me what judgments, if any, are required of the researcher at the onset.

9.30.2022 6:36am - (Replying to @DavidHirsh) Beautiful blacks have "racist" designation. Proud gays have "homophobe" designation. Proud Muslims have "Islamophobia" for hostile designation. Only we, proud Zionists, have no fighting word. This forces us to use "antizionism", which sounds almost as legitimate as Zionism.

9.30.2022 5:56am - (Replying to @MikeNew19512115 and @AvivaKlompas) The 1930's Judenrein Zones were established when Jews stood with trembling knees. Today's Jews can turn these racist enclaves into academic No-fly Zones. All it takes is leadership:

9.30.2022 3:08am - (1/2) My faithful calendar says we are two months ahead of November 29, the day in which history will celebrate the 75 yrs anniversary of its monumental miracle, the UN Vote to partition Palestine into a Jewish state and an Arab state. Two months from now,
9.30.2022 3:08am - (2/2) US cities will be ignited with celebrations, and university campuses will see Jewish and non-Jewish students joining in musical offerings to the Goddess of History, thanking Her for granting one nation what others take for granted - a homeland. @eishsadehy @blakeflayton

9.30.2022 2:28am - (Replying to @GaryMarcus @danbri and @ylecun) My experience has been the same, in the sense of realizing that scaling up without understanding can be disastrous and, simultaneously, that mathematics does have the power of bringing about a state understanding and spotting the pitfalls of would be disasters.

9.29.2022 11:52pm - (Replying to @emilykschrader and @Ilhan) @Ilhan can't get it to her skull that, to most women in the world, especially to Iranian women, Hijab is an icon of "no choice," though she enjoys Western privilege of wearing "by choice". I therefore suggested she take it down for 60 second, in public, just to show solidarity.

9.29.2022 9:57pm - An inspiring cover story about @GoldsteinBrooke and her grassroots org @EndJewHatred which are transmuting the character of the cognitive war against Israel, especially on US campuses. @EinatWilf @LekhtNaya @blakeflayton @CotlerWunsh @michaeldickson

9.29.2022 12:57pm - (Replying to @AlexTensor) What is the input = Data + Judgment ??

9.29.2022 5:12am - Dear @GovMurphy , Show me ONE Governor, or ONE University President who does not vow to fight antisemitism on US campuses. Only a handful succeed. Why? Because they are using outdated vaccines. Advice: Insist on Zionist representation in every EDI office on campus.

9.29.2022 4:52am - Berkeley Develops Jewish-Free Zones via @jewishjournal
9.29.2022 4:52am - (1/2) As frightening as these racist Zones are, our duty as academics is not to remain silent but to respond loudly and immediately: Any scholar scheduled to speak at any of the nine discriminating student groups at Berkeley Law School should receive hundreds of message from
9.29.2022 4:52am - academic friends and colleagues, urging him/her to rescind the invitation. Who should organize it? My natural inclination is to say: @ADL and @HillelIntl ; that's where the Jewish community has invested its resources and expectations. @JGreenblattADL @JNS_org @MiriamElman

9.29.2022 2:33am - I was told in Kindergarten it's not nice to laugh when someone else is angry. I tried to live up to it, but had to crack up (forgive me teacher) when Fatah declared another "Day of Rage" yesterday, to be added to the Calendar below:

9.29.2022 2:05am - (1/2) (Replying to @ylecun @artistexyz and @pabbeel) Mostly the bulging works in CI (causal inference) which provide theoretical characterization of reasoning tasks in terms of the questions they can answer and the kind of data they must be trained on. Here are a few examples:
9.29.2022 2:11am - (2/2) (Replying to @yudapearl @ylecun and 2 others) And here is another paper (w/ Yoshua as co-author) specifically characterizing the expressiveness of NN's including, I believe, proposals such as yours, based on Next-state-predictive models. But #Bookofwhy is a good and gentle primer..

9.29.2022 1:42am - (Replying to @KhaledAbuToameh) I don't recall a single day in the past 100 years that was not a "Palestinian Day of Rage", starting with the killing of the author Y.H. Brenner (Jaffa, May 1921). Hey, brother Abbas, how about a "Day of Reason and Co-Existence"? Just to try something different.

9.29.2022 1:35am - (Replying to @blakeflayton) I was about to retweet it, but had to pause. "Progressives" would say: No way! No "progressive" can ever be an antisemite", axiomatically. They are merely angelic "anti-Zionist". Why not name the racists for what they are?

9.29.2022 12:59am - Perhaps not coincidental, @HillelIntl has a job opening for CEO of Campus Climate Initiative: My advice, get some faculty of @SAFECUNY on the search committee, else you might hire another philosopher of antisemitism, susceptible to the tactic of inaction.

9.29.2022 12:43am - Thanks for the memories.

9.29.2022 12:34am - (Replying to @SAFECUNY and @HillelIntl) The same happened at USC 85 prominent professors complained about Zionophobic discrimination and USC's President anointed a "committee" to philosophize on "antisemitism", to the cheering sounds of Jewish organizations, granting her a license for inaction.

9.29.2022 12:21am - The honor were ours, and the LA Press Club who were looking hard for a Middle East Reporter with courage and integrity to report facts and avoid propaganda. Stay the course.

9.28.2022 11:53pm - Strange new partnership: @SAFECUNY with Hillel International on Campus Climate. Why is it strange? Notice the absence of the word "anti-Zionism" in the press-release, and absence of consultation with Jewish faculty. Why it matters:
9.28.2022 11:53pm - (2/2) It seems to me CUNY is using @HillelIntl as a fig-leaf to pacify community outrage over CUNY Law School exclusion of Zionist students and faculty. Again and again, students complain about Zionophobia and Jewish organizations betray them with a Kumbaya on "antisemitism."

9.28.2022 10:40pm - (Replying to @artistexyz @pabbeel and @ylecun) The problem with this slide is not absence citations (Pearl has enough of them) but its list of challenges, some of which were given theoretical underpinnings in other disciplines of AI research, largely inaccessible to DL folks.

9.28.2022 10:14am - I am getting signals that someone is trying to slander my colleague @KhaledAbuToameh. Let me state categorically that, having been reading his stories over the past two decades, I find him to be the most reliable and insightful source for all aspects of Palestinian society.

9.28.2022 10:04am - (Replying to @stephensenn and @learnfromerror) I am sure that if the alchemists explain what they want to do, (i.e., what the purpose is of analyzing "(fairly) simple experiments", what is given and what is assumed) Chemistry can help them get there faster.

9.28.2022 9:17am - I fully endorse @CotlerWunsh call "to add “Zionist” to the list of protected characteristics in existing hate speech policies, affording Zionists the very same treatment as any and every other protected characteristic." See why

9.28.2022 8:48am - (Replying to @MatthewEGunter @TuomasPernu and 2 others) I like this distinction: "intuitive causal models" vs. "analytical causal models", though sometimes the distinction lies not in the model itself, but in how it is used. For example, a DAG is "intuitive", but to find out if a set Z is "backdoor admissible" takes analytical work.

9.27.2022 11:08pm - (Replying to @omaclaren @DerriereLaLune and 2 others) Nice and clear. I think I begin to see where your generalization is heading. Need to digest.

9.27.2022 10:59pm - (Replying to @eoteromuras) Thanks for correcting the link.

9.27.2022 10:11pm - (1/2) (Replying to @DerriereLaLune @omaclaren and 2 others) My comments: It is a good and broad exposition of some problems in mediation analysis, uncontaminated by the dismissive attitude of the Imbens-Rubin school. But I do not see much difference between "mechanism" and "structural equations". In fact, the beauty of the First Law
9.27.2022 10:25pm - (Replying to @yudapearl @DerriereLaLune and 3 others) is in translating scientific knowledge about mechanisms (structural equations) to statements about counterfactuals and interventions. Strangely, only a few enlightened economists/statisticians have come to appreciate the power of the First Law. See:

9.27.2022 9:38pm - (1/2) (Replying to @instrumenthull @omaclaren and 2 others) You are right, I should have said "traditional 20th century statistician", with which our discussion started by D. Mayo. Rubin is semi-traditional in that he introduced non-statistical notation, Y_x. Still, he could not articulate when the equality P(y|do(x))=P(y|x) holds,
9.27.2022 9:49pm - (2/2) (Replying to @yudapearl @instrumenthull and 3 others) without invoking "ignorability" conditions which, as we have shown again and again, are thousands of miles away from the way scientific knowledge is stored. Indeed, potential outcomes penetrated social science only in the late 1980's.
9.27.2022 9:57pm - (2/2) (Replying to @instrumenthull @omaclaren and 2 others) He surely was. But his potential outcome notation was devised for one specific task (confidence intervals in RCT); he certainly did not consider E(Y_x) to represent "reality" (or "the world"), the topic where the thread started.

9.27.2022 8:52pm - (1/2) (Replying to @instrumenthull @omaclaren and 2 others) Sure! Neyman not only "talked" but also gave it notation. And Rubin (1974) further added the consistency Rule to Neyman notation. This is all in the history books and in #Bookofwhy. But it wasn't until the First Law was formulated that we were able to interpret your k's as
9.27.2022 9:05pm - (2/2) (Replying to @yudapearl @instrumenthull and 3 others) counterfactual statements, in Neyman's notation and, accordingly, use them to answer many policy questions, e.g., How many patients who died from Covid would have been alive today had we started the vaccine a month earlier. It's all in the First Law

9.27.2022 8:06pm - (1/2) (Replying to @omaclaren @learnfromerror and @stephensenn) If the k's represent something as ill-defined and mysterious as "causal indexing terms", how can we use them to answer meaningful policy questions? This is what Haavelmo asked. Now, the many bio-statisticians who are using these mystery models are probably asking: who needs
9.27.2022 8:14pm - (2/3) (Replying to @yudapearl @omaclaren and 2 others) to go formal? We are happy with our mystery models and once we estimate the k's we are done. What do we gain by being clear about the k's? The answer has many frills, but here is one of my favorites. Looking at the model, can you tell which of the k's can be estimated by OLS?
9.27.2022 8:37pm - (3/3) (Replying to @yudapearl @omaclaren and 2 others) (OLS = Ordinary Least Square). The answer is: Easy! It's known as "single-door criterion" see Corollary 1, For many other frills look up Section 3.2 of

9.27.2022 7:48pm - (Replying to @MatthewEGunter @IntuitMachine and @ylecun) Wondering: In what field can we find scholars capable of distinguishing "relevance" from "correlation"? (be it formally or conceptually)

9.27.2022 6:31pm - Gratifying to see red-blooded economists rebel against DML gurus and professors.

9.27.2022 6:11pm - (Replying to @eliasbareinboim @ylecun and 2 others) Thanks, this is the paper I had in mind. It also makes us appreciate the power of SCM to represent parsimoniously the function s(t+1) = g (a(t), s(t)) where a and z are any vectors of actions and state-features combination.

9.27.2022 5:21pm - The real fun begins when you ask them counterfactual questions about betting on the coins, as in Section 9.3.1

9.27.2022 4:50pm - (1/2) (Replying to @omaclaren @learnfromerror and @stephensenn) This is what happened in pre-causal econometrics too, until Haavelmo asked (1943) "what's the effect of an outside intervention? and the causal revolution emerged He also asked (1) Define k1 (in prob terms) and (2) what's the effect of X on Y? etc.
9.27.2022 5:08pm - (Replying to @yudapearl @omaclaren and 2 others) My point is that, because the answers to all these questions cannot be expressed in prob terms , they were not asked, and k1,k2,k3 got conflated with regression coefficients. Prominent statisticians even claimed: "K1 has no meaning" (1992) and worse.

9.27.2022 3:55pm - Say what you will about #GiorgaMeloni politics. If she manages to dig the truth despite all the Zionophobic propaganda around, she has some intellectual qualities that others don't.

9.27.2022 3:48pm - Some "storm troopers" seem more civilized than their fake victims.

9.27.2022 3:42pm - (Replying to @HenMazzig) Beautifully put. This is today's deepest root of antisemitism. We have been asking for protection for so long, and the world has gotten used to protecting us, sometimes heroically. We misbehave when we stop asking for protection and insist on doing it on our own, as equals.

9.27.2022 3:26pm - (Replying to @omaclaren @learnfromerror and @stephensenn) This is precisely the problem. It is easy to say, let P_1 be the prob. under intervention and P_0 the prob. under no intervention. Done! This is where 20the cent. statisticians stopped. What they missed was notation for the commonalities and differences between P_1 and P_0.

9.27.2022 12:09pm - (1/2) (Replying to @omaclaren @learnfromerror and @stephensenn) Trying to exonerate misguided Pearl from having assumed that "reality equals a prob distribution". Haven forgids. Pearl merely observed that statisticians, throughout the 20th century, had only ONE mathematical object with which to express knowledge about the world i.e, P(*).
9.27.2022 12:20pm - (2) (Replying to @yudapearl @omaclaren and 2 others) They were no dummies. They knew in their heart that knowledge is MORE than just P(*), but had no notation to express the MORE. Smart 21th century researchers may give MORE a probabilistic dressings, using "sets of P(*)'s", fine. But Pearl's observation of 20th Century is valid.

9.27.2022 6:12am - (Replying to @IntuitMachine @ylecun and 2 others) I am not insisting on world models to be of a special kind. But they need to have some machine representation. If a new representation scheme is proposed, I would be the first to welcome it and exploit its merits. Bring it over!

9.27.2022 6:01am - (Replying to @ylecun @GaryMarcus and @TiernanRayTech) Once training is done successfully, what would the final product be? And how would it be represented in the machine? A 3-variable toy example would be very helpful.

9.27.2022 5:44am - (Replying to @gamlambard) Argue about what? History of ideas? About the merit of SCM? I don't get it.

9.27.2022 5:33am - (Replying to @ylecun @BethCarey12 and 4 others) Not sure which topics are "new topics" and which are "more traditional ones". Is Rung-1 inference "new" or "traditional"? How about Rung-3 ?

9.27.2022 5:25am - (Replying to @RonKenett @stephensenn and @learnfromerror) I hope you can help us "get there", first by formulating "blocking" in the language of SCM. If this cannot be done, it would be the first concept uncapturable by SCM. If it can, what is it?

9.27.2022 5:18am - (Replying to @MaxIlse and @TacoCohen) Peace at all cost, of course. But peace without understanding is short lived. Why run away from performance guarantees? We have already a calculus that predicts performance given a world model; it shouldn't be hard to ask: what should the world be like for performance to succeed?

9.27.2022 2:12am - (Replying to @RonKenett @stephensenn and @learnfromerror) Ron, you seem to understand both frameworks, so, how about explaining to our readers in lay language what "blocking" is and, more importantly, whether the distribution engineered by "blocking" can be expressed in the language of SCM.

9.27.2022 1:53am - In honor of the Second Day of Rosh Hashana I am retweeting this inspiring sermon, on Youtube It should be supplemented, however, with one more ingredient, uniquely Jewish: Collective memory.

9.27.2022 1:40am - (Replying to @BethCarey12 @prem_k and 2 others) In the context of CI, the answer is clear: DL lifts us from finite samples to approximate distributions. The rest, going from research questions and world model to distributions, should be done outside DL.

9.27.2022 1:26am - (Replying to @TacoCohen) By all means, but it would have to be written in the style of Causality, i.e., accompanied by performance guarantees, toy examples, and explanations based on familiar theories of causation.

9.27.2022 1:14am - (Replying to @a_randomwalker @GaryMarcus and @TiernanRayTech) Agree. Most of these adjectives were authored to mean: "My program is 'better' than yours" as opposed to "This task is harder than the other", a differentiation that reigns the Ladder of Causation.

9.27.2022 12:46am - The nice thing about Chinese/Japanese readerships, they take Causality as a fresh "New Science"; they do not feel obliged to defend the honor of their professions and argue (as statisticians and economists do) "we have known it since 1890".

9.27.2022 12:17am - I take it as a hint that a 3rd Edition is due. I agree! It should contain Chapter 12, on Fusion, selection bias, missing data, individualized decisions, etc. and perhaps Chapter 13: "Juicy fights with Infidels on Twitter." Good idea.

9.26.2022 8:38pm - It's Monday! Time for us to disseminate last week's articles on causal inference:, and examine what's new in the field; new ideas, new applications, new players. Use and help use.

9.26.2022 6:12pm - (Replying to @jazchaz) Shana Tovah U'Metukah. Is this the same Rabbi Angela Warnick Buchdahl that contributed an essay to our book "I am Jewish" ?? If so, send her my hugs. Shana Tovah Anachnu Po

9.26.2022 5:42pm - (Replying to @ylecun @GaryMarcus and @TiernanRayTech) As one who advocates (and demonstrates) ML+CI (causal inference) symbiosis, I would be the last one to think that ML/DL conflicts with CI. The conflict I do see is in the prevailing practice of ML/DL which attempts CI without a world model, without even aiming to learn one.

9.26.2022 5:28pm - (Replying to @ylecun @GaryMarcus and @TiernanRayTech) If I were one of your 353K followers, I would have liked to hear from you whether the "next-state-predictor" you propose circumvents the basic limitations identified in the Ladder of Causation or, at the very least, whether I should even be concerned with those limitations.

9.26.2022 5:09pm - (Replying to @jacyanthis @ylecun and 2 others) Hybrid deep learning plus world models CAN ASCEND above Rung-1; this is precisely what the works on causal inference demonstrate. It also demonstrates that deep learning w/o world model CANNOT ASCEND. Human brains have a general template to fit world models into.

9.26.2022 4:59pm - (Replying to @carolw and @SAFECUNY) Begging to differ. Look what the haters accomplished with the word "Islamophobia". You can't say "peace" today without being labeled "Islamophobe" - the ultimate cardinal sin. Phobia is not merely "fear", it's any irrational, obsessive behavior that is socially unacceptable.

9.26.2022 11:35am - Muslim Brotherhood’s Qaradawi, who endorsed suicide bombings against Israelis, dies via @timesofisrael
9.26.2022 11:35am - My son Danny interviewed Qaradawi in 2001 for the Wall Street Journal, to see how "moderate" he really was. I mentioned him twice in my commencement speech, University of Toronto: No, he wasn't Mother Teressa.

9.26.2022 10:56am - (1/4) Thanks for the rebuttal. I don't have as strong a reason as @GaryMarcus has to be pissed by @ylecun 's latest discoveries; he has never dismissed me as one who writes only irrelevant papers; in fact, he has never mentioned any work on causal reasoning, with the exception of
9.26.2022 10:56am - (2/4) a passing remark, that Facebook has 12 PhDs working on causality, and they know what they are doing. I suspect the 12 never convinced him that the obstacles DL is facing (eg, explainability or generalizability) amount to mathematical impossibilities. One good thing we can
9.26.2022 10:56am - (3/4) expect from this debate is to see hundreds of DL devotees raise their eyes and ask: "Guru now says we need a 'world model', what is it? We thought all along that our NN is a world model! Isn't it?" This question would prove transformative in the lives of DL folks, realizing
9.26.2022 10:56am - (4/4) that a 'world model' requires more than a patch to some NN structure, but a whole new philosophy about knowledge, learning and cognition. They might even read #Bookofwhy to see how the new philosophy is concretized in a small corner of the wood called "causal inference."

9.26.2022 6:12am - Sister @Ilhan ! Sister @Ilahan ! Listen to what your tricky Jewish neighbors are praying tonight: "Sound the great shofar for our liberty, and raise a banner to gather our exiles, and gather us together from the four corners of the earth." Sounds like a Zionist conspiracy!!

9.26.2022 5:28am - (1/2) I am reading a new interview with @ylecun in which he professes some basic limitations of current approaches to AI. Alas, I get the impression that he hasn't read #Bookofwhy, since he never got to talk about the unassailable limitations presented by
9.26.2022 5:28am - the Ladder of Causation or by Radical Empiricism in general: I sometimes wonder if we are working towards the same goal (AGI), since even the discussion of Bayesian Networks is stuck on Rung-1. Looking forward to @GaryMarcus rebuttal. @TiernanRayTech

9.26.2022 12:02am - (Replying to @carolw and @SAFECUNY) They have a point. We have a word that highlights the specific racist character of anti-Zionism, Zionophobia, and we are not using it. They wouldn't dare tell us: "stop weaponizing Zionophobia" because they admit it and are proud of it.

9.25.2022 10:09pm - (Replying to @JerryLe00723434 @CNNSotu and @jaketapper) He would. He is not a coward, He just need to understand the ropes of the conflict and, confessedly, Jewish leaders do not show him the way either. They keep talking about borders, occupation, antisemitism, terrorism. All but the Elephant in the room -- Arab rejectionism.

9.25.2022 5:13pm - (Replying to @minus_1984) Zionophobes disregard of humanity is often accompanied with hate of the free world.

9.25.2022 4:03pm - Health permits, I'll be there Sept 28, and will call Roger Waters for a public debate on the topic "The racist and genocidal character of Zionophobia, from the standpoint of its victims." Join me!

9.25.2022 2:18pm - (Replying to @cubic_logic and @stephensenn) Havens forbid! I claim that no STATISTICAL model, not even perfectly adequate ones, have ever answered SOME basic questions about reality? Readers must be tired by now seeing the basic questions repeated. I suggest that folks who are still surprised by the claim try to refute it.

9.25.2022 2:06pm - The Ladder of Causation can be viewed as a concrete embodiment of Kopper's ideas, where the impossibilities he is talking about become mathematical impossibilities of climbing from Rung-i to Rung-(i+1) without the necessary information.

9.25.2022 1:51pm - (Replying to @ChristosArgyrop @cubic_logic and @stephensenn) But notice that in all Jaynes writings "reality" is equated to the probability distribution P that governs what we see, never beyond that. Max-entropy was devised to approximate P best we can. Next thing to notice, P, even known precisely, cannot answer: Was A the cause of B?

9.25.2022 3:58am - (1/2) Wishing all readers: HAPPY JEWISH NEW YEAR. May the new year make your wishes come true. My three personal wishes: (1) May our neighbors deem us worthy of peaceful co-existence. (2) May all oppressors be toppled by their oppressed constituents. (3) May we have the strength
9.25.2022 3:58am - (2/2) to endure life, in confidence, hope, creativity and dignity, until wishes (1) and (2) materialize. Shanna Tovah U'Metukah!

9.25.2022 3:30am - (Replying to @stephensenn and @learnfromerror) Perhaps we can see the problematic assumption if you state it in the context of the model, where we have only two halls, each offers one and only one diet? I don't recall using Ancova; I use conditional expectation, which is a property of a distribution, not of how it's estimated

9.25.2022 2:43am - (Replying to @stephensenn and @learnfromerror) I don't see any assumptions made about regressions, only about causal relationships, specified formally in the graph. Perhaps you can state the problematic assumption in the context of that graph, where we have only two halls, each offers one and only one diet?

9.25.2022 2:46am - @Alfrednguyen432 just posted a 2012 video of a seminar by Jim Heckman and myself which is highly relevant to our discussion of why statisticians can't let go of the statistical paradigm, and why economists may be able to do so.

9.25.2022 2:06am - (Replying to @stephensenn and @learnfromerror) The #Bookofwhy solution to Lord's paradox is beautiful and general: It does not depend on any unstated assumptions, if we stick to the original statement of the paradox, free of added complications, as shown here:

9.24.2022 9:11pm - (Replying to @learnfromerror and @stephensenn) I see nothing contradictory here: "If you think reality for Cox was MORE than prob. dist., show me the mathematical object he used for the MORE." Rephrased: If you think Cox deemed prob. dist. insufficient to capture some aspects of reality, what did he use to capture them?

9.24.2022 7:49pm - (Replying to @learnfromerror and @stephensenn) Quoting what I "alleged about what stat presupposes": "in traditional statistics, "reality" is equated with the probability distributions that govern the data we obtain." If you think reality for Cox was MORE than prob. dist., show me the mathematical object he used for the MORE.

9.24.2022 7:23pm - (Replying to @learnfromerror and @stephensenn) Show me a mathematical object used by Neyman, Pearson or Cox, capable of answering the question: What is the likelihood my uncle would be alive today, had he not smoked. Pearson, in particular, said: "It's all in the contingency table." I quote him here

9.24.2022 6:52pm - (Replying to @learnfromerror and @stephensenn) I am claiming that, in traditional statistics, "reality" is equated with the probability distributions that govern the data we obtain. "Reality" is more than that. It includes "roads not taken", i.e., cause effect relationships that permit us to generalize and explain things.

9.24.2022 6:26pm - (1/2) (Replying to @ChristosArgyrop and @stephensenn) He sure would, but the hypotheses he had in mind were all statistical. We are now in a new era, where our hypotheses are about causes and effects (including causes of effect) so all the beautiful ideas of Jaynes, Box, and others need reformulation. The degree to which this
9.24.2022 6:26pm - (2/2) (Replying to @yudapearl @ChristosArgyrop and @stephensenn) reformulation is "revolutionary" can be measured by the degree to which it is resisted or mocked by prominent statisticians to this very day. Why? See

9.24.2022 6:05pm - (Replying to @HL327 @stephensenn and @RogueRad) I think even Huxley would agree with us that you cannot slay a theory or hypothesis unless it is represented as a mathematical object. This is precisely what ML (and statistics) lack, a mathematical object representing reality, not merely data that reality graciously grants us.

9.24.2022 5:51pm - (Replying to @stephensenn) This distinction is an integral part of the scientific method which dedicates separate symbolic systems for (1) model of reality, (2) drawing conclusions from it and (3) testing it against empirical evidence. None exists in ML (or statistics) where data is mistaken for "reality."

9.24.2022 5:36pm - Thanking all readers who wished me well on my eye surgery. It went well, but it's still hard for me to return to normal activity, due to pain, patch and general discomfort. Will join the revolution slowly, so we can defend the faith together, and promote heresy when productive.

9.24.2022 10:18am - The Israeli-Palestinian conflict in 8 words: Lapid to UN: Two states, equally legitimate. Abbas to UN: Two states, unequally illegitimate.

9.24.2022 10:05am - The "unassailable" mathematics tells us what it takes to climb from one rung of the Ladder of Knowledge to the next. Our brain has a model of reality, however flimsy -- most ML systems do not; they operate on Rung-1 of the Ladder.

9.24.2022 9:51am - (Replying to @raedhamed1 @IlhanMN and @lsarsour) The struggling women in Iran are begging women who wear the Hijab by choice to show solidarity with those who are burning it as a symbol of no choice.

9.24.2022 2:05am - To all readers on this channel, I am undergoing a minor eye surgery tomorrow, so may be away for a day or two (hopefully not more). Stay the course and defend the faith. I gave the surgeon a copy of #Bookofwhy, of course, to keep his hands steady and his mind purposeful. Best, JP

9.23.2022 12:42am - (1/2) Luckily, I can escape the wrath against "AI critics". (1) I'm an "AI believer", I only critic AI's slavery to popular trends in ML. (2) My heretical claims are not mine, they come from unassailable mathematics. I admire heretics like @GaryMarcus and others who decry the
9.23.2022 12:42am - (2/2) waste, dangers and thoughtlessness of the imbalance created by the trendy ML dominion over education, funding and media attention. It makes the revolution so much harder to take off and reveal the potentials of AI systems that reason and understand.

9.23.2022 12:11am - (Replying to @rravi) Trying, but failing to parse the solution you propose, or any solution that is not based on uprooting deeply entrenched rejectionism to any form of co-existence that now dominates, to the last person, one side of the conflict.

9.23.2022 11:12pm - Back to School With Antisemites via @jewishjournal
9.23.2022 11:12pm - (1/2) Another brilliant back-to-school article by @thanerosenbaum , with a minor update: It's no longer sufficient for students to swear a symbolic oath that Israel is an odiously racist, colonial enterprise. To be part of campus life, students must translate their oath into active
9.23.2022 11:12pm - (3/3) hostility against Israel and join the armies of Zionophobic thuggery, often to the tacit approval of Jewish organizations, and always to the willfully blinded eyes of University Administrators. Our only hope: pushback Zionist organizations like @newzionists and @SSI

9.22.2022 8:57pm - (1/3) The ELEPHANT behind PM Lapid speech at the UN. Commentators express surprise that he explicitly backed a two-state solution. Rightly so, A decade has passed in which the 2-state dream has been simmering in the minds and hearts of most Israelis, yet absent from any diplomatic
9.22.2022 8:57pm - (2/3) discourse. The reason lies in the elephant: "We have only one condition: That a future Palestinian state will be a peaceful one. That it will not become another terror base from which to threaten the well-being, and the very existence of Israel." Here he states politely what
9.22.2022 8:57pm - (3/3) every Israeli knows too bluntly: Palestinian rejectionism is by far the main obstacle to peace, and any mention of "2-state" short of a plan to dismantle this obstacle is a ticket to futile negotiations and painful disappointments. I'm afraid Abas will ignore the ELEPHANT.

9.22.2022 8:14pm - Saluting Christiane Amanpour for her refusal to comply with Iranian President Raisi's demand to wear a headscarf which, to the women of Iran, is a symbol of women subjugation. The empty chair which Raisi refused to honor speaks more about him and his regime than 1,000 words.

9.22.2022 8:06pm - We can do more for the Iranian people, certainly in the public arena. For example, Muslim feminist leaders @IlhanMN and @lsarsour could take off their hijabs publicly, in a moment of silence, as a gesture of solidarity with the Iranian women who are burning their headscarves.

9.22.2022 4:50am - (Replying to @slasherbarbie and @blakeflayton) Beware! The last guy who called me "colonizer" was asked "Cite one Philistine poem or one Canaanite hero". He never showed up again.

9.22.2022 4:24am - (Replying to @David_Leavitt and @MassGov) Glad you discovered their ploy before it was too late. They would have sacrificed you to their false Gods and used your blood to prepare Hanukah Latkes. Praise to the Lord you are alive.

9.21.2022 7:24pm - First they claimed sole ownership of Tel Aviv, then sole ownership of our history, now they claim sole ownership of progressive values. Soon @RashidaTlaib will demand sole ownership of mathematics and human biology. Can someone teach her the word "share"?

9.21.2022 12:58pm - (Replying to @goldenrule108 @IlhanMN and 2 others) While support for the "mandatory hijab" is zero in the United States, solidarity with those in Iran would be empowering, if offered by those whom the world perceive to be the voice of American Muslims. No axe on Palestine, just on Israel = my litmus test of logical consistency.

9.21.2022 9:58am - (1/2) (Replying to @stu_frost @JohnHMorris and @jhurwitz) The gap is not between me and ML folks, it is between Rung-1 and Rungs-2&3 of the Ladder, which is a mathematical property of epistemology that some folks (not all) can't acknowledge, lacking a calculus to express it even if they wanted to. To acknowledge that we can't
9.21.2022 10:02am - (Replying to @yudapearl @stu_frost and 2 others) square a circle we need first to have a language in which to define a "circle". We can't define "causal effect" in the language of statistics.

9.21.2022 3:25am - (Replying to @sdandach @ilanomar and @RashidaTlaib) To give their followers/voters an illusion of principled consistency; most voters enjoy this feeling.

9.20.2022 10:52pm - (1/2) A readers noticed a glitch in my tweet: The tag of Rep. Ilhan Omar is @IlhanMN , and that of Linda Sarsour is @lsarsour ; both claim to be feminists, (see and both argue that the Hijab is NOT a symbol of women subjugation. Rep @RashidaTlaib does not
9.20.2022 10:52pm - (2/2) wear a hijab, but I am sure the protesting Iranian women would appreciate a show of solidary from her as well, since many of her Minnesota constituents tend to side with the Ayatollahs.

9.20.2022 9:51pm - As Iranian women burn headscarves in anti-hijab protests, see, I encourage American feminist leaders, Linda Sarsour, @ilanomar and @RashidaTlaib to show solidarity and publicly take off their hijab, for ONE solemn moment of silence.

9.20.2022 9:21pm - (Replying to @SAFECUNY) Love it!

9.20.2022 9:17pm - I don't think you have noticed how much further the politics of Islamophobia has advanced. Today, a speaker who utters the word "coexistence" is immediately branded an "Islamophobe". In truth, this word makes Palestinians feel terribly unsafe, and one should be considerate.

9.20.2022 9:07pm - (Replying to @blakeflayton) Welcome home! Blake!

9.20.2022 9:06pm - I forwarded it to my granddaughter (Med school, UBC) thinking: no student should graduate med-school without a glimpse at #Causal Epistemology.

9.20.2022 4:11pm - Don't be intimidated by military exercises and other dark clouds over Taiwan. It's October! If you find yourself strolling in Taipei on October 1, drop in to enjoy some good music and press freedom:

9.20.2022 3:04pm - (Replying to @PippaLally @PWGTennant and @statsmethods) Never give up! Your RQ is P(C|do(A)) =? The graph tells you P(C|do(A)) =SUM_B P(C|B,A)P(B) Done. See also Crash Course

9.20.2022 12:03am - My grandparents perished at Auschwitz, and I felt they expected me to scream yesterday, when Iran's Raisi spoke about "some evidence". Today I feel my scream was heard loud, even louder than what they expected, through the words of IDF Chief of Staff: "You have to be human".

9.20.2022 11:48pm - Review: Ken Burns issues a chilling warning with ‘The U.S. and the Holocaust’
9.20.2022 11:48pm - I have watched two of the three episodes. Not to be missed, both for subject and the method of exposition. I'll be back after episode-3, because I begin to sense that something is still missing.

9.20.2022 11:22pm - Causal Inference for fun, profit and science -- here are the articles that came into public view this past week:

9.20.2022 11:13pm - (Replying to @Ostrov_A) Thoughtless Jews of Discomfort ( are not irredeemable. For example, I'm told some signatories on the VanLeer Definition of Antisemitism have retracted their support once shown the consequences of that harmful document. I feel @YoBenCohen is redeemable.

9.20.2022 10:05pm - I don't think they are "a bunch of Jew hating hypocrites". They are simply "Jews of Discomfort", bullied to offer a "pound of flesh" to social acceptance culture, of the type nourished in many synagogues by Tikkun Olam Rabbis who forgot their "Birkat Hamazon". I know some.

9.20.2022 9:44pm - My personal tip to @YoBenCohen : If you wish your words to be taken seriously, distance yourself from Master-Zionophobe @Mehdirhasan on @MSNBC . He opposes everything you said about the rights of Israelis to life and liberty.

9.20.2022 9:32pm - I would not single out the Unions to blame for this pandemic. I would add the EDI industry (Equity Diversity and Inclusion) for the way they train their stuff nationwide.

9.20.2022 9:17pm - Rep. Rashida Tlaib: Israel has apartheid system, two-state solution won’t work via @JNS_org
9.20.2022 9:17pm - (1/2) Rejoice, Rejoice! @RashidaTlaib came up with a new reason to oppose a homeland for the Jewish people. It started with: "You are not a people", continued with "Not even the size of a postage stamp", then "You do not belong here", then "On our dead bodies! Rivers of blood!"
9.20.2022 9:17pm - (2/2) Then "No! No! No! (Khartoum 1967), then two intifadas. Today it's a new song: "We tried everything, it just doesn't work out, we must rethink things over". Rethink, Rashida, Rethink!

9.19.2022 4:37pm - To be indigenous to a country means to name your children after hills, valleys and rivers in that country; eg. Carmel, Chermon, Tavor, Ella, IZraeli, Jordan, Yarkon...etc. not after foreign religious or military leaders.

9.19.2022 2:15pm - (Replying to @bronzeagepapi) "Actual causality" is a subset of Rung-3. For example, PN, the probability of Necessity, differs from "actual causality", yet it is Rung-3.

9.19.2022 2:06pm - One would expect the Netufian-Canaanite fable to flourish in Stand-Up Comedy lane. No way! Some still think it strengthens BDS arguments. "A people without a history, must create one!" said King Solomon the Wise (a Netufian Imam)

9.19.2022 5:10am -

9.19.2022 5:10am - As everyone knows, Palestinian story-tellers always start their narrative with "Everyone knows". The classic was Chairman Arafat who, in Camp David, whispered in Clinton ear "Everyone knows the Jews never had a Temple in Jerusalem". As everyone knows, Clinton was not impressed.

9.19.2022 4:42am - In memory of Justice Ruth Bader Ginsburg, I am posting the essay she contributed to the book "I am Jewish":

9.19.2022 4:30am - I'm glad she put the two questions side by side, to equate the credibility and logic of both answers. Usually, there is a 3rd quality that correlates with Holocaust denial and Israel deligitimization: Repressiveness. I guess Leslie tried to save him the embarrassment.

9.19.2022 4:17am - (Replying to @FlorianWilhelm) I am not bothered, just asking for clarification. Do you think AI is synonymous to Deep Learning? I think that after "learning" comes "using", regardless how deep it is, and "using" is not always trivial.

9.19.2022 3:53am - Glad to see the first step accomplished: Demand that Zionism be recognized as "protected characteristic", replacing the fuzzy term "Jews" (even JVP claim they are jews). Next step: Declare Zionophobia a form of racism, of equal status to Islamophobia and White Supremacy.

9.19.2022 3:34am - Good news! UCLA is hiring! See invitation from our chairman If you are a super researcher in any area of AI, or you know of one, please consider this opportunity. And if your research is in Causal Inference, I can't wait to hear your talk.

9.19.2022 3:16am - Hilarious! Who said they don't teach selection bias in stat?

9.19.2022 2:05am - (Replying to @blakeflayton) This is partly because Zionist activists are still not using the word #Zionophobia, to let the "peacemakers" know how profoundly immoral BDSers are, much more repulsive than "academic or economic boycott".

9.18.2022 11:28pm - (Replying to @skeenscene @Claire_V0ltaire and @ReconRabbis) Brant Rosen exist? I thought he must be a Hamas operative making fun of Jewish holidays. In my home town he would be excommunicated but, given that there isn't anyone here to do it, let's ask the Chief Rabbi of Gaza. She must have the credentials.

9.18.2022 5:17pm - (Replying to @blakeflayton) It's not nice to talk like that against my hero, Lina Sarsour, who is a known Jew-lover (donated money to repair vandalized Jewish cemetery) and all the Jewish leaders competed for a seat on her band wagon: Please use her respectable title: a Zionophobe.

9.18.2022 4:38pm - I am going to watch CNN soon (5 pm Pacific), it's a painful topic I cannot easily dismiss. It bears similarities to the British blockade against Jewish refugees, 1939-1948.

9.18.2022 6:22am - This is very kind of you. I'm retweeting with the hope that some Trialists will take the time to read a few chapters of #Bookofwhy. I was shocked the other day to find out that some DID NOT, then learned that MOST did not.

9.18.2022 6:14am - (Replying to @LuisaBorrell) Why? Because the media, academia, and even some Congresswomen have bought into Palestinian logic: Israel is guilty of horrible crimes, Jews are likely to support Israel, hence every hoodlum is morally justified in acting to silence this criminal support! It's elementary logic.

9.18.2022 6:02am - Thanks @TriciaRoth18 for reposting this famous video, with Zahra Billoo (Dir. of CAIR-SF) warning us (38:00 min.) against "polite Zionist" organizations, ADL, Hillel, Zionist synagogues, who will befriend you today and betray you tomorrow (e.g., by advocating a 2-state solution)

9.18.2022 4:41am - (1/2) (Replying to @AviBittMD) Of course people have spoken jiberish about confounding since it was noticed (By Pearson (1890's) and Yule. Of course people proposed, for more than a century, all kind of wrong criteria for confounding, and continue to do so today. Are u trying to correct their thinking? If
9.18.2022 4:48am - (2/2) (Replying to @yudapearl and @AviBittMD) If so, what is your criterion for "correctness". I, too, have no dog in the fight; I just happened to know that you cannot correct their thinking in the language of statistics, void of either graphs or counterfactuals. No way! Same for "randomization" or "c-effect". Up to U.

9.18.2022 3:07am - (Replying to @AviBittMD) You seem to imply that "how confounding is being defined in book of why" is somehow unique to #Bookofwhy, and that there are alternative ways of defining it, perhaps more meaningful". I would be eager to hear! Can we start with a formal definition of "confounding" or "c-effect"?

9.18.2022 2:50am - (Replying to @MikeNew19512115 @ChancellorCUNY and 13 others) Thanks @MikeNew19512115 for recognizing the one, and only one way of unveiling the racist face of "diversity" officers in our universities. To my knowledge, no Jewish leader has dared doing it, which is beyond me: Why? Don't they understand that a Zionophobe can't deny its credo?

9.18.2022 2:23am - (Replying to @nagaraj_arvind) What happened? The media, academia, and even some Congresswomen have bought into Palestinian logic: Israel is guilty of horrible crimes, Jews are likely to support Israel, hence every hoodlum is morally justified in acting to silence this criminal support! It's elementary logic.

9.18.2022 12:29am - America 2022.

9.18.2022 1:18am - (Replying to @DavidDeutschOxf) I think clinging to Kings and Queens as symbols of commonalities that unite members of a culture explains the longevity of those cultures. Abstract ideas such as "common values" quickly lose their punch, unless they are attached to names of individuals, and turn "common stories".

9.18.2022 1:00am - Sad to learn the passing of Saul Kripke, a brilliant philosopher. I once tried to interest him in the necessity of causation, but he insisted on "Naming and Necessity". A loss to philosophy and to history of ideas.

9.18.2022 12:18am - (Replying to @Horror1403 and @Eve_Barlow) Exchange #jews with #zionists, this is what they ought to have answered.

9.18.2022 12:04am - This great piece by Bill Maher need be enjoyed together with this article on him: I'm surprised the Woke Forces allow him to stay on the air. Stay the course @billmaher

9.17.2022 11:12pm - Interesting thread. My two cents: None of the Journals mentioned has invited a survey article on Causal Inference or Graphical Models, thus constraining economics to another decade of [...] [I would let red-blooded economists fill in the appropriate adjective]

9.17.2022 10:38pm - Great read! @ShMMor ! I've also written a piece about Beinart, "Jews of Discomfort", on his first visit to UCLA: "Mother, I hate you, for making me feel so inadequate, unable to defend your honor except by joining the bullies in amplifying your blemishes.”

9.17.2022 10:22pm - After refusing accreditation of Israeli envoy, Chile reschedules ceremony
9.17.2022 10:22pm - We had some discussion with Chilean's readers about their President's behavior, whether he is a racist or just "democratically elected". I'm glad he apologized for the incident, but I'm not sure Israel's ambassador was tough enough. Of course, I've never been in Ambassador shoes.

9.17.2022 9:48pm - Watching this 12-16 hours queue on BBC today made me ponder where nationhood and peoplehood reside in the human brain. Plus, it made me ask: How did we, Jews, survive without a monarch for so many centuries? Are the Biblical stories of King David a sufficient substitute?

9.17.2022 9:38pm - (Replying to @DavidHirsh and @centre_as) I wish I could take part in this important conference. I am hearing only good things about it. Stay the course.

9.17.2022 9:28pm - I would not blame any President for child exploitation in his country, because I havn't studied the issue in sufficient details to attribute blame to one factor or another. But I can confidently charge a President with Zionophobic racism, a topic I have studied for a few decades.

9.17.2022 9:04pm - Kudos to Professor Lax @CUNY_Prof for so clearly stating the charge: "discrimination against Zionist students and professors." To exonerate herself from allegations of partiality, all Saly Abd Alla has to do is to publicly state her views on Zionism. Challenge her to the task!!!

9.17.2022 5:58pm - (Replying to @Drg_elness and @gabrielboric) I the era of Putin and Ebrahim Raisi, we should not fall for the words "democratically elected." We should look at policies, norms and attitudes. President @gabrielboric failed the latter and "democratically elected" arguments do not make him less of a racist.

9.17.2022 4:41pm - (1/2) Last week I had a Zoom session with a higher-up in the FDA and the same question came up. I cast my answer in a positive mood: Solutions to chronic weaknesses of RCT have been proposed: Selection bias, integrating diverse studies, generalization to target population, etc.
9.17.2022 4:41pm - (2/2) It is time for the FDA to assemble a task force to examine these (theoretically sound) proposals, and remove the language barriers between "trialists" and "causal inference" researchers. Step-1, read #Bookofwhy. Step-2, read, the rest will follow.

9.17.2022 3:05pm - (Replying to @ShaharAzani and @Israel) And to think that we grew up 300 meters from each other, unaware of the potential, only 30 years apart, makes you hair stand up in awe.

9.17.2022 5:16am - (Replying to @robwilliamsiii) Sorry, Twitter tells me "This Tweet is unavailable" Evidently someone blocked me.

9.17.2022 5:12am - (Replying to @DavidDPreiss and @gabrielboric) Yet your countrymen claim it's shameful to compare @gabrielboric to Maduro. Can't they see where he is heading with his animosity towards Israel, and his determination to please the Zionophobic voting-base in Chile?

9.17.2022 3:53am - Glad to see the first step accomplished: Demand that Zionism be recognized as "protected characteristic", replacing the fuzzy term "Jews" (even JVP claim they are jews). Next step: Declare Zionophobia a form of racism, of equal status to Islamophobia and White Supremacy.

9.17.2022 12:09am - (Replying to @quantadan and @DrMayaPetersen) Pleading innocent on both charges. Scrapping half-finished dissertation is worse than converting to a new religion. Stay the course.

9.17.2022 12:00am - To Chilean Pres. @gabrielboric , Readers on this channel promised that you are not another Maduro. Please apologize!

9.16.2022 7:19pm - Highly recommended May this be the last video I retweet before lighting the Shabbat Candles with my family. Shabbat Shalom.

9.16.2022 7:08pm - (Replying to @GadSaad) Stay away from walking Zionophobes. Shabbat Shalom.

9.16.2022 7:06pm - (Replying to @GalitPeleg @AlEmbassyIsrael and 7 others) Be'Birkat Techezakna.
9.16.2022 7:06pm - (Replying to @Claire_V0ltaire) Plus, it is genocidal in effect, given that Israelis are adamant about defending their sovereignty tooth and nail.

9.16.2022 6:58pm - (Replying to @YiMaTweets) If you think the "world model" is stored separately from the data-fitting optimizer, let's pull it out and offer it as an alternative representation to the models currently used by causal-inference folks. This would be of great scientific value; but can we?

9.16.2022 4:37pm - Correcting dates. The phone call came Feb 24, 2002.

9.16.2022 4:28pm - (Replying to @WvanAmsterdam @MaartenvSmeden and @GelovenNan) An important distinction between "individual effect" and "average effect in a subpopulation resembling an individual". I hope your distinction matches ours:

9.16.2022 4:18pm - Gelman and Imbens argue that "it can make sense to ask questions about the causes of effects." Sadly, their dogmatic rejection of the First Law, leads them to erroneously conclude "the answers to these questions will be in terms of effects of causes."

9.16.2022 4:08pm - When @ShaharAzani showed me this clip, yesterday, I felt the same electricity as I felt on Feb 1 2021, when the US Consul in Pakistan read to me (on the phone) Danny's last words. I don't think there was a scholar then who didn't revise his/her theory of identity at that point.

9.16.2022 3:52pm - @tdietterich , your tweet has invited many NN folks into the thread who haven't heard that NN have uncrossable limitations. For them, I'm retweeting and, and, to calibrate claims of "NNs can do it too."

9.16.2022 3:33pm - (Replying to @wikibinator @NickRMorgan and 2 others) I'm not convinced we gain anything by calling all these tricks and architectures "neuralnets". Why not just do things correctly, following the scientific method, then choose fashionable titles to gain the right social acceptance.

9.16.2022 3:25pm - (Replying to @dancaron @trunghlt and 2 others) Not without doing it the way it's done by causal inference methodology, then calling it NNs. The key to the former is having an explicit representation of "the world", separate from your data-fitting optimizer.

9.16.2022 3:59pm - Scientists salute the wisdom of the scientific method as the greatest achievement of homo-sapiens in the past 50,000 years. Yet, when presented with a computational model of the method, they doubt its virtues, and prefer the wisdom of owls and snakes, modeled by neural networks.

9.16.2022 3:18am - (Replying to @AnoopRKulkarni and @azandisarietta) Statisticians at the time, including inter-disciplinary researchers, could not appreciate his revolutionary step: assigning a mathematical symbol (arrow) to a causal relation. Some statisticians (and economists) still do not appreciate it, even today. See

9.16.2022 12:13am - Now I am torn between the urge to think backwards (Rung-3) and assess our unimaginable loss had you not read Causality in 2002, or reason forward (Rung-2) and list all the marvelous things we are about to do next 34 years -- personalized medicine, free will, automated scientists

9.15.2022 8:28am - Admittedly, I was also skeptical about the authenticity of the #AbrahamAccords but, once you listen to the music at this wedding, you know it's genuine.

9.15.2022 8:18am - Nides: Biden administration taking Abraham Accords ‘from start-up to real operation’ via @timesofisrael
9.15.2022 8:18am - The importance of words: "Abraham Accords" vs. "Normalization Agreements". Biden administration, belatedly, is waking up to the opportunity! My concern: Nides' zeal to get the Palestinians on board may obscure hard reality: "normalization" for Palestinians mean "identity theft".

9.14.2022 7:57pm - Bill Maher said it better than I: Will President Biden denounce hate on both sides? Will any of the dignitaries dare denounce the hate that @IlanOmar and @RashidaTlaib bear against my people in Israel, and against my Zionist students at UCLA? Will @ADL ?

9.14.2022 7:30pm - Disagree, @JNelsonLDF . Those who fight hate in the abstract, w/o identifying its specific nature, are wasting their time. Here is your proof: Zionophobia, the ideology through which @ilanomar and @RashidaTlaib hate Israel and Zionism is missing from your list! @JGreenblattADL

9.14.2022 7:13pm - "Going just after the 'Zionists'" is not less racist than "going after others too." But @Cornell is redeemable, because its no-nonsense President, Martha Polack, is familiar with their "just after" tactics, and knows how to handle it. See last paragraph of

9.14.2022 6:48pm- via @ @i24NEWS_EN
9.14.2022 6:48pm- Condolences to all readers who expected Maccabi Haifa to win this game, especially to those how paid $5,000 a seat to see Messi play in Haifa. I was impressed by the Israeli soccer fans and the little girl who greeted the players before the game.

9.14.2022 3:51pm - (Replying to @hmatejx) He is right. Given the data alone, we can't tell this DAG from many others.

9.14.2022 5:18am - (Replying to @NikBles and @davidchalmers42) Not so simple. "have deep understanding = being able to answer question at all 3 levels of the knowledge hierarchy".

9.14.2022 3:26am - (Replying to @JFronczekMD) The conditioning may be involuntary. E.g., when samples are selected for a study based on the severity of the outcome (eg hospitalization or hiring).

9.14.2022 12:22am - The fault is not entirely theirs. Jewish leaders, from @HillelIntl to @ADL fail to insist on this recognition, preferring instead the seemingly safer "antisemitic" card. See my interview with Moment Magazine @Ostrov_A

9.13.2022 11:38pm - (1/2) Of the many examples demonstrating "causation without correlation" I especially like this: X--->Y, X-->Z<--Y, in which it is easy to produce: P(y|x)=/=P(y) and P(y|x, z)=P(y|z). Thus, conditioning on Z, a non-cause of Y, renders Y independent of its cause, X. For numerical
9.13.2022 11:38pm - (2/2) example, set 1-a^2-b^2 =bc in Eq.[20] of "The Microscope" (Selection Bias).

9.13.2022 11:11pm - Those who spent hours trying to understand Stanford's Foundation Models would take one glance at this picture and scream: "Eureka! "

9.13.2022 3:12pm - What does it take for a politician to critique "the fringe of both sides"? A public the rejects one-sidedness.

9.13.2022 2:29pm - If any of our readers plans to attend @ReichmanUni this coming fall, don't miss this course by @EinatWilf , even if it means dropping statistics 101. It's also time for @ADL to adopt this course as part of their effort to educate university administrators. @JGreenblattADL

9.13.2022 5:19am - The paradigmatic shift that @ChristinABuelow is describing parallels the traumatic experience I went through in 1988, upon realizing that statistics and probabilities, no matter how sophisticated, would never capture causality, not even the simple concept of 'confounding'.

9.13.2022 3:23am - (Replying to @the_heruman) Hari Seldon is futuristic, while the #Bookofwhy, so I hope, is about our everyday reality. But looking at the history of the causal revolution from antiquity till today, I agree that there is an element of tectonic struggle here, which is still raging.

9.12.2022 8:43pm - Remember the conversation I had with @MomentMagazine , about Zionophobia, antisemitism, AI and more? The video is available now, and I am happy to share: We didn't get to the AI part - coming shortly.

9.12.2022 8:14pm - Time to examine this week's production of causal inference articles: Some of the titles are extremely enticing, but I do not have the time to sort them all. Please alert us if you find new breakthroughs or new perspectives.

9.12.2022 5:07pm - (Replying to @arthur_alb1 @soboleffspaces and 5 others) I find here 15 authors & 193 references in pursuit of "patient-centered estimates of outcome risks with vs. without the intervention, taking into account all relevant patient attributes simultaneously". I don't see how this pursuit differs from standard Rung-2 Causal Inference.

9.12.2022 12:06pm - (Replying to @stuz5000 and @csabaveres) Is this a confession or an accusation? Please unpack.

9.12.2022 12:03pm - (Replying to @hohl_jeremy @truth4_today and 11 others) We are not talking "history", we are talking daily practices towards their own citizens, this day and age.

9.12.2022 6:18am - No, NOT a joke! I often think about the #HRC51 delegates from enlightened countries, say UK or France. What do they tell their children when they come home, and are asked: "Dad, did you sit with these guys in the same room and play the game without puking?" I feel for them.

9.12.2022 5:56am - (Replying to @PavlosMsaouel @WvanAmsterdam and 7 others) This is a paper even I could understand, though it is billed as "predictive", in the PATH category. I hope every oncologist reads it, including some from the PATH enterprise; it proves the usefulness of causal modeling even when data are purely RCTs.

9.12.2022 2:55am - (Replying to @WvanAmsterdam @arthur_alb1 and 6 others) Suggestion: Some readers (like me) do not know what "Offset" is, and it would be helpful to define CR-CATE earlier than bottom of page 3. It is, after all, the essence of the paper.

9.12.2022 2:43am - (Replying to @haaretzcom) Sorry, but you cannot brush off this idiotic subtitle so lightly, without informing your readers what steps were taken to prevent its author from repeating. @Ostrov_A

9.12.2022 2:33am - (Replying to @WvanAmsterdam @arthur_alb1 and 6 others) I may have been misled by the term "clinically meaningful HTE" which, for poor me, connotes "causally meaningful." Re-thinking, I still believe doing correlational modeling outside causal modeling is not wise, because we lose tools for fusion, selection-bias and missingness.

9.12.2022 2:11am - Strange, but the only thing that makes these speeches different from the ones I heard in my childhood (say 1940-1949) is the word "occupation". Change "occupation" to "Yahud (Jews)" and you reconstruct the 1947 texts of Azzam Pasha and Fawzi al-Qawuqji . What's new?

9.12.2022 1:39am - (Replying to @guicho271828 @tdietterich and @YiMaTweets) Our history of myths only attests to our craving for causal explanations and our willingness to enslave ourselves to the comfort provided by these explanations. We don't find such comfort in statistical associations.

9.12.2022 1:26am - (Replying to @WvanAmsterdam @arthur_alb1 and 6 others) I've been searching for such issues, that are particular to health care, and could not find any which would not show up in a general causal model. Even with purely RCT data, causal interactions among covariates need be modeled if one seeks criteria for clinically meaningful HTA.

9.12.2022 12:52am - Whenever I hear an anti-Israel diplomat regurgitating: "The two-state solution has reached a dead end" I am reminded of that legendary fisherman talking to the fish in his net: "I guess, buddies, our swimming exercise has reached a dead end". @Ostrov_A @EinatWilf @LawfareProject

9.12.2022 12:28am Two-state solution has reached a dead end, former Jordanian ambassador to Israel says
9.12.2022 12:28am I sort of like what the Ambassador says: "The new Jordanian position therefore needs to... focus on equal rights for both sides as a basis for any future solution", assuming of course that the right of both sides to a national homeland is counted among the many other "rights".

9.11.2022 7:46pm - (Replying to @yudapearl @arthur_alb1 and 6 others) On a 2nd reading of the Tuft's paper, its main claim seems to be: "Accounting for multiple variables simultaneously, .. is foundational to the concept of personalization in EBM". This task is standard since 1995 so, perhaps one of the 15 authors can explain: What have we missed?

9.11.2022 4:07am - And the educators among us would value this introduction: where @smueller and A. Forney share their experience teaching Causal AI for undergraduate classes.

9.11.2022 2:40am - Many new readers ask this question and, as I've done in the past, I would recommend Primer which is available free (chapter by chapter) and logically coherent, starting with the First Law and ending with Tool Kits for Mediation and Attribution.

9.11.2022 2:21am - Hilarious! They actually sing it in the UK: "O Lord our God arise, Scatter her enemies And make them fall; Confound their politics," I'd hate to be in the Queen's enemy camp and see my politics confounded by one stupid backdoor path. Unfair!

9.11.2022 2:07am - "He is a Sabra"! My aunts and uncles said when they had to apologize for my spiteful behavior. Superheroine Shira Haas, so I hope, would need no apologies.

9.11.2022 1:45am - It may not make Eleanor Roosevelt happy, but it would surely exhilarate @hrw and @amnesty . And @KenRoth will sing Halleluiah, in high falsetto!!!

9.11.2022 12:19am - The clock has just struck midnight, and has brought with it the gravity of September 11, a day that refuses to match our understanding of homo-sapiens. I cannot find a better expression to 9/11 gravity than the monument below, built in Israel, honoring every victim by name.

9.10.2022 9:34pm - A wiseman once said: "A nation without history is a nation without a future." Shortly thereafter, scholars at Al-Quds University discovered that Adam and Eve were two Palestinian philosophers who were ethnically cleansed by white colonizers while fighting for a Right of Return.

9.10.2022 2:35pm - (Replying to @AdamMilstein) It's not Marouf fault, who was just doing what he learned to do at home and in school. It's Ahmed Hussen's responsibility, who hired Marouf, and who is orchestrating the agenda of this whole department.

9.10.2022 5:07am - (1/2) As we are watching the coronation of King Charles III we should remember the poetic connection between the monarchy and causality, as orated by Bertrand Russell (1911): "The law of causality, I believe, is a relic of bygone age, surviving, like the monarchy, only because
9.10.2022 5:07am - (2/2) it is erroneously supposed to do no harm.” My thought: Had Russell been with us today witnessing the rebirth of the science of causality, he might perhaps have changed his opinion about the usefulness of the monarchy.

9.9.2022 11:05pm That's not the end of the story. The sad part is that as soon as we hear a University choking on: "These labels are antisemitic" we find dozens of Jewish organizations praising them for showing such consistent and forceful commitment to fighting antisemitism. @StopAntisemites

9.9.2022 5:30pm - This double standard is written into the code of every university that I know, and into the charter of every Equity Diversity and Inclusion office that I have had a chance to work with.

9.9.2022 5:12pm - (Replying to @blakeflayton) Where do you collect these draggers?

9.9.2022 4:33pm - (Replying to @Shayadler2) Sure, but one of Britain's narrow interests was to appease the Mufti who was leading the 1936-39 riots in Israel, and organizing the anti-British uprising in Baghdad 1941. "Enlightened intellectuals" can't accept the idea that Zionism if a home-coming endeavor - not in textbooks.

9.9.2022 2:16pm - (Replying to @WvanAmsterdam @MaartenvSmeden and 4 others) I was under the impression that "making sure the prediction models actually estimate the interventional distributions (p(y|do(x),w)) " is a Rung-1 exercise, so, we can call the resulting distribution "estimand". Why 'predictimands'?

9.9.2022 6:25am - (Replying to @stephensenn @jwbelmon and 5 others) My hope is that someone who understands both SCM and "practical real-life" RCT jargon will be able to articulate the latter in the language of SCM, and things would fall into place. This was done in (real life) epidemiology, social science, economics and more. Why not RCTs?

9.9.2022 6:14am - (Replying to @jwbelmon @stephensenn and 5 others) I could not see any methodological difference between prognosis and diagnosis. Willing to accept practical differences, but why do we have more patient-specific data available in diagnosis? Is it because the patient is already in a hospital environment?

9.9.2022 6:09am - (Replying to @stephensenn @jwbelmon and 5 others) I can't outguess what you mean by "new formulation" I need variables, events, data available, quantities needed to be estimated, etc. SCM offers us a flexible language to express any research question I can think of. Why not use it?

9.9.2022 6:01am - (Replying to @Shayadler2) Balfour did a lot for the establishment of Israel, (see but then Herbert Samuel (trying to appear neutral) appointed Amin El-Hosseini as Mufti of Jerusalem which turned Britain against the Zionist project and started a 100 year conflict.

9.9.2022 5:47am - (Replying to @stephensenn @jwbelmon and 5 others) Because I still can't cast the research question in a language that would permit scientific analysis: variables, events, what is given, what is needed, etc.

9.9.2022 5:38am - (Replying to @stephensenn @jwbelmon and 5 others) "How does this work?" is not a research question. You probably mean: "Can we eliminate the difference between experimental findings and what we wish estimated." This is the "selection bias" problem solved. If u have a different problem in mind, articulate and it will be analyzed.

9.9.2022 5:18am - (1/2) I was born in the British mandate. My father traveled on a British Commonwealth passport. We screamed our heart out: Down with the British Government! Down with White Paper! (in the bus to summer camp). British paratroopers searched our apartment (for illegal weapons).
9.9.2022 5:18am - Ettlee, Bevin, and King George VI were my sworn enemies (for blocking Holocaust survivors from reaching Israel). Yet, I feel only reverence, warmth and admiration towards Queen Elizabeth II, and sadness on her departure. She worked for a cause higher than herself. Gun Salute!

9.9.2022 2:47am - (Replying to @MaartenvSmeden @soboleffspaces and 3 others) Given that you are familiar with both SCM and "predictive models", can the problems tackled in the paper above be cast as problems in SCM? Why not? For example, can the notion of "predictimand" be translated into SCM? I did not find it well defined (formally or conceptually).

9.9.2022 2:20am - (1/2) (Replying to @stephensenn @jwbelmon and 5 others) I assume you want to find out whether tools developed in causal inference are applicable to tasks in your areas of interest? The general answer is: We define each task scientifically and appeal to SCM to find out if it can be accomplished formally. For example, question 5
9.9.2022 2:34am - (2/2) (Replying to @yudapearl @stephensenn and 6 others) pertains to a task called "selection bias"; it is defined using the language of SCM here:, analyzed and given a solution in Summary Results-2, and more powerful results here: The same protocol applies to the other questions

9.9.2022 1:58am - Excited to see this paper tackling the century-old problem of finding "the root causes of outliers" using Structural Causal Models. My only question: Why not use "probabilities of causation" to rate the many "root cause" candidates? e.g.,

9.9.2022 12:19am - (Replying to @elderofziyon) As they once said in our places: History with no truth to people with no history.

9.8.2022 11:49pm - (Replying to @MaartenvSmeden @soboleffspaces and 2 others) This paper re-enforces my interpretation here: Am I wrong? I can see a few points of intersection with causal inference: 1. explanability, 2. meta-analysis and 3. missing data. Do you see more?

9.8.2022 11:09pm - (Replying to @SaraReyi) This is what Israelis are proud of. I was born in one of those cities, Bnei Brak, which was named "Ibn Ibrak" till 1924, and Bnei Brak till 140 CE. Believe me, you wouldn't want to grow up in Ibn Ibrak, nor did the few shepherds who welcomed the water pump my family installed.

9.8.2022 10:53pm - (Replying to @blakeflayton) Why is it only "gaslighting" and not a "license for genocidal plans"? as I argue here:

9.8.2022 9:15pm - (Replying to @adrianachavira and @LAPressClub) Thank you Adriana for your mentorship and dedication to generations of students at the Daniel Pearl Magnet High School. We sincerely hope that this unfortunate incident ends with the withdrawal of all charges against you and with no spec on your record.

9.8.2022 3:35pm - I don't think it's "thoughtful" to take a country, a people and a civilization and call it a "political system" so as to make it more palatable to advocate its demise. Moreover, given how determined this people is to cling to its sovereignty, it's a license for a genocidal plan.

9.8.2022 1:41pm - Sharing a press release, jointly with the LA Press Club, regarding censorship at Daniel Pearl Magnet High School, followed by suspension of a great journalism teacher.

9.8.2022 4:54am - Call me an antisemite, or racist, or whatever, but I do feel a pinch of sympathy to these Law students. "Freedom of speech" is no reason to stop a Boycott of Zionist speakers if Zionism is indeed as dangerous as charged. It is about time Israel advocates learn to sell Zionism
9.8.2022 4:54am - on its merits, not as a necessary evil in a world that must bow to "freedom of speech". The BDS circus will continue from bad to worse until Jews gather the Chutzpa to propose a new initiative: "No Zionophobic speakers at Berkeley". I'll vigorously campaign for it @bandlersbanter

9.8.2022 4:18am - (Replying to @Ostrov_A) If you noticed, @PeterBeinart just adopted the language of the Ayatollahs. Israel is not a "country" but a "political system", a "Zionist regime" or a "Zionist entity". It must be nameless to fit into Beinart's ideology that Israel is NOT the country his grandfather prayed for.

9.8.2022 4:07am - (Replying to @sometimes_data @kilnerevonnet and 5 others) The word "useful" is rooted in "use", which connotes "actions" taken, the results of which may be good of bad. "useful" then is translated into that which helps us choose the more meritorious action, hence causation.

9.8.2022 1:56am - (Replying to @stephensenn @soboleffspaces and 3 others) Surely it is a common distinction in science, and surely a beautiful theory was developed to formalize and algorithmitize these two steps. If there is an interest among people who are dealing with these two steps in practice I'll be happy to answer any questions.

9.8.2022 1:42am - (Replying to @soboleffspaces @statsepi and 3 others) It is hard for me to admit, but we, in the "causal inference" framework, are suffering from a major problem: We see things clearly and we can't understand why other disciplines have hard time defining things as clearly as we do. When we point it out, we are accused of arrogance.

9.8.2022 12:53am - (Replying to @stephensenn @soboleffspaces and 3 others) Is this prelude a necessary step or just frill? Can it be skipped or replaced by something else?

9.8.2022 12:46am - (Replying to @statsepi @soboleffspaces and 3 others) Agree that this is obvious. Can we assume then that scholars who advocate "predictive models" have developed ways of distinguishing "usefully predictive" from "uselessly predictive"??? I would be surprised if this can be done without causal models. I'm actually convinced it can't

9.8.2022 12:30am - (Replying to @VicMaz7 and @bobjcarr) Respectfully, Mr. @bobjcarr , observe for yourself what diabolical mental aberrations your tweets tend to inspire. (E.g., @VicMaz7 .) I know it was not your intention, but "the wise has eyes in his head" said King Solomon.

9.7.2022 11:01pm - The transition from non-Zionists to all-out Antizionists is a natural self-defense mechanism whereby mental discomfort "is elevated to a noble feeling of moral purity. They [ @PeterBeinart & other Jews of Discomfort] used to feel guilty for Israel’s actions while conscious

9.7.2022 11:01pm - (1/2) The transition from non-Zionists to all-out Antizionists is a natural self-defense mechanism whereby mental discomfort "is elevated to a noble feeling of moral purity. They [ @PeterBeinart & other Jews of Discomfort] used to feel guilty for Israel’s actions while conscious
9.7.2022 11:01pm - of her problems. No More! Elevated in virtue, they now see every blemish on Israel’s face as “the litmus test” for her impure personality — hers, not theirs." (Quoted from

9.7.2022 7:58pm - I have a different take on "un-Jews". I call them "Jews of Discomfort"

9.7.2022 7:37pm - (Replying to @JamesKukula) If person B only had the decency of asking: "What does Z mean to you?" and the patience to listen to the answer, I have no doubt person B would say: "Oh, in which case Z is an important part of my identity too, sorry for jumping to conclusions based on anti-Z propaganda."

9.7.2022 6:20pm - Careful @bobjcarr , the words "chance of a two state solution" make you an "enemy of the Palestinian people" as they did to many Westerners with good intentions, who did not realize how offensive these words are to every Palestinian of good standing. My father was one. @Ostrov_A

9.7.2022 6:01pm - I haven't been to my office in weeks, I wonder if I would need armed escorts next academic year. As Kafka asked: "They eat horses, don't they?"

9.7.2022 5:37pm - (Replying to @BTreehornThug) Person A spoke for HIMSELF.

9.7.2022 5:03pm - This is an interesting philosophical question. Person-A says: Z is an important part of my identity. Person-B says: No it isn't. Question: Who is the racist?

9.7.2022 4:04pm - (Replying to @soboleffspaces @MaartenvSmeden and 2 others) I have similar set of questions, but I would summarize them as: (1) in what sense is statistical modeling "not predictive". (2) How can clinical study NOT be predictive?

9.7.2022 11:51am - And for those who have trouble understanding my Israeli accent, here is the text of my commencement speech "On Science and Freedom": It sounds my better from the paper.

9.7.2022 4:27am - And for those who have trouble understanding my Israeli accent, here is the text of my commencement speech "On Science and Freedom": It sounds my better from the paper.

9.7.2022 12:31am - (1/2) My 86th birthday has lured some readers to ask which paper, lecture, or interview I am most proud of. It took me two days to decide, and I now dare to share my choice: It is a 2007 commencement speech I gave at the University of Toronto,
9.7.2022 12:31am - (2/2) titled "Science and Human Freedom". It is not published anywhere but, since the message cuts across all the sciences, including the humanities, and since I still find it inspirational, I thought it would be useful to share with readers of the educational channel. Enjoy.

9.6.2022 10:40am - (Replying to @PhilDawid) There is no gap, because the goal was not to reach consensus but to minimize the number of things people need to agree on for full consensus to prevail, and to make those things MEANINGFUL. I'm not the one who labeled them "meaningful", it was human language, from Abraham to Hume

9.6.2022 2:12am - (Replying to @PhilDawid) Of course it must be "parial". We know from the ancients that Rung-3 claims cannot be tested by data alone. But having testable implications is merely a bonus on top of the real miracle: having a parsimonious and meaningful object to agree on before we agree on everything else.

9.5.2022 11:31pm - (Replying to @DKedmey) This question belongs to the "Probability of Causation" variety also known as "Causes of Effects" , formalized through PS PN PNS and a few more.

9.5.2022 9:29pm - This bookshelf is even more impressive than the one I have at home and, confessedly, I too find some of the old papers not easy but extremely enjoyable. What a strange combination! Thanks for this unique birthday gift. #Bookofwhy

9.5.2022 6:58pm - (Replying to @soboleffspaces) I would say "causal effect must be defined in other ways...". They can continue and argue that those "other ways" are built on assumptions they do not need in RCT. Fine. We are dealing with definitions, not estimation. This distinction is basic to the scientific approach.

9.5.2022 6:53pm - (1/2) Good question, the answer to which revolves around personalities and ego trips, so it would take 100 tweets. Imagine what a trauma it must be for Rubin to see all his Potential Outcomes, their properties and intricacies emanate from something as simple as a causal model,
9.5.2022 6:53pm - (2/2) invoking only "observable variables". It must be "bad science". Same with Angrist and Imbens, who vowed to follow Rubin no matter what. See Or, think about Heckman or even some epidemiologists & more.

9.5.2022 6:28pm - (Replying to @soboleffspaces) Agree. The backdoor criterion and the adjustment formula do not need the details of SCM. They emanate from the DAG and the DATA alone.

9.5.2022 6:21pm - It's a great historical page by @Martin_Kramer , to which I can only add that for us, who grew up in pre-State Israel, the name was as natural as mother milk; we had no doubt, nor a second thought that we are FROM ERETZ ISRAEL, what else?

9.5.2022 5:33am - 50 years to the Munich Olypics massacre compel us today to remember 7 under-reported facts about the event that has changed how we view terrorism.

9.5.2022 3:24pm - (1/2) Trite? Saying that "two people who accept Euclid axioms of geometry would also agree on all theorems about triangles" is not Trite but a Miracle. Same as saying "two people who accept Kolmogorov's axioms of probability would also agree on Dawid's axioms of conditional
9.5.2022 3:24pm - (2/2) independence" is not Trite, but a Miracle of logic. Halleluiah! The miracle is further amplified in the case of counterfactuals, because they are the building blocks of scientific thought and moral behavior. Praise to the Lord! Amen!

9.5.2022 3:09pm - (1/2) The statement may appear trite, once we accept the First Law of Causal Inference But given the number of prominent researchers who resist the First Law, or fail to appreciate its power, the statement is the opposite of trite; it is in fact a hymn to
9.5.2022 3:09pm - (1/2) the power of logic to produce answers to such a gigantic space of counterfactual questions from such a parsimonious, transparent and (partially) testable representation. Look: Y_x = Y_{M_x} the rest is just math. Halleluiah!

9.5.2022 1:23am - (Replying to @AndrewDGarrett) "hence the trial" is what we are trying to explicate. Can we examine more closely what the trial tells us about "safety features"? Can it be articulated in scientific language? Is it in any way related to Prob(harm) defined in or is it another animal?

9.4.2022 5:46pm - (Replying to @SAFECUNY) CUNY is not entirely at fault. "Jew hatred" and "antisemitism" are (purposely ) slippery concepts. Zionophobia is clear, well defined and undeflectable. CUNY should be called out for what they can't deny.

9.4.2022 5:31pm - (Replying to @DingWynn) Glad you undertook this broad survey of causal inference. My only reservation is its breath. By trying to give a voice to anyone who expressed an opinion about causation the impression is created that disagreements rule the day when, in fact, the field enjoys widespread consensus

9.4.2022 5:10pm - (Replying to @AprileBernardi) I was skeptical too, till I was told that RCTs have so many imperfections that all they can deliver is a "comparison" between groups eligible for comparison, and all attempts to rectify imperfections are based on assumptions known to be false (in the imperfect world)

9.4.2022 4:35pm - (Replying to @stephensenn) I think even a stupid guy like the one you are mocking should be listened to when he/she points out a few terms in the RCT jargon that could benefit from scientific interpretation. In particular, what RCT findings promise the public, aside from lists of individuals compared.

9.4.2022 3:04pm - (Replying to @EpidByDesign @robertwplatt and @RWJE_BA) "May Differ" does not mean "differs". S expresses subjective suspicion in possible differences. More importantly, we are dealing with "definition", not identification. If it so happens that S is too large to allow identification, researchers need to know it too.

9.4.2022 12:41pm - (Replying to @artistexyz and @soboleffspaces) It's a little hard for me to envision how system 1 can be more accurate that system 2, when the former must sprout from the latter.

9.4.2022 11:59am - (Replying to @dataengines and @stephensenn) Tiny clarification. Pearl does not say answering the other questions is not worth RCT expenses. He says absent Q5 guarantees, it behooves trialists to explicate why answering the other questions is sufficiently relevant to justify RCT expenses.

9.4.2022 1:46am - Readers seeking additional proofs, references and background material for our "Crash Course on Good and Bad Controls" may find it in this extended survey of "Confounder Selection" by Guo, Lundborg and Zhao (2022).

9.4.2022 1:05am - Agree, the cartoon does not show the line up of the other 10%, the Giordano Brunos in our community, who are driven not by thirst for confirmation, but by an ingrained suspicion of conventional wisdom.

9.3.2022 11:19pm - (Replying to @AmbDennisRoss and @Ostrov_A) I wonder what was it that we have done to give the Iranians the impression that we are so gullible? And.... aren't we still doing it?

9.3.2022 8:20pm - (Replying to @IsabellaGhement) Amazing flower. And every leaf thinks it is i.i.d. !!!

9.3.2022 8:16pm - (Replying to @blakeflayton) This will go on and on till...till we learn to charge @rogerwater and his likes with Zionophobic racism, not deflectable antisemitism. Even @rafaelshimunov would not be able then to say: "He ain't one"

9.3.2022 6:26pm - (Replying to @artistexyz @soboleffspaces and 5 others) My first reaction to the "missing data" myth was: Show me a problem that is not "missing data", eg is Chess a "missing data problem"? My latest reaction: "missing data" is a causal, not statistical problem, as shown here: it's hopeless w/o causal graphs.

9.3.2022 6:15pm - (Replying to @AdanZBecerra1) I would be enlightened with much less, e.g., if a trialist could just say who should pay attention to the trial results. The fact that "trial relevance" isn’t rigorously addressed amazes me - but doesn't surprise me given the notional and conceptual deficiency in the field.

9.3.2022 4:07pm - (Replying to @AMULETAnalytics and @uclaextension) "Tough read"? How about de-toughening it with a softer book? It shouldn't be difficult once you get the main message: speak science not data.

9.3.2022 3:59pm - (Replying to @artistexyz @soboleffspaces and 5 others) Rubin does mention graphical models, 57:00 min into the video: "The graphical guys, they mix everything up... using observed values.. confusing..where is the insight?.. where is the science?..." My wish: May his words be available to historians of 21st century statistics.

9.3.2022 1:38pm - This construction gives us X=const1 Y=const2, but the CBN: X Y says: X is a random variable distributed as P(X) (same with Y). We need a repair.

9.3.2022 8:08am - (Replying to @artistexyz @gottfriedmath and 6 others) I don't see how we can define SCM from CBN. Here is a CBN: X Y no arrows. What is the SCM associated with it?

9.3.2022 1:26am - My My! What a wonderful gift for my 86-th birthday tomorrow. Knowing that some bookshelves on planet earth carry my thoughts makes me feel immortal and doubly committed to spending the next 34 years of my life searching for the calculus of our mind. Thank you!

9.3.2022 1:13am - (1/2) Clarifying. Of course student groups have a legal right to be selective. But University administrators have a legal obligation to set norms of discourse on campus and tell them publicly: "You are acting against the values of this great university". Care to see it in action?
9.3.2022 1:13am - (2/2) Change "Zionists" to "Palestinians", and watch how haven and earth roar with indignation as Ken Stern is barred from ever speaking on a US campus again.

9.3.2022 1:09am - (Replying to @Claire_V0ltaire) Roger is right, he is just a Zionophobic racist, not an anti-Semite. What allows him to continue and fool the world with this mask is Jewish leaders refusal to treat Zionophobia as an uglier form of racism.

9.3.2022 12:37am - Bad analogy, @Ostrov_A . It invites Ken Stern to argue: Aha! Race and gender are immutable, Zionism is just a political belief. But change "Zionism" to "Palestinians" and watch how haven and earth roar with indignation, as Ken Stern is barred from speaking again on any US campus.

9.2.2022 10:36pm - I have often been thinking how strange it must be for a Palestinian pupil to sit in a "Palestinian History" class and listen to teachers who never had such classes.

9.2.2022 8:50pm - I would agree with this analogy if you can present a mathematical alternative to SCM. I am not aware of any such alternative, unless we count an exhaustive list of all counterfactuals as an alternative.

9.2.2022 5:19pm - I think the source of the confusion lies in assuming that an SCM is defined for one individual u. It is in fact defined for EVERY conceivable u. So, nothing is missing nor hidden.

9.2.2022 4:50pm - (Replying to @artistexyz @soboleffspaces and 5 others) Mohan & Pearl's paper on missing data concludes that missing data problems are causal problems, namely solved within the SCM paradigm, not the other way around. IOW, we do not have a calculus for missing data that is independent of the calculus of SCM.

9.2.2022 2:37pm - (1/3) I agree. I don't know of any causal quantity that remains undefined in SCM, assuming of course that we take the relation "listens to" as a primitive for defining SCM. As for RCT, the proof in #Bookofwhy is the only defensible proof I know that RCT delivers (asymptotically)
9.2.2022 2:37pm - (2/3) the Average Causal Effect (ACE) of treatment on outcome. It's strange indeed that ACE isn't treated in the RCT literature as the ultimate goal of the RCT exercise. Taking "comparison" as the goal of RCTs hardly justifies their "gold standard" stature, not to mention expense.
9.2.2022 2:37pm - (3/3) This is even less justified when the goal is cautiously declared to be "comparison between the two groups participating in the trial" with no reference to a policy decision, nor to the target population eventually affected by the conclusion of the RCT study.

9.2.2022 6:20am - (Replying to @cristina__m16) I am not sure what PCA is. But a prequisite for causal interpretation of regression coefficients is either (1) physical intervention or (2) causal model.

9.2.2022 3:34am - I remember how this "3-No's" resolution triggered the West-Bank settlement movement: If we are destined to live by the sword to the end of time, we better do it from a position of strength. Official Arab League renunciation is way overdue.

9.1.2022 7:19pm - (Replying to @YanChan98780882) Which napkin? the old or the new?

9.1.2022 2:46am - I believe we tried to submit to a more general journal, and it got rejected. I fully agree; should be available to every researcher who resorts to causal models. Please help us retweet it. @AmJEpi , @tweet_epi @TheEconomist , @causalinf @EpiEllie #Bookofwhy

9.1.2022 7:18pm - (Replying to @EpiEllie @AmJEpi and 3 others) Not sure I see the confusion. [By definition, confusions are hard to see]

9.1.2022 7:06pm - (Replying to @soboleffspaces @stephensenn and 3 others) I have noticed the same. As a matter of fact does anyone know of a proof that RCT delivers ATE? The only attempt I know of is Rubin's, which assumes Y(1),Y(0) || X , an assumption manufactured to get the proof going. Any trialist for rescue?

9.1.2022 6:49pm - (Replying to @jaredlcm) Not sure who the "grab-bags" are, but I surely adore Holmes' hat and Bayes' advice, though I'm only "Half Bayesian":

9.1.2022 6:40pm - (Replying to @rjherber) The reason J. Street criticize him, they believe his allegations are not totally removed from reality, so, as yester-Jews, they must exonerate themselves from the terrible "crimes" that Israel is committing. They wouldn't know how to answer another comedian. @JeremyBenAmi

9.1.2022 5:08pm - Arab-American Comedian Frustrated in Attempt to Bait Brooklyn Jews Over Israel-Palestinian Conflict
9.1.2022 5:08pm - I almost guessed correctly! I thought Mula was a Jewish comedian trying to impersonate Palestinian mentality. He turned out to be impersonating is own mentality on an Orthodox Jew, who knew precisely how to bounce back Zionophobic cliches: What war? In Ukraine?

9.1.2022 3:15pm - Honored to be invited as keynoter for the Causal Data Science Meeting 2022, and meet our co-principal-investigators from across the isle. In 2023, we won't call it "Causal Data Science" any more -- it would be totally redundant.

9.1.2022 1:12pm - That's the best answer i've seen thus far. I would only replace "antizionism" with "Zionophobia", to make it clear we are not talking philosophy, but about a morally disgusting, genocidal ideology.

9.1.2022 1:45am - (Replying to @angelamczhou @artistexyz and @pablogerbas) Hines et all is a good survey, thanks.

8.31.2022 10:41pm - After weeks of uncertainty, 1st day of school begins for 2.5 million Israeli pupils via @timesofisrael
8.31.2022 10:41pm - Nostalgia! My heart jump from joy, re-experiencing this most memorable day -- the first day of classes, for the children in Israel. No strike, no politics, just a smiling teacher welcoming us with: Shalom Kitah Aleph. (Shalom, first class!)

8.31.2022 8:00pm - Hurray, our Crash Course is finally published:, which makes it not only "looks super useful" but "truly super useful". Confessedly, I often wake up wondering how perfectly good people can walk through life without it. (Names deleted: reviewer-phobia)

8.31.2022 7:34pm - (Replying to @angelamczhou @artistexyz and @pablogerbas) Is there anything special about the form of your density functions that would prevent you from using the method to estimate ANY density function from its samples? If not, what makes your method "causal" or "counterfactual", rather than "statistical"?

8.31.2022 7:15pm - (Replying to @angelamczhou @artistexyz and @pablogerbas) I missed this part in your paper, partly because I could find no graphs there, nor an ignorability assumption usually invoked by graph-avoiding authors.

8.31.2022 3:27pm - I heard Biden no longer reads Fox News, so I've hoped CNN would echo this urgent letter. All I see on CNN is Trump Trump Trump, as if the Ayatollahs promised not to attack Israel till the US elections.

8.31.2022 2:58pm - Hilarious! I'm pretty sure the guy on the right is a comedian, impersonating Palestinian mentality. And the guy on the left (with the hat) is the producer of the clip: You mean Ukraine? Hilarious!

8.31.2022 4:20am - (Replying to @CotlerWunsh) Why? Do you know a West Bank Mayor who thinks differently?

8.31.2022 4:11am - I once presented this question to a reasonable colleague I've chanced to meet. His answer: Pedagogically, it's seductive to start with "If I treat someone, he either dies or not. Same if I don't," The rest is just missing data.

8.31.2022 3:28am - The economic situation in Lebanon must be super rosy for its Energy Minister to take time off and play kindergarten near the Israeli border. Let's hope he is allowed back to office by Nasrallah.

8.31.2022 1:15am - (Replying to @soboleffspaces @artistexyz and 2 others) To make it even clearer, I would say: "L3 answers another question: How likely is it that *THIS SPECIFIC* outcome would have been different if the treatment had been different? Once we condition on the observed outcome, we can no longer express the question in do-notation.

8.30.2022 11:00pm - Are they considering twinning with Shderot? After a decade of Hamas rockets they can surely use some Scottish support.

8.30.2022 10:27pm - (Replying to @stuz5000 @artistexyz and 2 others) Thanks for clarifying. I get many annoyed tweets which I cannot decipher.

8.30.2022 9:10pm - (Replying to @angelamczhou @artistexyz and @pablogerbas) I just noticed that you are a co-author on this paper, so perhaps you can resolve our puzzle. What is the input to your system? Do you estimate ATE from RCT data? Or from some other kind of input. If the former, what makes it "causal"? Isn't it standard stat. estimation?

8.30.2022 3:20pm - One more mini-tweet: "And the U.N. expects humanoids to takes its other "commissions" seriously!"

8.30.2022 3:04pm - (Replying to @RabbiWolpe) The other one is "Chutzpa". Which only Larry King was able to define: "Hadassah organizing a fund-raising event in Khadafi's Libya", not as a provocation, but because it's a really "noble cause."

8.30.2022 10:43am - (Replying to @analisereal @artistexyz and @soboleffspaces) Can you give us a 280 characters description of what "influence functions" are, where they come from, and what ordinary mortals should think about when someone tells them "influence functions".

8.30.2022 10:00am - Beg to disagree. Pearl does not claim that "you can only do counterfactuals using SCM." He claims that "if you want to understand what you are doing, e.g., defend or test your assumptions, then you need to know that counterfactuals come from SCM." See

8.30.2022 9:48am - (Replying to @artistexyz and @soboleffspaces) ATE too is Rung-2. To avoid such confusions, I recommend using do-expressions for Rung-2 and subscripted (or parenthetical) expressions for Rung-3. But isn't the paper really Rung-1, given the identification is assumed away?

8.29.2022 11:49pm - The long abandoned children of Shderot truly deserve a visit of their heroine Gal Gadot. She grew up among them (figuratively), we all did, and we all know what they had to experience past year. Still, the NYT considers their plight unworthy of a mention. Thank you, @GalGadot

8.29.2022 10:09pm - Glad to receive a confirmation from someone who understands the paper. I didn't, because even the Interventional mean in Example 1 is ill defined; the covariate X must be back-door admissible -- an assumed-away criterion that probably applies to all examples.

8.29.2022 9:59pm - It's linking to Causality chapter 1

8.29.2022 9:51pm - Anyone claiming to be doing counterfactuals using Bayesian Networks (Rung-2) should be challenged with the example on p. 35-36 here: Jordan's paper defines counterfactuals as E[Y(1)] which is Rung-2. I do not see any R-3 counterfactuals, say E[Y(1)|Y].
8.29.2022 9:51pm - It's linking to Causality chapter 1

8.29.2022 9:16pm - I hope I would be able to talk more about this balcony tomorrow, for @MomentMagazine , 1:30 pm ET Perhaps also to explain why "Zionophobia = Racism", not "anti-Zionism = antisemitism", is a more effective vaccine against the new virus.

8.29.2022 9:00pm - The delegates at yesterday's Congress made a mistake. The boasted about going there to celebrate, not to pass resolutions. The should have passed one resolution: "We hereby declare anti-Zionism to be a racist ideology, to be named Zionophobia". We need a Herzl to push it through.

8.29.2022 8:48pm - Well put: "Erasure and demonization". BDS cronies are doing the latter, while University administrators are doing the former - perfect symbiosis. In vain you'd search for the words "Jewish identity" in the records of any Equity, Diversity and Inclusion officer in any university.

8.29.2022 8:32pm - Herzog reenacts photo of Herzl on balcony while attending Zionist congress anniversary via @timesofisrael
8.29.2022 8:32pm - I can't take my eyes from these two photos. The left was always hanging on the wall in my school days, the right one was taken yesterday in Basel. What one balcony can accomplish!!!

8.29.2022 5:01pm - A Zionophobic initiative at UC Berkeley calls for "No Zionist Speakers" It has yielded two positive results: (1) The Dean spoke in first-person: "I am a Zionist, can I speak in my own school?" (2) The actual definition of Zionism finally came to light.

8.29.2022 4:19pm - Time to examine this week's harvest of Causal Inference articles: Many of the titles are seductive to my curiosity, hoping to find new ideas, methods or applications. But time constraints behoove me to seek your help in this exploration. Keep us informed.

8.29.2022 3:28pm - My faithful calendar sings: Tomorrow, Aug. 30, I will be talking (Zoom) with @NadineEpsteinDC about antisemitism, Zionophobia and even AI. Moreover, why we should fight each virus with its own potent vaccine. @JGreenblattADL @DanaBashCNN @wolfblitzer @deborahlipstadt @jaketapper

8.28.2022 11:41pm - (1/2) My dear friend @MissDiagnosis has just found a causal explanation to a problem I could not reconcile: "Why do reasonable progressives stay silent when Zionists are unfairly attacked?" Her answer: They are afraid to be accused of racism and lose
8.28.2022 11:41pm - (2/2) their jobs if they defend someone already accused of racism. So, pause before you decide to defend someone like me who, together with my Zionist colleagues at UCLA, were accused of such horrific crime by a visiting lecturer. Don't risk your job; instead, let me confront her.

8.28.2022 9:59pm - (Replying to @JGreenblattADL) Herzl: “Zionism is a homecoming to the Jewish fold even before it becomes a homecoming to the Jewish land.” Amazing how right he was!

8.28.2022 9:48pm - (Replying to @GeraldNGOM) Herzl: “Zionism is a homecoming to the Jewish fold even before it becomes a homecoming to the Jewish land.” Amazing how right he was!

8.28.2022 9:38pm - (Replying to @IzaTabaro) Herzl: “Zionism is a homecoming to the Jewish fold even before it becomes a homecoming to the Jewish land.” Amazing how right he was!

8.28.2022 9:30pm - (Replying to @JGreenblattADL) Herzl: “Zionism is a homecoming to the Jewish fold even before it becomes a homecoming to the Jewish land.” Amazing how right he was!

8.28.2022 9:16pm - (1/2) Unable to personally join today's historic celebration in Basel, Switzerland, I wish nevertheless to contribute a milligram of reflection, and retweet my story about the First Basel Congress that took place 125 years ago: Historians seem to miss
8.28.2022 9:16pm - (2/2) Herzl's final words, which he made on the morning of Aug. 29, 1897. After 15 minutes of wild cheering, Herzl took the stage and said: “Zionism is a homecoming to the Jewish fold even before it becomes a homecoming to the Jewish land.” Amazing how right he was!
8.28.2022 9:16pm - (Replying to @yudapearl) Done

8.28.2022 5:47am - (Replying to @MariDTakashima and @ChristinABuelow) Welcome to the island of commonsense. The food is not exactly on the gourmet side, and the sharks take their toll, but the good company makes up for it.

8.28.2022 5:07am - (Replying to @ChristinABuelow) Keen observation and, true, I don't think I would ever get to ask the questions I did ask if I were say a statistician or economist. It was quite early that I was shaken by the realization that we are causal machines and, therefore, AI should become a causal technology.

8.28.2022 12:45am - To me it does not sound antisemitic. To me it sounds like saying "People of wisdom, moral integrity and deep commitment to peace control the US government" which is a blessing of good fortune for America. I wish it were true. @AdamMilstein

8.28.2022 12:13am - Great! I was always suspicious of "Degrees in Conflict Resolution". All the failed "peace makers" (John Kerry, Denis Ross, etc) had such "degrees". But with Maitha graduation I begin to think: Perhaps there is a solid body of skills and knowledge deserving of the title!

8.27.2022 11:07pm - (Replying to @elderofziyon and @jjz1600) We must add @jjz1600 to the list twisted minds who rush to label Israel "apartheid" and who fail to understand the harm they are causing to those who truly experienced it. Sad to see a word once stood for human struggle for equality turn into a word soiled with twisted minds.

8.27.2022 10:14pm - This is a profound statement that has to do with the monotonic character we wish information to have. The things that a message excludes increase monotonically as more evidence accumulates, not so the things that a message permits.

8.27.2022 10:06pm - (Replying to @DrMikeH49 @AvivaKlompas and @Claire_V0ltaire) Thanks for the link. Chemerinsky's semi-encouraging words can be explained by the kind of Zionist he is. Instead of saying "I am a proud Zionist myself, who isn't?" he had to bow down and beg for legitimacy saying that he “condemn[s] a lot of Israel’s policies". Now he is Kosher.

8.27.2022 9:04pm - (Replying to @blakeflayton) How proud He would have felt.

8.27.2022 7:19pm - (Replying to @harrysfrand and @ilovemilfrryy) Some people can't grasp that "Peace sign for Palestinians" is an insult to the "Palestinian narrative" and a death warrant to the "Palestinian struggle". "From the River to the Sea, Peace will not replace us!" @Eve_Barlow

8.27.2022 6:20pm - Agree. Taking counterfactuals out of history books would be worse than taking Algebra out of science.

8.27.2022 3:57pm - (Replying to @tdietterich @khademinori and @ChrSzegedy) The question is: What have we learned from the outcome of the first game, that we did not know before hand? If the answer is: That the opponent has weakness W, then the question is counterfactual. Otherwise, the first game is redundant and our question is: find effect of move M.

8.27.2022 3:46pm - (1/3) To explain the virtues of graphs in missing-data problems, here is what we write in our paper: "The transparency associated with graphical models stems from three factors. First, graphs excel in encoding and detecting conditional independence relations, far exceeding the
8.27.2022 3:46pm - (2/3) capacity of human intuition. Second, all assumptions are encoded causally, mirroring the way researchers store qualitative scientific knowledge; direct judgments of conditional independencies are not required, since these can be read off the structure of the graph. Finally,
8.27.2022 3:47pm - the ultimate aim of all assumptions is to encode “the reasons for missingness” which is a causal, not a statistical concept. " We've added this paragraph to pacify a reviewer who claimed: "I don't need graphs to think about conditional independencies". Yes, reviewer, you need!

8.27.2022 3:30pm - (Replying to @tmorris_mrc @dailyzad and @stratosinit) If it's true, then you would be the first humanoid I met who claims to feel more comfortable thinking about conditional independencies (among observed and unobserved events) in his head than with causal graphs. Can we test it on some missing-data problem which we both understand?

8.27.2022 3:13pm - (Replying to @yisongyue @tdietterich and 2 others) This is "effect of action" not a counterfactual question, unless you learn something from your actual move that you did not know before hand.

8.27.2022 5:06am - (1/2) You see why I fight Zionophobia instead of antisemitism? If you fight antisemitism you are automatically "a threat to human rights" and @PeterBeinart says: "I'm Jewish, it ain't me". If you fight Zionophobia, Beinart knows that it is precisely he who is charged with
8.27.2022 5:06am - (2/2) this disgusting form of racism and, funny thing, he can't say "It ain't me". It's you! @PeterBeinart ! @EinatWilf @JGreenblattADL @AJCGlobal @MiriamElman @michaeldickson @DanaBashCNN @jaketapper @bandlersbanter @BretStephen_NYT @bariweiss

8.27.2022 2:56pm - (Replying to @SamelDavid and @PeterBeinart) There is a typo in your tweet. The belief that Jews should have equal rights with others in called "Zionism", also going by the slogan: "Two states to two peoples, equally legitimate and equally indigenous", which Zionophobic bigots resist tooth and nail.

8.27.2022 2:34pm - Any debate related to #LordsParadox is worth watching, especially one that concerns inference about "change". #Bookofwhy

8.27.2022 1:14pm - A Machine Learning class in Computer Science that contains a session on Causal Inference may serve as a role model to ML instructors everywhere. It even comes close to the First Law of counterfactuals.

8.27.2022 2:25am - (Replying to @dailyzad @tmorris_mrc and @stratosinit) We speak different languages. I am a problem solver, not a promoter anyone's work. When a speaker discusses a problem to which a solution exists it is very natural for those who know the solution to comment constructively and respectfully on such existence.

8.27.2022 1:38am - (Replying to @tmorris_mrc and @stratosinit) I can't imagine anyone not being courteous to me. But I am not talking etiquette. I am talking answering questions about missing data methods that cannot be answered by conventional frameworks. So, I enjoyed thinking how ecstatic the audience would have been with this progress.

8.27.2022 12:01am - The "Haavara" agreement between the Jewish Agency (ie Arlozoroff) and Germany not only saved 60,000 Jews from certain genocide, but also created jobs that enabled 200,000 Eastern European Jews to escape, without whom Israel would not have survived the genocidal assault of 1948.

8.26.2022 10:32pm - (Replying to @PollakDan and @mishtal) That's possible, but it could not be the whole story; I have had some loud Zionophobic followers 2-3 years ago who resigned to silence, once they understood how they are perceived in the rational world.

8.26.2022 5:10pm - (Replying to @alexwareb9) If every entry in a Table says "regression," it is perfectly justified, I think, to summarize the Table as "just statistics". I would love to hear another interpretation which sees the Table as a mixture of statistical and non-statistical considerations.

8.26.2022 4:30pm - (Replying to @WickedViper23) Agree. Causal inference cannot be treated as a patch to Machine Learning (ML), nor can it be taught as a special variant within the ML paradigm. It must be seen as a new data-science paradigm, in which ML routines play a symbiotic role with model-based theories.

8.26.2022 12:45pm - (Replying to @WilliamWonkaJr) The whole point of using the word "phobia" is to exclude certain opinions from the sphere of "political differences", i.e., modifiable by rational arguments, and highlight their "religious" of "immutable" character. Zionophobia has this irrational character in some people.

8.26.2022 1:43pm - (Replying to @Eve_Barlow) Shabbat Shalom princess, Avi Hitpalel Va'Yechulu Hashamayim VeChol Tsavaam

8.26.2022 1:39pm - The references I have for RCT failures are: (1) Deaton and Cartwright (2018) and, (2) My commentary on (1),

8.26.2022 12:56pm - Thanks for resurrecting this entry from @Quora : It is relevant to our current discussion of ML vs. Stat, and faithfully reflects my current understanding of the issues.

8.26.2022 12:45pm - (Replying to @WilliamWonkaJr) The whole point of using the word "phobia" is to exclude certain opinions from the sphere of "political differences", i.e., modifiable by rational arguments, and highlight their "religious" of "immutable" character. Zionophobia has this irrational character in some people.

8.26.2022 12:31pm - Several readers called my attention to a tiny error in Eq. [9]; the summation should range over z, not over s. This, and other errors have been corrected, in red, on the final version of our PNAS paper

8.26.2022 10:24am - (Replying to @konohas_itachi1 and @WickedViper23) We all agree that the equation ML=stat is wrong. My question was/is: "How come we haven't seen massive protests from #datascience folks against the table that equates ML with a bunch of regressions?"

8.26.2022 3:42am - A beautiful summary of what being a Zionist means, and why the word "zionophobia" drives the conversation from what Israel does to what Israel is, casting the former in the context of its neighbors' rejection of the latter.

8.26.2022 2:53am - (Replying to @WickedViper23) "ML = stat" is nasty because it reduces ML to tasks achievable by statistical methods. This rules out many ambitious goals on the aspiration list of ML researchers, for example, causal analysis, data fusion, transfer learning, explainability etc

8.26.2022 1:21am - Chapter 6 of Causality should be instructive to our discussion It is titled: "Why there is no statistical test for confounding, why many think there is, and why they are almost right"

8.26.2022 12:32am - This overview confirms a nasty allegation that machine learning is just statistics on steroids. I hope someone from #datasciece rebels.

8.25.2022 11:43pm - Insightful analysis of a recent Count decision. My take: employees who object when the boss peddles disparaging comments about their race should argue rather than sue. Not finished: Students who find a teacher doing it should demand the President's resignation, rather than sue.

8.25.2022 9:10pm - (Replying to @osazuwa @soboleffspaces and 2 others) "Inductive bias" is a term used by some deep learning folks to say: We know that certain tasks must be done by causal modeling but, since we feel awkward doing causal modeling, let's call it "inductive bias"; it sounds like home-grown "deep learning" and makes CI part of DL.

8.25.2022 4:34pm - Just Like with Afghanistan, Biden Must Have His Deal with Iran via @jewishjournal
8.25.2022 4:34pm - Yet, unlike afghanistan, Biden won't have the excuse: "No one warned me".

8.25.2022 4:21pm - (Replying to @osazuwa @soboleffspaces and 2 others) Last time we discussed "inductive bias" I felt that it is a catch word that does not covey more information than, say, "be careful" or "some theories are a priori preferred to others" or "it's all magic". I prefer therefore to talk SCM, and identify features needed for each task.

8.25.2022 1:29pm - The twisted minds who decided to label Israel "apartheid" did not foresee the harm they were causing to those who truly experienced it. Sad to see a word standing for human struggle for equality turn into a word soiled with twisted minds.

8.25.2022 11:48am - (1/2) Students asking for "empirical rule for the presence of confouding" are probably seeking some pattern in non-experimental data that detects confounding. Such a pattern does not exist, since "confounding" is a causal, not statistical notion. Eq. (8) is only valid if we have a
8.25.2022 11:48am - (2/2) causal model according to which Z is backdoor sufficient. If they allow experimental data, then the inequality P(y|do(x)) =/=P(y|x) for any (y,x) should give them the empirical rule they are looking for. I would call it "empirical test".

8.25.2022 1:29am - As we remember the 1929 Hebron massacre, we should not forget that it was ignited by the Mufti Haj Amin El-Husseini when he came up with an ingenious theory "The Jews are destroying Al Aqsa Mosque." The Mufti is forgotten, but his invention remains with us to this very day.

8.25.2022 1:19am - It's the first time in my life that I wish I could believe Saudi sources.

8.25.2022 1:16am - (Replying to @GadSaad and @NatureHumBehav) I've been laboring for two decades to define what "harm" and "benefit" means in toy problems, with 3-4 variables. Now, these editors claim they can predict "harm" and "benefit" for the entire human race, for the foreseeable future and beyond. I wish I was an "Editor".

8.25.2022 12:44am - (Replying to @Eve_Barlow) Deplorable? Yes, but they have been saying it for the past 100 years! There is virtue in openness - others plan in sobriety.

8.25.2022 12:32am - USC Hillel Vandalized via @jewishjournal
8.25.2022 12:32am - (1/2) Glad the USC story made it to the press and, reading through the responses of Jewish leaders, I seem to be the only one to focus on the core issue: "Zionophobia". Moreover, as I read the convoluted way Dean Soni Varun defined anti-Zionism, I get the strong suspicion that
8.25.2022 12:32am - faculty members were not consulted on the final version of the document. It views Zionism as a rare Jewish affliction, begging protection, rather than an organic part of Jewish identity that is a valuable campus resource. Where do they train Deans?

8.24.2022 8:34pm - It is not rocket science. The propensity score criterion is good old fashion "backdoor admissibility", as explained here: Perhaps this is why Donald Rubin saw it as "nonscientific ad hockery"

8.24.2022 6:43pm - (Replying to @elderofziyon) Makes me wish to be a "settler".

8.24.2022 6:39pm - (Replying to @blakeflayton) Natali and @MassarahMikati are exhibit one of what being "pro-Palestinian" means to those who claim to be ones. The first jokes about Jews in the ovens. The second, more refined, bemoans the firing of the first and tells her readers "Injustice! Just for being pro-Palestinian"

8.23.2022 8:23pm - Reminds me how I got into computer-chess (in 1980's) because I wanted to learn about how our mind works.

8.23.2022 8:17pm - We are witnessing an interesting tectonic shift: As religious leaders become more tolerant, secular "scholars" become more intolerant. Here is an example from a "scholar" of Middle East Studies:

8.23.2022 7:43pm - (1/3) This incidence coincides with USC's President Folt's Letter, expressing her commitment to "continue to work" with a committee she appointed -- not one word on whether Zionist students are welcome to USC. The committee recommendation also came out today
8.23.2022 7:43pm - (2/3) It's much better than expected, for it mentions anti-Zionism 12 times. However, instead for declaring anti-Zionism an unacceptable form of hate, it recommends: "Acknowledge explicitly that anti-Zionism can sometimes be experienced as antisemitism in responding to issues and
8.23.2022 7:43pm - (3/3) incidents." In other words, the committee acknowledges explicitly its commitment to remaining as vague and inexplicit as possible in order to keep the "committee" in session, and resist defining anti-Zionism as an assault of one's identity. Sad. @bandlersbanter @DionJPierre

8.23.2022 4:29pm - Great joke/cry! I bet it would have taken them four years to deliberate had the employee worn a blue-and-white yarmulke. Life-affirming symbols are incompatible with the Museum's goal of "neutrality". Holocaust Museums in the US suffer from a similar "neutrality" hang-up.

8.23.2022 3:00pm - (Replying to @PavlosMsaouel @statsepi and 2 others) Agree. Antagonistic dismissals such as "unprincipled.. confused...non-scientific ad hockery" have mellowed down in statistic journals to read: "not useful ... unhelpful". For background context, here is my rebuttal of Rubin's paper: (rejected by SIM).

8.23.2022 2:21pm - Apropos, our paper on Personalized Decision Making got updated and Arxived, so here it is for your continued enjoyment: A conceptual introduction in full nakedness.

8.23.2022 2:08pm - Sharing two new papers from our brewery: * A. Li and J. Pearl "Unit Selection with Nonbinary Treatment and Effect," * A. Li and J. Pearl "Probabilities of Causation with Nonbinary Treatment and Effect," Enjoy, share and use.

8.23.2022 12:54pm - What is Zionophobia and how is it different from Antisemitism—and what can be done about it? I am grateful to @MomentMagazine for starting a public conversation of these (and other) questions.

8.23.2022 12:13pm - (Replying to @statsepi @MaartenvSmeden and @robertwplatt) I took the time, not to cast dispersion, but to congratulate in good faith what I thought was a change in editorial attitude. Evidently, the change has occurred already w/o my knowledge. Thanks for correcting my perception. No need to jar, let's enjoy the change, and build on it

8.23.2022 11:00am - (Replying to @JGreenblattADL @ADL and 2 others0 I'm surprised @mitchellreports and especially @MSNBC did not cut you off when you mentioned anti-Zionist hate coming from their buddies on the Left. Expect @MehdiHasanShow to make up for it.

8.22.2022 9:37pm - I'm honored to mutually follow @Kasparov63 , a master of thought and a champion of freedom. I hope I could help in areas where I've spent some time exploring. #Bookofwhy, *

8.22.2022 6:06pm - (1/2) It is Monday again, our day of examining the causal inference papers that this week has produced: Let us know if you spot a conceptual breakthrough. My eyes got caught at the last paper "Causal Pathways Analysis.." only to discover a whole field of
8.22.2022 6:06pm - political science research, struggling with "pathways" problems without d-separation, without backdoor criterion, and without counterfactuals. I speculate that they could be far ahead today if the literature of causal inference was made available to them. Anyone knows this field?

8.22.2022 4:02pm - (Replying to @statsepi @MaartenvSmeden and @robertwplatt) What does it have to do with getting a space to rebut the "unprincipled" allegation?

8.22.2022 1:45pm - Typical of racist minds is telling others where "home" is, contrary to where they "feel at home". My late wife came from Iraq, a place she always remembered as "never felt as home." Go explain it to the authors of "this must eventually happen."

8.22.2022 1:31pm - (Replying to @robwilliamsiii) I asked it in a very specific context, where the absence of an arrow from X to Y means "Y is not listening to X"

8.22.2022 1:23pm - (Replying to @robertwplatt @statsepi and @MaartenvSmeden) As I wrote to @statsepi , I am taking back my "more receptive" to read "receptive". I haven't recovered yet from Stat in Med rejection of my rebuttal to Rubin's where he labels DAGs "an unprincipled and confused theoretical perspective." Time flies.

8.22.2022 8:07am - Joining this Intl. Day of Remembrance and Tribute to the Victims of Terrorism, I remember my son, Danny, and I re-post my WSJ article "Daniel Pearl and the Normalization of Evil" "When will our luminaries stop making excuses for terror?"

8.22.2022 7:17am - (Replying to @statsepi @MaartenvSmeden and @robertwplatt) Thanks for the survey. I am taking back my "more receptive" to read "receptive". I guess I haven't recovered yet from Stat in Med rejection of my rebuttal to Rubin's where he labels DAGs "an unprincipled and confused theoretical perspective."

8.22.2022 6:34am - (Replying to @DanaBashCNN and @CNN0 @DanaBashCNN , I am glad to see that, like me, the majority of your followers have issue with hesitating expressions such as "nuanced" "emerging" "more complicated" when applied to the Zionophobic component of contemporary antisemitism. I hope you agree that a follow up is due.

8.22.2022 3:46am - (1/3) I watched it twice, @DanaBashCNN , and I believe we need a follow up, to focus more meaningfully on contemporary antisemitism, its anatomy and dangers. While it was refreshing for me to hear several speakers explicitly addressing antisemitism on the left (especially Alyza
8.22.2022 3:46am - (2/3) Lewin and @cassieblotner ), no one mentioned the most pervasive form of contemporary antisemitism: The obsessive denial of Israel's right to exist, the harassment of those who promote that right, while taking cover from charges of antisemitism by expressing infinite love
8.22.2022 3:46am - for Jews in general. Painfully missing were actionable ideas on protecting students affected by this form of contemporary antisemitism. The ideas presented, extensions of outdated traditions, have proven ineffective and new ones must be heard. @JGreenblattADL @deborahlipstadt

8.22.2022 2:44am - Congratulations to @MaartenvSmeden for assuming an associate editor role at Statistics in Medicine. I hope it marks a more receptive environment for papers in #causalinference

8.22.2022 2:26am - (Replying to @blakeflayton) I've noticed that, sadly, no one interviewed by @DanaBashCNN even mentioned the dangers posed by contemporary antisemitism: The denial of Israel's right to exist and loving non-Zionist Jews. Worse yet, @rabbijilljacobs gave this danger an explicit pass. Sad!

8.22.2022 2:07am - (Replying to @SAFECUNY @PSC_CUNY and @EricDinowitzNYC) Your president is getting closer to understanding "Jewishness". While he is wrong stating: "Fidelity to the state of Israel in itself becomes a kind of Litmus test for one’s Jewishness,” seeing Israel as the homeland of the Jewish people is a natural attribute of being Jewish.

8.21.2022 8:12pm - (Replying to @TriciaRoth18 and @DanaBashCNN) I feel it was a missed opportunity to deal head-on with the plague that Cassie Blotner described: Anti-Zionism that delegitimizes Israel & does NOT crossover to classical antisemitism. @rabbijilljacobs saw no evil in that kind of racism, all she cared about were AS tropes. Sad!

8.21.2022 5:59pm - (Replying to @FJnyc and @Bundeskanzler) We perfectly understand why Abbas needs to blame someone for the Holocaust -- he knows the active role Palestinian leadership played in blockading the Mediterranean shores, and preventing Jews from escaping Europe. If I were in his position I would blame mother Theresa.

8.21.2022 5:49pm - (Replying to @FJnyc @erikbiz and 2 others) Of course antizionism is antisemitism, why else would they deny Jews normalcy among nations? Trouble is, even a hard Zionophobe like Bella Hadid insists on being a lover of Jews, namely, Jews who know their place. The word Zionophobia sums it all up.

8.21.2022 4:01pm - I understand that @DanaBashCNN documentary on antisemitism will air tonight, 9pm EST, on CNN. She also has an interesting personal story Both King and Bash will occasionally be missing the key issue, lacking the key word: Zionophobia. I'd forgive them.

8.21.2022 7:41am - (Replying to @thomanish @FredOswald and @CortRudolph) Agree. And there is an important expansion of path-tracing in Highly recommended for SEM researchers and other sailors.

8.21.2022 7:28am - (Replying to @Ostrov_A) Good point. According to @HRW , there aren't any Palestinian terrorist organizations in existence, since they all call themselves "peaceful civil society", and none such organization has ever been suspected by @HRW . Lucky are Israeli children watched by HRW

8.21.2022 6:57am - (Replying to @BeckyCNN @hrw and @saribashi) Come on, @BeckyCNN , you spoke to @HRW program director @saribashi about anything related to Israel? You could not find someone more hostile? I would check her records first:

8.21.2022 4:33am - (Replying to @elderofziyon) Body counts do not define a Holocaust, intentions do.

8.21.2022 3:44am - (Replying to @tunguz) You are a singularity in your field; please do not give up. That is why our oath, in causal inference, is: "think deeply about what actually needs to be done" before you start. #datascience

8.21.2022 3:43am - Ottawa to probe 'disturbing' tweets by consultant on government-funded anti-racism project | CBC News
8.21.2022 3:43am - Only in Canada can a racist be hired as " senior consultant on an anti-racism project." Why? Because Canada has a "Federal Diversity Minister". In 2022, the words "diversity, equity, inclusion" should signal an alarm.

8.21.2022 2:58am - (Replying to @YaaraSaks) To those with decades of experience in campus antisemitism, a tweet that starts with: "I have been working with colleagues..." is automatically suspect of immaturity or complicity. Try to start with your own actionable goals, not with close doors "working with colleagues."

8.20.2022 10:15pm - To the best of my memory, this is the first time that CNN is addressing the Zionophobic virus on college campuses. Thanks, @DanaBashCNN for bringing it to public awareness. But, please, do not stop here, move on to University administrators; I can give you a few insider's leads.

8.20.2022 3:06pm - (1/3) Honored to be quoted with Olaf Scholtz @Bundeskanzler , Chancellor of Germany, on Abbas "50 Holocausts" mentality. The Chancellor overlooked, however, one tiny fact; Abbas did not "deny" the Holocaust. He in fact labeled it “the most heinous crime in modern human history,"
8.20.2022 3:06pm - (2/3) then used the credibility of the German Govt invitation to propagate his Zionophobic narrative, that Israel is 50 times more criminal then the Nazis, hence must be dealt with the same way. Chancellor Scholz, there is only one way you can honor the victims of the Holocaust:
8.20.2022 3:06pm - (3/3) Tell President Abbas the truth. Tell him that it's time for his people to internalize the permanent and rightful existence of Israel in their neighborhood, and that this existence cannot be overturned by slender.

8.20.2022 5:56am - Thanks for having me on your show, Gad. It is always a challenge, and great fun, to address new kind of audience, from a different background, and different angle on the playground we call "science."

8.20.2022 3:53am - (Replying to @Dam_Nuwen and @Sjnjerak) Shocked? Why? It's just commonsense. For next step, I'd recommend Primer Stats textbooks? For haven sake, not in 2022!

8.20.2022 2:57am - (Replying to @DChristensen73 and @ProfMattFox) Most SEM researchers are still stuck in the linear universe of SEM's founding fathers, so they use product terms as nodes in the graph. Non- parametric analysis assumes all possible interactions are possible a-priori, and goes ahead to quantify their magnitudes, IF NEEDED.

8.20.2022 2:48am - We are a week away from celebrating 125 years to the First Zionist Congress, Aug. 29, Basel, 1897. Time flies, but @GilTroy 's story will make your time freeze: and my humble take will add an unexpected color to that pivotal event:

8.20.2022 2:15am - (Replying to @KhaledAbuToameh) The supreme arbiters of right and wrong have developed a quantitative measure for crime severity, and they can't understand why some people chuckle.

8.20.2022 12:36am - (Replying to @b_surial @doaalsaleh and @ProfMattFox) Extremely unlikely is not a reason to leave an arrow out; in many situations the whole task is aimed at identifying those unlikely factors that spoil the null assumption.

8.20.2022 12:32am - (Replying to @JonMinton and @ProfMattFox) Every DAG imply the existence of some function Y=f(pa(y)) behind every parents-child family. If all are binary, f() must be a Boolean function, namely, a logic gate.

8.20.2022 12:28am - (Replying to @MD3_Phd @WillBall12 and 2 others) Its a solid paper, though a bit ambiguous on the meaning of "within levels of", i.e., do we hold the level constant? Or condition on it.? More generally, why label a variable "effect modifier" instead of estimating the quantity of interest, say, P(Y|do(X), do(Z)) ?

8.20.2022 12:16am - (Replying to @MKellyIrving and @ProfMattFox) Interaction? Yes, with a hidden, unnamed variable. "How best to feature that in a single DAG?" Retain the arrow Z-->Y. In DAGs, interactions are assumed to hold by default.

8.20.2022 12:08am - (Replying to @DChristensen73 and @ProfMattFox) It is "interaction", but between A and a hidden variable that may not have a name, but characterizes some unit, say those having an unknown type of allergy.

8.20.2022 12:05am - (Replying to @sweichwald and @ProfMattFox) No need to invoke/exclude higher moments. See example on page 35-36 of Causality. Causation exists even when Prob(Y=y|do(A=a)) = Prob(Y=y|do(A=a')) for all y, a and a'.

8.19.2022 11:55pm - (Replying to @PWGTennant @AndersHuitfeldt and @ProfMattFox) Another such example, involving three binary variables, is discussed on page 35-36 of Causality. It goes even farther: 'causation does not necessarily imply average causal effect (ATE)'

8.19.2022 11:45pm - (Replying to @EpiEllie and @ProfMattFox) "faithfulness" is assumed in causal discovery (ie, constructing the DAG from data alone), not in causal inference where a DAG is constructed from expert knowledge. The knowledge needed in the construction is individual-level, which some call "sharp null hypothesis".

8.19.2022 11:37pm - (Replying to @soboleffspaces and @ProfMattFox) Glad Elwert got it right, saying: "to anyone"

8.19.2022 11:22pm - Just checked @bellahadid Twitter. Wow, she hasn't block me! Though I explained to her clearly what it would take for her to be invited again to Friday dinners - chant the holy oath: "Two States for Two Peoples, Equally Legitimate and Equally Indigenous."

8.19.2022 10:42pm - (Replying to @mishtal and @Edinburgh_CC) The way to fight it is to propose an amendment: Twin with two cities, one being Shderot. This reasonable amendment might draw some support from the decent camp of the City Council, and adamant resistance from the Zionophobes, shelving the motion, and exposing its shady motives.

8.19.2022 5:02pm - We are in agreement that "Terms such as benefit, harm, kill, cure are not defined properly" in medical practice. I attribute it to the language in power, i.e., statistics, where those terms are indeed provably undefinable. You attribute it to "separate outcomes". I need examples.

8.19.2022 4:17pm - (1/2) The answer is YES. If we can think of any individual whose idiosyncratic characteristics would make her Y sensitive to A, an arrow should be drawn from A to Y. Reason: the DAG is an abstraction SCM, which is a collection of functions capturing individual sensitivities. Also,
8.19.2022 4:17pm - (2/2) for a practical explanation of "why": Assume we have a name to that characteristic (say allergy) which we would like to recognize as a variable C in the DAG. Wouldn't we like the DAG to tell us that Y depends on A given C? It won't if we deprive it of the A--->Y arrow.

8.19.2022 3:59pm - In three hours we will be sitting down for Friday night Kiddush, and mention poor @bellahadid loneliness, and her heartbreaking confusion why a racist campaign against the legitimacy of another people should exclude her from a table that celebrates the life of that very people.

8.19.2022 2:47am - The Western Press was so happy to dismiss President Abbas "Holocaust speech" as a slip of the tongue. And here comes this massive parade and tells them: "No tongue and no slip, these are out deepest unshaken beliefs, our mother milk and our water supply. This is who we ARE!"

8.18.2022 11:08pm - Students booted from sexual assault support group for being Jewish | World Israel News

8.18.2022 10:12pm - An important and captivating paper on a Turing Test for "agency". Still reading, but it occurred to me that a more credible counterexample to agency would be a thermostat, not a rock. Would a thermostat meet your definition of a "non-agent"? Still reading.

8.18.2022 10:01pm - It's my signature, I confess. And "Chazzak V'Nitchazek" (get strong and we all will) used to be my code for Jewish solidarity. I didn't think anyone would take me seriously and make Aliya; you are much stronger now, and so are we. Thanks for making me feel an accomplice.

8.18.2022 6:42pm - (Replying to @kareem_carr) My best tip: check if the book (video, course, totorial) starts with the Ladder of Causation. If not, I would hesitate, because you would be missing the Rosetta Stone of data-science.

8.18.2022 3:13pm - An unbelievable story! It's the most inspiriting solace I've received since Danny's death. Thanks. As Malcolm Hoenlein once told me: "He chose life." Now I see what he meant. Danny is still living up there, still inspiring strangers to take a look at where the planet is heading.

8.18.2022 4:55am - I tried, but your system says I am in already - can't vote twice. Too bad. In Poland, before WWII, professors could vote twice (so my mother told me).

8.18.2022 4:31am - (1/2) (Replying to @HL327) A verbal example of how PNS leads to different decisions from "the usual" is give in p. 3 of We see an RTC with zero treatment effect and PNS analysis that decides whether (1) the treatment has no effect on any individual or (2) it kills some and
8.18.2022 4:32am - (2/2) (Replying to @yudapearl and @HL327) saves others. If you are concerned about "Can this treatment harm my patient" (a counterfactual question) you would act differently under (1) vs. (2). Conclusion: You would act differently from "the usual", assuming "the usual" is acting according to what RCT tells us.

8.18.2022 2:24am - (1/4) The question that @HL327 is asking troubles many practitioners that I've talked to: How can mathematical analysis of hypotheticals (eg counterfactuals) assist us in making everyday decisions in the real, not hypothetical world? My answer: The questions you ask in real world
8.18.2022 2:24am - (2/4) decisions are in fact counterfactual in nature but, sadly, traditional education has not given us a language to express those questions honestly and directly, forcing us to ask them in roundabout ways, using statistical surrogates (invoking for example notions such as
8.18.2022 2:24am - (3/4) "risk factors", "odd ratios" "correlation" "attributable fractions", etc.) Today we have the language to capture those questions directly (eg, will it harm this patient), express them as counterfactual expressions (eg PN, PS, PNS) and submit them for analysis and estimation.
8.18.2022 2:24am - (4/4) In other words, counterfactuals help practitioners be more honest about their decision problems, articulate them freely & naturally as they come to mind, undistorted by appeasement of statistical needs, and submit them for mathematical analysis using the available tools.

8.18.2022 12:33am - (Replying to @akbarzamir) "why was it necessary to put a direct causal link between Sex and Final Weight?" Because we cannot rule out the possibility that the effect of W_I on W_F differs for boys, compared with girls. Oh, Lord, Kumbaya!

8.17.2022 7:34pm - Cover Story: AIPAC Doubles Down on its Mission via @jewishjournal
8.17.2022 7:34pm - From the day Andy Levin (D-Mi) lost his bid for re-election, Israel detractors have not stopped complaining about a shady organization named AIPAC using "dark money" to subvert US elections. This cover story by David Suissa has given me a picture of what the outcry is all about.

8.17.2022 12:56pm - (Replying to @elderofziyon) I don't think you get Abbas arithmetic. If the Holocaust is "the most heinous crime in modern human history" and Israel has committed 50 Holocausts, go figure out how evil Israel is. "It's elementary" says President Abbas. "It's elementary" say mama, papa, teachers, clerics..

8.17.2022 12:23pm - Israel, Turkey return to full normalization
8.17.2022 12:23pm - Next is Saudi Arabia, not for the sake of normalization, God forbid, but for the sake of "helping our Palestinian brothers”; who could object? How about Pakistan?

8.17.2022 11:45am - (Replying to @elderofziyon) Amazing! From his photo, one would swear he did finish high school. He is making a stupid mistake writing in Arabic; he could find many more English speaking readers, especially in the press, willing to swallow his stuff.

8.17.2022 3:19am - I'll weep for you tonight, Malki Roth. And when I weep for you, I weep for my son Danny too - your brother in pain. Malki and Danny, two treasures of humanity, crushed in Barbarian claws, to the cheering silence of those who can't spell "Barbarians".

8.17.2022 2:02am - I love the slogan "All regressions are 'true'". How true!

8.17.2022 1:40am - Same goes for this man's advice (2) "Listen to voices outside the foreign policy mainstream." In our region, this amounts to listening to what Israelis are saying: Mutual acceptance first, diplomatic horizon second.

8.17.2022 1:17am - (Replying to @JFronczekMD and @smueller) The example on page 3 of demonstrates that such reversal is not unlikely. Note that the RCT study not discarded, but *combined* with the observational study; the two together imply reversal.

8.17.2022 1:06am - (Replying to @davidboxenhorn @soccerdhg and @EinatWilf) The most "defining moment" of all is talking to any Palestinian intellectual today, and observing the certainty with which they hold Israel's demise.

8.17.2022 12:54am - (Replying to @davidboxenhorn @soccerdhg and @EinatWilf) For me, and for many Israelis of my generation, the defining moment was the famous "3 No's" of the Arab League Summit in Khartoum (Sep. 1967) in response to Israel's "Land for Peace" overtures.

8.17.2022 12:37am - "Dangerous consequences" will not deter PA President Mahmoud Abbas from future Holocaust distortions. What will? Reminding him of the causal connection between Palestinian's actions of 1939 and the Holocaust consequences of 1942.

8.17.2022 12:25am - (Replying to @PHuenermund) I always ask people "where are you from?". In that way I convey to them how proud I am, and how proud they should be as carriers of cultures that add color and luster to our lives.

8.16.2022 11:48pm - The word "Holocaust" spoken by Mahmoud Abbas always reminds us of Palestinian's role (1936-1939 riots) in preventing 6 million European Jews from fleeing the Nazis, blocking their escape routes (White Paper, 1939), and sealing their fate in the death camps.

8.16.2022 6:58pm - (Replying to @HL327 and @soboleffspaces) As I read your introduction to diagnosis reasoning, I can't help by asking what the research question is. IOW, are we searching for a factor X that maximizes the causal effect of X on Y? Or, perhaps, one that maximizes the conditional probability P(x|y), or PN, or PS of X on Y?

8.16.2022 1:39pm - (Replying to @soboleffspaces) If #causal reasoning cannot capture clinical reasoning, what can? Is clinical reasoning beyond capturing?

8.16.2022 1:34pm - (Replying to @duffy_maths) For "causal inference in economics", I recommend: And, if you can take the time to solve the 12 toy- problems presented (and solved) there, you'll be ahead of 85% of all living economists. @causalinf @econtweeter , @VC31415 , @instrumenthull

8.16.2022 5:21am - (Replying to @soboleffspaces and @prem_k) The first sentence raises the question: What practice is not "evidence based". Do you find it satisfactory?

8.16.2022 5:14am - (Replying to @JFronczekMD and @smueller) This thread is longer than a Govt manual. Can you summarize the doubts and concerns that you find compelling?

8.16.2022 5:04am - (Replying to @stephensenn) Unfortunately, no progress yet. Travel, conferences, deadlines etc. hijacked my thoughts. It's simmering there, and will I hope yield some resolution.

8.15.2022 6:17pm - If any of us was kept sleepless, wondering why days go bye with no word from CAIR (Council on American-Islamic Relations) on the Rushdie attack, @TheIPT has gotten to the bottom of it: Put simply: CAIR is on the side of the attacker, spiritually, of course

8.15.2022 5:48pm - (Replying to @artistexyz @esMktDigital and 2 others) Can you describe the "synthetic control" setup in the language of Bayesian Networks? Such description would be very helpful for communicating with economists.

8.15.2022 4:03pm - (1/2) I am more impressed by the paper's resolution of two conceptual puzzles: (1) Why should population data provide any information whatsoever on an individual response and, (2) Why should non-experimental data add any information (regarding individual response) to what we can
8.15.2022 4:03pm - (2/2) learn from an RCT alone? Resolving conceptual puzzles is one of the most satisfying advantages that causal modeling has over black-box methodologies.

8.15.2022 3:27pm - (Replying to @DKedmey) It may choose to speak Swahili, but the logic will be SCM logic, I am convinced.

8.15.2022 2:21pm - It is Monday again, our day of looking at the causal inference papers that this week has produced: Please let us know if you spot a breakthrough among the trees.

8.15.2022 1:55pm - This paper provides a conceptual introduction to Personalized Decision Making. It is one of the few to distinguish, both formally and conceptually, maximizing an individual's chances to END UP WELL from maximizing his/her chances to GET BETTER.

8.15.2022 1:16pm - It was great meeting Bernhard Schoellkopf Saturday, and discussing causal machine learning in Thomas Mann's personal office Needless to say, singing Bach's corals at Mann's own piano, was an inevitable part of the conversation:

8.15.2022 12:31pm - I've just received @EinatWilf 's new book - a MUST for those of us who are still pondering: Is there a path to peace in the ME? Note, she could have broaden the titled to read "The Path to Peace"; she didn't, boldly asserting: If you fear the path, you'll never walk it. Kudos!

8.15.2022 2:41am - @MissDiagnosis is one of those Pakistani friends to whom I must apologize for being unable to celebrate Pakistani 75th birthday. @AsraNomani and Prof. @akbarahmed479 are likewise being asked to forgive me, together with the many journalists I got to know.

8.15.2022 2:25am - (1/2) I wish I could join Jinna's "story of great human ideals" and celebrate Pakistan's 75th birthday with the many Pakistani friends that I have made through the years. But I can't. Pakistan's freeing of my son's murders, and her persistent anti-Israel stance in the
8.15.2022 2:25am - (2/2) international arena compel to pause and ask forgiveness of those friends -- I simply can't.

8.15.2022 1:40am - Strengthening my claim: There is no understanding/teaching of machine learning, or data science, without an understanding of the Causal Hierarchy. The hierarchy is perhaps more important in machine learning than in causality, where it is built into the inference engine.

8.15.2022 12:42am - Palestinian bus-shooter Amir al-Sidawi was incited by Palestinian Islamic Jihad (PIJ), the rockets of which were not condemned by mainstream media (eg, Mehdi Hasan of msnbc), thus conferring legitimacy onto PIJ's ambitions and practices.

8.15.2022 12:09am - Congratulations! And I thought good math professors, like the one I had in high school, no longer teach math in Israel. So proud to be proven wrong! I am eager to see the questions.

8.14.2022 10:45pm - (Replying to @Valerie_Harper1) The credibility of Amnesty's "fact finding" in Ukraine, parallels the credibility of their "fact finding" about Israel. Both are products of same institutionalized culture in which facts are enslaved to ideology.

8.14.2022 10:14pm - That was in the old days. No more. Come to college campuses and you'll get unambiguous descriptions of what these terms mean. UC Irvine pioneered this disambiguation with the a week-long celebration of "A World w/o Israel", followed with "Allah is a Terrorist"; more to come.

8.14.2022 9:55pm - False Accusations and Ideological Bias
8.14.2022 9:55pm - I have friends who've, for years, been supportive of Amnesty International - I'm often wondering if they still do. I've other friends who are telling me: It's impossible that such a noble-aimed NGO could mislead its donors-supporters for so long, so many of them. IT IS POSSIBLLE!

8.14.2022 5:45pm - Great thread! To be expanded perhaps with "probabilities of counterfactuals"; e.g., My uncle, a heavy smoker, died at 87. What's the prob he would have reached 90, had he not smoked. See

8.14.2022 1:29pm - Highly recommended! #causaltwitter, #causalinf, #EconTwitter, #Statistics,#instrumentals , #Bookofwhy, #causality, #MLB, #DeepLearning

8.14.2022 12:23pm - There is no understanding/teaching of causality without an understanding of the Causal Hierarchy. So much so, that I've started calling it Knowledge Hierarchy. @eliasbareinboim etal chapter formalizes and exemplifies the hierarchy to make it understandable to all researchers.

8.14.2022 11:17am - (Replying to @PavlosMsaouel @gary_lyman and 8 others) Looking forward to read the complete paper.

8.14.2022 10:43am - Antizionism/Zionophobia is a despicable form of racism even it were a totally new phenomenon, unconnected to ‘traditional’ antisemitism.

8.13.2022 3:41pm - (Replying to @soboleffspaces @stephensenn and 5 others) This is one of the greatest victories of scientific thinking. I am wondering what AI architecture could generate such an argument, and from what kind of experience/input. This exercise should be taken seriously by those who argue about symbols vs neurons.

8.13.2022 3:23pm - Senior UN official loses her post after tweet condemning PIJ rocket fire at Israel via @timesofisrael
8.13.2022 3:23pm - Warning! To work for the UN you must deny truth and sing foul. I am surprised at Mrs. Muscroft though, did she think her co-workers will not betray her for posting truth? Didn't she smell the stench of deceit before posting?

8.13.2022 12:51pm - (Replying to @soboleffspaces @eliasbareinboim and 4 others) I don't get it. Who are the "algorithmicists believing in data"? What are 3B-parameter models? Lost in guesses.

8.13.2022 12:45pm - (Replying to @eliasbareinboim @soboleffspaces and 4 others) I agree. The paper cited is strictly level-1. The authors state so explicitly: ANN is used for estimation and BN for "interpretation".

8.12.2022 5:17pm - Let's not overlook the fact that "distribution drift" has been solved completely in the microcosm of causal inference which, under normal circumstances, would trigger the curiosity of Titans from the AI macrocosm to ask: How? Can it be generalized? Can DL do it too?

8.12.2022 4:57pm - (Replying to @matloff and @MichaelALewis10) I've given this trend more that a serious thought. Can you imagine me using the terms "revolution" and "New Science" without giving it a serious thought? Additionally, CI is not a "Trendy New Thing", it is a New Thing turned trendy after 3 decades of hard struggle with stats.

8.12.2022 3:12pm - (Replying to @matloff and @MichaelALewis10) Please, do not "leave it at that". Please tell us what exactly is "unfairly dismissive of stat". We need to understand why statisticians feel offended when offered tools to express what they wanted to express but couldn't. See my interview with D. Hand

8.12.2022 12:56pm - (Replying to @MichaelALewis10 and @matloff) I would add to it: (1) causal assumptions is what's needed for CI, nothing else would do. (2) causal assumptions are defensible, because we think in causal, not statistical, assumptions, and definitely not "ignorability" assumptions.

8.12.2022 11:07am - (Replying to @MarielzaTalks @GaryMarcus and @ylecun) Which raises two interesting questions: (1) Why are people interested in debates more than, say, new polynomial time algorithms? (2) Looking back at your personal research, how much have you learned from debates compared to solving intellectually challenging problems?

8.12.2022 6:40am - (Replying to @johnmark_taylor) Bottom line: Identification = Symbolic, Estimation = Deep Learning. More details: Also:

8.12.2022 6:22am - Thanks for correcting the link.

8.12.2022 6:17am - (Replying to @DialSquareNZ @idatheleader and @mishtal) The number gap has always been against us, yet the world came to see our side in Nov. 1947, when we proudly touted our cause and exposed our enemies intentions. Today, we are moaning: Oy Vey! Antisemism! (which no one takes seriously) Instead of shouting: Zionophobes! Get out!

8.12.2022 5:57am - An interesting debate between @GaryMarcus and @ylecun on the tension between neural vs. symbolic architectures in AI: ttps:// For us, in causal inference, the symbiotic harmony and division of labor between the two is self-evident.

8.11.2022 9:52pm - (Replying to @artistexyz @soboleffspaces and 2 others) When the Titans are fighting up there, on Mt. Olympus, it's a perfect time for us, mortal sailors in causal inference, to sail the revolution ship forward, many miles forward.

8.11.2022 8:47pm - (Replying to @kareem_carr) And thank you, for giving us a proof that racists can still be handled as in the old days.

8.11.2022 8:30pm - Great! I think every respectable university should hire a couple like this, to help professors advertise new classes. Gee, just think what they could do with Stanford's "Foundation Models".

8.11.2022 6:53pm - (Replying to @robwilliamsiii) I would second your recommendation of #Bookofwhy which, btw, is not very charitable to Reichenbach's principle of common cause -- he forgot about colliders. (He is forgiven though, since, like most philosophers, he didn't have graphs to check ideas) @PessoaBrain

8.11.2022 2:39pm - Very interesting paper, thanks for shouting out to me. I am wondering whether all these features: "causality, role, knowledge, objective foreseeability, capability, intent, desire, autonomy, and character" are independently needed for computing "responsibility." Poor Jury.

8.11.2022 2:22pm - (Replying to @EliKohn3 and @isaacdecastrog) Gee, and my Bible goes all the way to Abraham: "To the land that I'll show you.." I bet it was printed by AIPAC.

8.11.2022 2:10pm - (Replying to @mishtal) When there is no speaker calling "Zionophobes out of London" passerby-s think these thugs are human rights advocates, and then some even join them. Words count, and we fail to use them.

8.11.2022 11:57am - (Replying to @dataengines) Sorry, I did not understand what your argument was from the "thermometer" metaphor. I scanned your paper and its not clear to me whether your PNS (prob of necessity and sufficiency) is the same as in and how you compute its bounds from data.

8.11.2022 3:47am - (Replying to @yudapearl @holdmytowel and 2 others) This trend towards "Inflated Expectations" was also noticed by @YiMaTweets who called it "the explosion of new names". But I'm not sure the choreographers of "Foundation Models" really have inflated expectations; surely the funders have.

8.11.2022 3:23am - I meant to say: Thank you @SadhguruJV .

8.11.2022 3:20am - (Replying to @elderofziyon) I don't believe @CAMERAorgs captured the real difference between @MehdiHasanShow and @AvivaKlompas . He calls her "racist" 3-4 times, and she treats him as a rational unbiased reporter, not once labeling him "Zionophobic racist"; you see why Mehdi won't debate me?

8.11.2022 2:59am - (Replying to @TsviArii and @Claire_V0ltaire) What do you think has gotten us out of Egypt, being Bundists?

8.11.2022 2:54am - (Replying to @holdmytowel @stu_frost and @jhurwitz) It's the peak of "Inflated Expectations".

8.11.2022 2:50am - Thank you @JFREJNYC for reminding us of things we tend to forget about Israel. And to imagine that all this was created under 75 years of relentless assaults by Zionophobic neighbors and other haters, @vinojininair .

8.11.2022 2:27am - (Replying to @yudapearl @skeenscene and 4 others) Those who understand what BDS stands for, especially its Zionophobic aims, are Ex-Jews, the rest are "useful idiots" - exonerated by reason of thoughtlessness.

8.11.2022 2:22am - (Replying to @skeenscene @juliajassey and 2 others) I looked at @JFREJNYC 's Pinned Tweet and I am not sure the slogan "all Jews are Jewish" makes any sense. Zionophobic Jews are certainly Ex-Jews because they renounce the central component of Jewish peoplehood. What about BDS-supporting "Jews"? Good question. Those @AdamSerwer

8.11.2022 1:43am - (Replying to @VivianBercovici and @Claire_V0ltaireZ) The titlte "Jew Hater" is bad because no "Jew Hater" would admit to being one. Compare with the title Zionophobe, of which they are all proud.

8.11.2022 1:39am - (Replying to @DrewBHeadley) Gelman's book could be a starting point for writing the badly needed "Regression Analysis: a causal perspective" . My opinion of Gelman's style of causal inference is summarized here:

8.11.2022 1:09am - (Replying to @stu_frost and @jhurwitz) It's wonderful to see "causal ai" being noticed by technology futurists but, for some reason, I do not feel comfortable seeing "causal ai" placed on the same curve with "foundation models."

8.10.2022 6:08pm - Speaking of Zionism, Judaism, Jewishness and Israel -- here is the most profound lecture on the topic, by Yossi Klein Halevy: Main motto: "We didn't survive the Holocaust, we defeated it!" @JGreenblattADL @michaeldickson @Ostrov_A @EinatWilf @blakeflayton

8.10.2022 5:53pm - Now, suppose they are right. Suppose "Zionism isn't Judaism". Does this make Zionism less noble? Less inspirational? Does this make Zionist Jews less deserving of rights to dignity, normalcy and self determination? Less deserving of admiration for what they have accomplished?

8.10.2022 5:38pm - It's time to examine this week's harvest of papers in Causal Inference: I se increased interest in "fairness" and "responsibility".

8.10.2022 5:28pm - Time to this week's harvest of new papers in Causal Inference: I see increasing interest in "fairness" and "responsibility".

8.10.2022 4:53pm - (Replying to @juliajassey and @JGreenblattADL) This is all very true in theory; in practice, the anti-Zionist will internalize this truth only when our leaders start using the word Zionophobia, the only word capable of making the anti-Zionist look at the mirror and comprehend the moral deformity of what he/she sees: ME?

8.10.2022 4:42pm - (Replying to @Claire_V0ltaire) This is partly due to our timidity and unwillingness to use a fighting word, Zionophobia, to assert our readiness to defend the nobleness of our cause and to expose the moral deformity of anti-Zionism.

8.10.2022 4:34pm - (Replying to @lackboys3 @bennettjake1 and @IlhanMN) The way to test it is to donate "out of area" money to defeat @ilhaMN and see if it turns "shadowy & dark". I just did. And I can't wait for the elation I'm going to feel when she is defeated. Just thinking about the possibility of such feeling is already worth the money.

8.10.2022 3:23pm - Jokes aside. Statistics is in dire need of a textbook on "Statistical Regression Analysis: A Causal perspective". It should be easy to write, just take the most popular regression textbook written, and explicate the thought that the author has been trying to hide. A best seller!

8.10.2022 2:31pm - I have been waiting for a re-run of Moshe Vardi's Lecture: "How to be an Ethical Computer Scientist." It is now available on youtube: The painful lessons for us, computer scientists, begin about 20 min. into the video.

8.10.2022 12:08pm - (Replying to @causalinf) When was this workshop?

8.10.2022 11:37am - Beautiful quote by Leonardo! Which I would use to argue for SCM: "understand the model and you will have no need of the experiment, be it "controlled", "natural", "quasi" "targeted" or "well-defined".

8.10.2022 11:28am - (Replying to @shiwalimohan) Lots of wisdom in your advice

8.10.2022 11:23am - (Replying to @AnoopRKulkarni) Welcome to the causal revolution and, as Churchill once said: “I have nothing to offer but [ ... ] commonsense and victory" (revised to fit the circumstances.

8.10.2022 7:01am - (Replying to @YiMaTweets) I suspect you overlooked the causal revolution, where you will not find explosion of new names, but explosions of new ideas and new solutions of problems that were lingering for generations. Including, BTW, ML problems such as transfer learning, data fusion, and explainability.

8.10.2022 3:49am - (Replying to @causalinf) It was an economist (Arthur Cecil Pigou) who first noticed collider bias (see #Bookofwhy, p198) but Economists lost direction once they accepted PO philosophy.

8.10.2022 3:34am - (Replying to @axel_red_third @HumzahAlkindi and 4 others) It was me. I call Zionophobic Jews Ex-Jews or Yester-Jews, stressing the observation that Zionism is a key component of Jewish identity, perhaps THE key.

8.10.2022 3:28am - (Replying to @HumzahAlkindi) I forgot where we stand on the 3rd Edition. Cambridge U. Press was essentially closed during the pandemic. I'll try to check again. Thanks for your interest.

8.10.2022 12:49am - I'm not sure if I shared with you my March 8 interview with Roi Yozevitch, who believes that Bayes Rule in the key to scientific thinking. It is available now on youtube (In English). We cover probabilities, causation, Hume, Eratosthenes, and more.

8.9.2022 11:42pm - (Replying to @robwilliamsiii) Blaming is a dangerous occupation. I would rather blame it on commonsense, which #Bookofwhy tried to capture to some extent, and which so many researchers hoped would sprout on its own, given more data.

8.9.2022 7:50pm - I remain cold, silent, skeptical and somewhat surprised at any attempt to do causal analysis without (1) the First Law (2) the Ladder of causation (3) Back-door criterion etc. etc. It's like shopping in a strange city without a city map. You may find a few bargains, but ...

8.9.2022 7:27pm - This is a more comprehensive survey, which covers some of the latest advances in the field, like transportability, sensitivity analysis and fairness. An important addition to the Causal Machine Learning avalanche would be

8.9.2022 7:13pm - Short but good paper to wet the appetite of RS readers for the emerging field of Causal Machine Learning. Missing ingredient: The First Law, i.e., that all potential outcomes can be derived from SCM, without which the impression is created that SCM and PO are non-intersecting.

8.9.2022 6:11pm - (Replying to @FabItMart @EpiEllie and 6 others) I've glanced at the paper. It seems restricted to linear analysis, so the basic issues we have discussed on the definition of natural effects do not surface.

8.9.2022 5:37pm - (1/3) I'm not sure how much of it is conspiratorial, and how much is deep genuine beliefs. The writer, @marwanbishara , is a great source for learning what Palestinians truly think about Israel. He might even be willing to engage with us to examine what myths we should both shed.
8.9.2022 5:37pm - (2/3) Hey, Marwan, are you willing? I like you conclusion: Israel could still turn all that hatred into tolerance,... if only it has the courage to atone for its violent past...and start treating [Palestinians] with the respect and honor they deserve as equals,.. in their homeland.
8.9.2022 5:37pm - (3/3) It mirrors (in reflection) so perfectly what I hear from my Israeli friends, reflecting almost word by word what I have been writing ("equally legitimate and equally indigenous"), perhaps there is room here for a dialogue? What do you think, Mr. Bishara?

8.9.2022 12:31pm - (Replying to @kareem_carr) That's awful! Did you try complaining to @Twitter ? How have you been handling it thus far. Namely, what percentage of them did you bother to answer?

8.9.2022 7:57am - (Replying to @JBSTVORG) Thank you @JBSTVORG for giving me the opportunity to share with your audience a few thoughts that could not be conveyed on Tweeter. Keep on the good work and I hope to see you again.

8.9.2022 1:00am - If a nomination is in process, I would be honored to join in support. Plus, a secret prayer:

8.9.2022 12:23am - Last week, when Ayman al-Zawahiri was killed by US forces, I was interviewed on JBS about this event, as well as Jewish on identity and campus #Zionophobia. It is available now on Youtube: I hope you can watch and share with others. @michaeldickson

8.8.2022 11:17pm - As I am sharing this video, I am also sharing a secret prayer: "Oh God, please confirm my hope that somewhere in this sophisticated system there is a tiny component, perhaps a smart sensor or an activator, that leverages one of my algorithms in its operation." Just a quiet prayer

8.8.2022 10:42pm - (Replying to @EpiEllie @pietrobiroli and 5 others) Recall that when dealing with dynamic systems, an experimental "unit" U=u must specify both a subject identity i and the time t. If, we wish to apply the NDE definition for a subject i, with t unspecified, we need perhaps some assumptions.

8.8.2022 7:46pm - (1/3) As predicted, Israel-bashers jumped on Gaza Fighting as a bowl full of cherries. Here is what Linda Sarsour writes on Facebook: "The Israeli government is massacring Palestinian children in Gaza and playing victim." They don't get it through their skulls that deceit and
8.8.2022 7:46pm - (2/3) hyperboles erode any credibility they still have in reporting grievances for the Palestinian people. They reinforce the theory that the only grievance Palestinians truly care about is the existence of Israel. Indeed, Nerdeen Kiswani (CUNY commencement speaker)
8.8.2022 7:46pm - (3/3) writes on Instagram: The only way to get justice for these children will be to end Israel." As I predicted: a bowl full of cherries.

8.8.2022 6:55pm - (Replying to @pythiccoder and @yanaiela) Any paper the leverages back-door should be recommended, especially coming from NLP modeling.

8.8.2022 6:47pm - "Ahmad! Turn the camera away! Rockets are being launched towards the sea!" Born in deceit, lives in deceit.

8.8.2022 6:21pm - (Replying to @EpiEllie @pietrobiroli and 5 others) We simply call the value T before the intervention T=0, and we need no assumptions. In my 2001 paper I was not even thinking about "intervention". I go to an arbitrary variable X that undergoes a transition from X=x* to X=x, for any reason whatsoever, and I ask about Y.

8.8.2022 6:10pm - (Replying to @EpiEllie @pietrobiroli and 5 others) Promise to read and to comment.

8.8.2022 5:58pm - (Replying to @EpiEllie @pietrobiroli and 5 others) See Equation (9) in Reading: NDE measures the expected increase in Y as the treatment changes from T = 0 to T = 1, while setting the mediator variable to whatever value it would have attained (for each individual) prior to the change, that is, under T= 0.

8.8.2022 5:44pm - (Replying to @EpiEllie @pietrobiroli and 5 others) I do not see any assumption needed for this innocent definition: "The expected change in Y while keeping all mediators at whatever values they had before the change (of the exposure)" I frankly do not see any threats coming from "time-varying causes" or "time-invariant causes".

8.8.2022 5:13PM - (Replying to @EpiEllie @pietrobiroli and 5 others) It's not a matter of "intention" but of equality. The pre-intervention value of Z is equal to the value that Z *would have had* under the other exposure, name under no intervention. Since "other exposure" is X=x*, as opposed to X=x, the prevailing value of X.

8.8.2022 3:43pm - (Replying to @EpiEllie @pietrobiroli and 5 others) As one who coined the term "Natural," please allow me to define it the way it was defined in 2001, Definition 4. Adding, of course, the phrase: "for that specific individual", since we are dealing with a "unit level" definition.

8.8.2022 3:27pm - (Replying to @yudapearl @EpiEllie and 6 others) Before you jump to another hasty conclusion, let me reiterate: Definitions have nothing to do with estimation. The definition I gave is valid for experimental as well as observational studies, with or without confounding, with or without identifying assumptions. And it's so easy.

8.8.2022 3:21pm - (Replying to @EpiEllie @pietrobiroli and 5 others) The Natural Direct Effect (NDE) of X on Y is the expected change of Y in the transition from X=0 to X=1, while keeping the mediator Z fixed, i.e., at whatever value Z had before the transition. It is easier than CDE, where you have to decide and justify at what value to fix Z.

8.8.2022 3:00pm - (Replying to @EpiEllie @pietrobiroli and 5 others) What makes you jump to the conclusion that I cannot present a clear explanation? The Natural Direct Effect is much easier to explain then the Controlled Direct Effect. And, as we both know, explanations have nothing to do with estimation. So what made you jump to such conclusion?

8.8.2022 2:15pm - A people's right to exist as a people expires upon denying that right to one's neighbor. Palestinians right is renewable anytime, but it would take a UN Civil Rights Lawyer to remind them that renewal entails explicitly rescinding all vows to eliminate Israel.

8.8.2022 1:38pm - Exciting news. Looking forward to snorkeling there again regardless of explanation. Still, would 2nd @DavidDeutschOxf in the quest for explanation.

8.8.2022 1:33pm - (1/2) One comment on @ShMMor 's accurate summary of the latest Gaza crisis. Israel's improved mastery of the information war (eg. immediately releasing the video of PIJ rockets that killed Palestinian children) still doesn't match the craftmanship of Western news editors in turning
8.8.2022 1:33pm - (2/2) Israel's innocence into guilt. By printing visual images of Palestinian children casualties, and none of Israeli's children, Israel's guilt is implied and imprinted in readers' mind, regardless of whether the rockets came from Shderot or from Gaza.

8.8.2022 12:35pm - (1/2) To substantiate my statement that the potential outcome framework is still firmly attached to the umbilical cord of "treatment assignment", I counted: The word "treatment" appears 500 times in Rosenbaum's book "observation and experiments", 220 times in Angrist & Pischke's
8.8.2022 12:35pm - "Mastering Metrics" and ??? times in Imbens & Rubin's book. What's the point? By rejecting the structural definition of Y(1) (e.g., the PO framework tied itself back to experimental setups. Take away the word "treatment" and you turn off PO semantic.

8.8.2022 11:09am - (Replying to @artistexyz @chiragjp and 4 others) I've never heard of "Vibration of Effects"

8.8.2022 11:07am - (Replying to @EpiEllie @pietrobiroli and 5 others) I can't think of even one example of a cause that *does't have* a natural direct effect, so, whose redflag is it.

8.8.2022 11:02am - (Replying to @EpiEllie @pietrobiroli and 5 others) What's wrong with the enzyme example on page 23 of ?? The purpose of science is to take "magic" and turn it into "actual examples".

8.8.2022 10:46am - (Replying to @EpiEllie @pietrobiroli and 5 others) Wait a minute. How a quantity is estimated in "actual experimental study" is what determines its having or not having "real world analog"? What's wrong then with the Mediation Formula for observational/experimental studies? See for "real world" examples.

8.8.2022 10:25am - (Replying to @instrumenthull @MeganTStevenson and 7 others) It took many decades for potential outcomes to be recognized as a valid tool for non-human manipulations, yet, despite the battle Rubin 76/Holland 86 were fighting, they're still attached to the umbilical cord of "treatment assignment", refusing the structural definition of Y(1).

8.8.2022 10:04am - (Replying to @EpiEllie @pietrobiroli and 5 others) I couldn't find reference to "Natural Direct Effect" in this paper so, naturally, I couldn't see why you find it "lacking of real world analogue." Is it related to the old crave for "well defined interventions"??

8.8.2022 9:50am - (Replying to @EpiEllie @PHuenermund and 5 others) Lucky me, and I was telling all my friends to DO MEDIATION ANALYSIS; didn't realize it's considered "lacking of real world analogue" by some real world people. Is it?

8.8.2022 9:27am - (Replying to @pietrobiroli @instrumenthull and 7 others) Surely we can talk about the effect of sex on "observer-1 perception of Sex" which, in turns, affects "observer-1 expectations" etc. Also, we can talk about the effect of sex on the secretion of certain hormones at a certain age. I don't see the problem.

8.8.2022 9:20am - (Replying to @HumzahAlkindi and @axel_red_third) For a definition of "Jewishness", see my 2019 commencement speech: For definition of Ex-Jew: "Born Jewish + Turned Zionophobe."

8.8.2022 8:37am - (Replying to @TriciaRoth18 and @MehdiHasanShow) He can't help it, given his upbringing in Al-Jazeera, but how can MSNBC have such a low opinion of its viewers' intelligence?

8.7.2022 11:21pm - Roger Waters knows his history right. Taiwan has always been part of China, Ukraine has always been part of Russia and Tel Aviv has always been part of Palestine. And if you do not know that, you are not reading enough. One thing I don't know, why CNN finds him so interesting.

8.7.2022 6:52pm - (Replying to @MehdiHasanShow and @MehdiHasan) What a creative solution!!! End the blockade! End the occupation! No wonder no one thought of it. It's so simple! Does @MehdiHasan truly believes what he is saying? Hasn't he read @EinatWilf about what the people of Gazza really want? I'm willing to go on his show for rehearsal.

8.7.2022 9:34pm - He doesn't really mean it. "He is only using this rhetoric to gain attention, and inspire his comrades with dreams and hopes" -- so say all Western analysts. @MehdiHasanShow even called (today) to lift Gaza blockade to allow him rockets and airplanes. He doesn't really mean it!

8.7.2022 6:24pm - (Replying to @EricRWeinstein) @EricRWeinstein , a friend advised me to tune in your Twitter thread. Glad we see equally clear the Israeli-Palestinian conflict and the inevitability of a 2-State solution. I see that many of your readers are trying to push you into confusion. Wanted to tell you: I am behind U.

8.7.2022 4:41pm - (Replying to @SAFECUNY and @CUNYKCCPRES) I don't believe she can be that callous; she is just unable to spell Zionism. Partly because no one has asked her before. Help her out; she must be thinking it's radioactive.

8.7.2022 4:19pm - Such a tectonic mind-shift, from "refugees" to "permanent residents" would take two generations to sink in. But, since it is the ONLY solution possible, Israel's Gov't should start bringing it to the table in EVERY diplomatic initiative, and every international forum. @EinatWilf

8.7.2022 4:05pm - (Replying to @soboleffspaces and @GuyattGH) Sorry, admitting ignorance, I've not been around EBM's HQ's lately. Perhaps today is a historical day for EBM, to shake hands with Causal Inference. Can you imagine what it would do for personalized medicine? #Bookofwhy #EBM

8.7.2022 2:28am - #Bookofwhy devotes a number of pages discussing species like owls, eagles or snakes lacking the ability to predict consequences of certain actions. Evidently, some humanoids are afflicted with a similar deficiency; they call themselves "journalists" and "Middle East Advisors."

8.7.2022 1:55pm - (Replying to @SAFECUNY @CUNYkcc and @CUNYKCCPRES) The litmus test on whether @CUNYKCCPRES really means to "address it" is whether she spells out what "it" means. If she says "antisemitism", forget it, she seeks a license for inaction. If she says "anti-Zionism", give her a chance, but make sure she spells it on CUNY's website.

8.7.2022 12:17pm - (Replying to @conjugateprior @a_strezh and 6 others) Another clear and thoughtful paper on race and discrimination, resisting manipulativity. The conclusions are too cautious, in my opinion. I've hard time thinking of any case where race is understood at all, be it formally or conceptually, and where it cannot be modeled causally.

8.7.2022 11:49am - Latest from Israeli TV: Cease Fire agreed on in 2 hours (11:30 pm Israeli time). Assessment: The best round of Gazza fighting ever; Islamic Jihad is crashed in both aims, men and infrastructure. The winners: the scientists and engineered who developed Iron Dome, 97% success rate.

8.7.2022 6:49am - (Replying to @dataengines) Cryptic to me.

8.7.2022 6:45am - (Replying to @WilliamWonkaJr and @RabbiLindaGold1) I'm sure too. But she represents the mentality of some Yester-Jews, especially from @IfNotNowOrg and @NewIsraelFund , that talking to her makes me feel like talking to them, or about them.

8.7.2022 3:30am - They say that when you get up there and meet Saint Peter, at heavens gates, he pulls out your voting record, raises his eyebrows and ask: "And which gate do you think you are going through?"

8.7.2022 2:56am - Follow the red ring around the Jabalia rocket that Al-Jazeera has identified as a "war crime".

8.6.2022 10:23pm - David and Goliath in front of our eyes. @Israel information bureau, with 800K followers, trying to correct a @CNN headline sent to its 59 Million followers. As David later chanted: "Truth sprouts from the ground" (Psalms).

8.6.2022 9:25pm - Agree. NYT reporters are unwittingly giving DAGs the publicity they so badly need. Endless reporting of data with no context, rockets instead of intentions, dead bodies instead of aims and goals, increases our appetite for DAGs, and the kind of knowledge they can represent.

8.6.2022 6:38pm - (1/2) Tonight, as we are entering the saddest day of Jewish calendar, 9th to the month of Av, we remember not only the destruction of the First Temple (586 BC, by the Babylonians) and the Second Temple (70 AD, by Titus, the Roman), but also the miracle of redemption --
8.6.2022 6:38pm - (2/2) the State of Israel -- tired of being hunted, but as if not concerned, will wait for the end of the chase. As Napoleon said when he saw Jews in French synagogues morning on this night: "A nation that morns it glorious past for 1800 years will surely regain it in the future"

8.6.2022 5:49pm - (Replying to @MeganTStevenson @instrumenthull and 7 others) Lily Hu's paper is unique in trying to bridge between those claiming: "this is NOT the 'true social meaning' of discrimination" and those trying to capture, be it formally or conceptually, what this "true social meaning" is all about. Highly recommended:

8.6.2022 4:34pm - The entire world understands that guns in the hands of the police are a barrier to peace because it enables a power imbalance in which store owners have zero incentive to do the hard work of negotiating with organized crime. It's also unfair.

8.6.2022 4:05pm - (Replying to @MeganTStevenson @instrumenthull and 7 others) The burden of changing "deep rooted instincts" is on those who know what discrimination truly is, to define it in different terms, perhaps avoiding variables altogether, but still enabling society to distinguish discriminating employers from fair ones.

8.6.2022 3:45pm - Speaking of "fairness", here is a "fair" request made by my Rabbi @RabbiLindaGold1 : The Mayor of every US city should make sure the quality of weapons in the hands of the mafia equals that given to the police. We must "create some parity", said the Rabbi, "It's truly a disgrace".

8.6.2022 11:12am - To me, the difference is much more than notation. The difference lies in whether counterfactuals must be defined in the context experimental "treatment assignment," or are they part of our commonsense knowledge about the world, eg SEM.

8.6.2022 10:59am - (Replying to @eliasbareinboim @geomblog and 7 others) (1) A toy example is needed to illustrate the new decomposition. (2) Would the new decomposition pacify social scientists demands to include "constitutive factors" (however defined), as in

8.6.2022 10:45am - (Replying to @instrumenthull @geomblog and 7 others) I don't have a favorite empirical analysis. I don't know what insight such analysis can provide before we agree on how discrimination is defined. The useful information from such studies lies in the data themselves, perhaps well summarized and organized, not in remedial actions.

8.6.2022 9:33am - (1/2) Latest from Tel Aviv. It took just 170 seconds for IDF to surprise Islamic Jihad (Arabic speaking Iranians) and eliminate the head of its military arm, Mahmoud Jabrie, together with 15 other saints and establish a new equation: intelligence-verified decision to attack Israel
8.6.2022 9:33am - will meet the same response as an executed attack. Key question: Will Hamas stay out of the fighting and, more importantly, will the new equation create long-term relief for Israeli civilians? Rocket sirens sounded in Tel Aviv and other Israeli cities - so far no casualties.

8.6.2022 9:08am - (Replying to @geomblog @instrumenthull and 7 others) Interesting paper, ending with an unresolved dilemma: Can "constitutive relationships," those defining what discrimination IS, be represented by, or derive from DAGs and/or counterfactuals? I believe they can, and the paper does not (yet) convince me that they can't.

8.6.2022 12:15am - (Replying to @MeganTStevenson @instrumenthull and 5 others) It is outdated to talk about "the Dangers of Counterfactual Causal Thinking" with Rubin's PO framework in mind. Legal counterfactuals are defined in SCM by the First Law where manipulativity is a red herring. See #Bookofwhy p.312 Carson v. Bethlehem Steel

8.5.2022 5:39pm - Just talked to my sister in Tel Aviv. She's going to sleep with cloths on -- the shelter is level 1, and she lives at level 4 of the building. I tried to lift her spirit, reminding her how worse things were in 1948, under Egyptian bombing. Price of independence. Shabbat Shalom.

8.5.2022 5:24pm - (Replying to @soboleffspaces) Who is Olga? Does she speak Russian? Does she sing BOW?

8.5.2022 2:15pm - (Replying to @TomPepinsky @instrumenthull and 4 others) Yes, because there is no back door path from race or gender to any other variable. However, recall that in discrimination cases we are usually interested in the direct effect (not the (total) causal effect) which is a counterfactual notion. See

8.5.2022 1:54pm - (Replying to @JohnMullahy @PHuenermund and @_MiguelHernan) Forcing people to think in a certain way may be helpful to some people. I would rather think my way, without being forced, since we now have a language to capture our natural way of thinking.

8.5.2022 1:46pm - (Replying to @instrumenthull @PHuenermund and 3 others) So, after "precisely imagining" and after thinking "natural experiments", do they get the same answer we get from graphs and counterfactuals?

8.5.2022 7:39am - Do organizations such as @IfNotNowOrg decay and perish in their own stench? Or they need an adult to tell them: "Look what you have become"?

8.5.2022 6:52am - (Replying to @prem_k) They talk money, not science, so I am unsure of what they are doing. One thing I am sure about: our generation would draw greater benefit from investing this money in an academic center for causal inference.

8.5.2022 6:41am - I've mentioned Simpson in passing, but the more I think about it, the more I see it as a profound litmus test of "understanding" in machine learning: Would a learning system ever reject a drug that is "good for men, good for women, bad for persons," when the data support it?

8.5.2022 6:06am - This paper should be read together with the Appendix to which demonstrates how causation among non-manipulable variables permeates ordinary scientific discourse.

8.5.2022 5:30am - (Replying to @PHuenermund) Conclusion: Stop working, your income won't change.

8.5.2022 5:27am - (Replying to @Claire_V0ltaire) I love the expression "divine religions". Can't think of a "non-divine religion," unless they have "Zionism" in mind.

8.5.2022 4:51am - (1/2) A new lucrative source of income for Israel advocates! An Irish court ordered Prof. Ronit Lentin, to apologize and compensate Yoseph Haddad, a pro-Israeli Arab activist, after she defamed him on social media: Just imagine if every Israel basher is
8.5.2022 4:51am - (2/2) is forced to compensate all those whom he/she has defamed or maligned, both personally and collectively. The UN treasury would surely go bankrupt.

8.5.2022 4:28am - Readers who asked for a tutorial article on the front-door criteria might find this one to be instructive: It's my first article on the topic, written 30 years ago, which has retained the freshness of the idea, and its relation to Instrumental Variables.

8.5.2022 4:17am - INN is a brain-child of the @NewIsraelFund , launched in 2014, assuring the Jewish community that members of IfNotNow have a deep sense of commitment and ethical clarity. Sadly, self- righteousness combined with juvenile mentality may spawn a different creature, as seen below.

8.4.2022 11:01pm - (Replying to @artistexyz @LeCun and @GaryMarcus I agree that the NN-Symbols debate may not be the issue of interest, and that data-fitting vs model-based dichotomy can tell us more about the theoretical limitations of NN systems. For example, can NN's reject Simpson's drugs: good for men, good for women, bad for persons.

8.4.2022 9:52pm - (Replying to @thinkyparts @ADL and @JGreenblattADL) "both sides"? No way!! In my school, a pupil who made a derogatory remark against Arabs was sent home to bring his parents. The frequency of the word "peace" in Hebrew folk-songs exceeds that of "I love you". Please don't tell me about "both sides". I'm tuned to their textbooks.

8.4.2022 6:40pm - Hilarious conversation! "American Jews think/act like Christians without Jesus."

8.4.2022 6:02pm - (Replying to @thinkyparts @ADL and @JGreenblattADL) Thanks for the warning. I usually us "mother-milk deep" but, given what I see and hear from our neighbors, I'm not sure which is deeper.

8.4.2022 5:57pm - (Replying to @Rondo2 @ADL and @JGreenblattADL) I wouldn't give up so easily on @ADL . We recently heard some sparks of understanding from @JGreenblattADL , and he might even clarify here what @ADL 's position is on "pure" Zionophobic bigotry, one that does not rattle any anti-Semitic tropes American Jews habitually fear.

8.4.2022 2:05pm - (Replying to @thinkyparts @ADL and @JGreenblattADL) 2-State solution has always been Zionists goal and, while settlements make this goal difficult, what makes it IMPOSSIBLE is Arab's rejectionism: "not even the size of a postage stamp". There's ample ambiguity in defining antisemitism, anti-Zionism is crisp: No Jewish sovereignty.

8.4.2022 1:21pm - I was happy to see the @ADL 's report pointing to the danger of anti-Zionist candidates. Alas, the only fault @ADL finds in those candidates is invoking "Jewish power and control" rhetoric. Does @ADL deem Anti-Zionism in itself to be a legitimate form of bigotry? @JGreenblattADL

8.4.2022 11:59am - (Replying to @Big_Worker @EliSennesh and 2 others) Evidently, some people take seriously what these 2 UN "experts" say. Some even quote them (Lynk and Rajagopal) on Tweeter, as if apologizing: "This is where we get our education on the ME conflict. Don't blame us for being so ignorant."

8.4.2022 10:00am - (Replying to @kareem_carr0 Why "think" when it is easy to test. The part that comes from the mental model is not available to a machine, so, lets give the same data to a machine and ask her questions about "what's going on". That's what we did mathematically and found: "No causes in, no understanding out".

8.4.2022 9:10am - (Replying to @waseemkiani and @MyNutritionSci) Let me assure you, and I am unbiased, in this case it's not "jut marketing." When you are done, I'll explain what makes "revolutionary" different from just "useful".

8.3.2022 11:02pm - I am willing to forgive @jstreetdotorg for their fight with AIPAC, if only they would change their insulting logo: "pro-Israel, pro-peace", as if Israel, if left alone, is not craving for peace; as if Israelis need J-street push and supreme wisdom to fulfil their lifelong dream.

8.3.2022 10:33pm - (Replying to @ESMDcan123 and @artistexyz) My point is that the study consists of two stops: The consenting subjects are first assigned randomly to treatment and control arms, and then subjected to forced intervention, depending on the assignment. This second step corresponds to the letter 'C' in RCT.

8.3.2022 10:13pm - @NaviPillay1 seems determined to bring the reputation of the UN Human Right Council to below zero.

8.3.2022 10:00pm - Who said Americans would give up their hard earned freedoms?

8.3.2022 6:43pm - I believe it came about to prevent people from assuming that you can get similar results by doing post-treatment randomization. IOW, to emphasize that an RCT entails two steps: randomization and forced intervention.

8.3.2022 6:34pm - Hold your head high, @Andy_Levin . You now have a choice: (1) Bask in @PeterBeinart admiration of your heroic stand, (2) Mull over a realistic assessment of your mistake: Palestinian freedom entails honest co-existence, a word detested by the squad members you have embraced.

8.3.2022 5:19pm - (Replying to @TheOtherSimo) The ideology that equates the survival of Israel with "hating Palestinians" is precisely the ideology that has prevented Palestinians from accepting co-existence with Israel (since 1948), which would have brought peace, independence and prosperity to both sides.

8.3.2022 3:28am - We are informed of an interesting discussion group on causality from Neurips viewpoint. It has Rung-3 of the ladder on their icon, so they must know where they are heading.

8.3.2022 3:05am - (Replying to @RabbiLindaGold1 and @JenChangeFL) I see great future for him, once he realizes that Jewish voters do care about Israel, and are no longer fooled by populist slogans J-Street style.

8.3.2022 2:49am - (Replying to @StaceyEBurke and @JenChangeFL0 Zionophobes can't tells lies from truth, their entire universe is built on the former.

8.3.2022 2:43am - (Replying to @JenChangeFL) It's 2022 and people are still touting: "There is a difference between anti-semitism & anti-Zionism", not realizing that voters deem the latter uglier and more dangerous than the former, and that the real issue in the case of Andy Levin was his embrace of the dangerous latter.

8.2.2022 9:05pm - (Replying to @jacobmbuckman @kareem_carr and 2 others) Here is the proof, @kareem that everyone, including deep learning folks are interested in "understanding", not just prediction. Do you still see such a big dichotomy between stats and ML? Isn't the latter of great value to the former?

8.2.2022 7:48pm - (Replying to @elderofziyon) I don't understand why good people try to minimize AIPAC's role in Levin's defeat. How else could people like me, a resident of California, express the danger they see in having another Squad sympathizer enter the Hall of US Congress, if not by openly supporting AIPAC?

8.2.2022 6:27pm - (Replying to @kareem_carr) I'm not sure what "stats" you have in mind for which it's fair to say: "In stats, understanding is almost always the goal." My definition of "understanding" makes stats a discipline for which "understanding is almost NEVER the goal". Where do we differ? On stats or understanding?

8.2.2022 4:54pm - The racist always walks around with a genetic test-tube, to test his neighbors linkage, especially racists lacking history, songs, language or holidays; they got nothing to base claims on, but soiled, polluted test-tubes.

8.2.2022 4:35pm - (Replying to @MilenaAlmagro) Worth asking: What Kept the Cowles Commission at Bay? See, p. 168

8.2.2022 4:17pm - I couldn't believe @RBReich actually said this, until I checked it myself, he used to think before speaking. @RBReich , Berkeley professorship does not absolve you from thinking. Do you really believe AIPAC's goal is to "take down progressives"? Not to "take down Zionophobes"?

8.2.2022 2:16pm - (Replying to @78tiger and @blakeflayton) The rumor I mentioned was, of course, poetic. But not too far off. To the best of my recollection, it was @IfNotNowOrg that said Kaddish for both "Israeli and Hamas victims" in the latest Gazza clash -- any distinction between good and evil makes them feel too "ethno-centric".

8.2.2022 11:07am - And if anyone objects to your definition of "indigeneity" ask him/her the questions that Lord Balfour asked professor Rashid Khalidi in this poetic encounter:

8.2.2022 10:48am - ‘I Solemnly Demand That Anti-Zionism Be Banned in France,’ Parliamentary Deputy Declares
8.2.2022 10:48am - The first Jewish leader to recognize that our current fight against antisemitism is misguided, futile, and only gives cover to our real enemies - the Zionophobes. @JGreenblattADL , @AJC , @almog_doron , @noatishby , @EinatWilf , @newzionists , @blakeflayton @SAFECUNY

8.2.2022 10:20am - Our neighbors always side with the oppressed freedom-fighters: Putin, China, Ayatollah, Isis, Taliban,... As the Mishnah says: Your most important choice = Good neighbor

8.2.2022 10:06am - Significant step in re-thinking religion's role in society.

8.2.2022 9:00am - Interesting details on Ayman al-Zawahiri's life, before last week's recall to answer his creator.

8.2.2022 2:09am - Sharing our weekly harvest of causal inference papers: Many of the titles look enticing, but I didn't have a chance to look at them yet.

8.1.2022 11:30pm - (1/2) Background: Here is a sweet letter USC @PresidentFolt sent to 45 Zionist professors who complained about the treatment of Rose Ritch Guess what word is missing! "Zionism" - the core of the complaint. Why? Because dealing with
8.1.2022 11:30pm - (2/2) "antisemitism" is so much safer; all the Jewish organizations praise you for inviting them to the "Advisory Committee on Jewish Life," and give you 5 years to philosophize and pretend you can't spell "Zionism," the toxic word you refuse to embrace in Jewish identity. Sad!

8.1.2022 11:00pm - Jewish leadership still hasn't learnt the futility of whining "antisemitism" and the power of indicting "antiZionism". See how they melt away at USC, and give @PresidentFolt the license she needs for inaction. @bandlersbanter , @ADL , @AJC , @noahfarkas

8.1.2022 9:36pm - (Replying to @blakeflayton) Rumor has it that @InNotNowOrg folks are saying Mourner's Kaddish tonight for Ayman al-Zawahiri, the greatest de-colonizer of all times, and the prophetic voice of "true Judaism."

8.1.2022 8:50pm - (Replying to @kevinsalwen and @MaxBoot) Thanks for writing, Kevin.

8.1.2022 8:11pm - I have a personal account to close with al-Qaeda, more so with al-Zawahiri the "ideologist", than with bin-Laden or Khalid Sheik Mohammed, the executioners of our son, Daniel. Today I feel an inch closer to closure. Angels in heaven tell me they feel that way too.

8.1.2022 7:39pm - (Replying to @HumzahAlkindi) Thanks for sharing the article in Haaretz, but I do not see any contradiction between Popper's view and mine. My position that a nation state can be both successful and open to criticism is attested in the form of Israeli society, which neo-Marxists feel obliged to delegitimize.

8.1.2022 7:09pm - This is not an endorsement (hate to lose my tax status) only my moral duty to let readers know the names of some election candidates who are not Zionophobic - a gift of fresh air in squad-ruled politics.

8.1.2022 6:49pm - (Replying to @HumzahAlkindi) Oh, I've confessed my views here when a reader asked: where does this moral deformity comes from?

8.1.2022 6:33pm - (Replying to @pdddavid1) Thanks.

8.1.2022 3:07pm - (1/2) Readers who inquired on foreign publications of #Bookofwhy, here is what our publisher tells us: 1. US Basic Books, 2. UK Penguin Books , 3. SPANISH Pasado y Presente, 4. HOLLAND Maven, 5. FINLAND Terra Cognita, 6. PORTUGAL Temas y Debates, 7. JAPAN Bungei Shunju,
8.1.2022 3:07pm - (2/2) 8. KOREA ScienceBooks, 9. TAIWAN The Walk Publishing, 10 CHINA CITIC, 11. ISRAEL Kinneret-Zmora Dvir Publishing, 12. RUSSIA AST, 13. POLAND Copernicus Center Press, 14. TURKEY Ginko Publishers You're right, no French publisher found thus far; amazingly behind!

8.1.2022 9:47am - (Replying to @dmkalman and @skdh) I'm generally unsatisfied with "trial and error" until we define what "error" is and how to choose the next "trial". Once we do that, bingo, we can systematize the process, hopefully even automate it.

8.1.2022 9:43am - (Replying to @sean_welsh77 and @David_Gunkel) So you see any impediments to emulate these functions on a computer? I don't. Especially "caring" "morality" and "curiosity", where we currently have the beginning of a thread to their secrets.

8.1.2022 9:35am - (Replying to @StillTr05207382 and @YiMaTweets) You have provided is a fairly accurate description of what has happened in #ML in the past 15 years: Build machines to do tasks, without defining what the "task" is, what success is, and what theoretical limitations are there on the success. e.g.,

8.1.2022 1:32am - (Replying to @sean_welsh77 and @David_Gunkel) From practical AI viewpoint, faking it would be sufficient and extremely useful (and dangerous). From a philosophical viewpoint, just saying "its more than just xyz" whenever xyz is less than an actual human being is not very interesting, unless we hear what function is missing.

7.31.2022 9:22pm - (Replying to @ProudUSJew) Daniel's Yahrzeit is February 1st. If you have a chance, ask your minyan tomorrow in shul to say: Avinu SheBashamayim, Tsur Israel V'Goalo, Barech et Medinat Israel, Reshit Tsmichat Geulatenu... Daniel would appreciate using his name to inspire Jewish pride, resilience and hope.

7.31.2022 7:46pm - (Replying to @_brohrer_) How can you tell a cult of complexity that is a cover up for lack of substance from complexity that arises when science matures? My answer: The latter is accompanied with a toy problem that illustrates the idea with 3-4 variables. I bemoan the lack of toy problems in #ML.

7.31.2022 6:57pm - Correction. Biden said "50 years", not "25 years". Indeed, his first trip to Israel was in 1973

7.31.2022 6:48pm - I would stick to "mimicking" as long as no one explains to me what she means by "ACTUAL phenomenal consciousness". (Recall, faking it is having it, because it's so hard to fake w/o having) And I wouldn't knock "symbol manipulation according to rules" - this is how my brain works.

7.31.2022 6:37pm - (Replying to @bonnieshouses and @Claire_V0ltaire) Thanks, I also like this article, but Wikipedia won't allow me to reference it -- Sister Linda is a protected species. Please try to cite it and see what happens. Perhaps it's only me that she fears.

7.31.2022 6:24pm - (Replying to @ramespaulus and @IfNotNowOrg) I'm going to ask @Twitter for reinstatement, present my Jewish credentials, and pretend I am a Jew of Discomfort. I can even put on an Arafat Yarmulka, chant Havdalah at noon like an idiot. As long as they don't ask me to support Rep. Andy Levin -- I cannot betray my people.

7.31.2022 5:45pm - There are many Americans who are great friends and allies to Israel. But Biden was unique in telling the hecklers so precisely why their screams "Occupation! Occupation!" sounded so idiotic in light of his 25 years search for a single Palestinian leader willing to accept Israel.

7.31.2022 4:03pm - My! My! I am blocked by
----Quote ----
You’re blocked You can’t follow or see @IfNotNowOrg ’s Tweets. Learn more
As Joe Biden noticed, only Jews who cut their historical roots can join these juvenile hecklers.

7.31.2022 3:18pm - (1/2) This should be the main topic of discussion in the next international conference of Clinical Psychiatrists. Some say it's morphed antisemitism - inability to stomach Jews, whose historical place is down below, suddenly act as sovereign equals. But I believe it's more a matter

7.31.2022 3:18pm - (2/2) of social ideology. Israel's success makes a mockery of neo-Marxist textbooks, according to which a nation state must be evil, corrupt and unjust. If it isn't, we must make it look like one. I wish some university would establish a separate department to study this pathology.

7.31.2022 2:42pm - (Replying to @Abel_TorresM @eliasbareinboim and @kerstingAIML) Can you share an example of a "well-known" model = ?

7.31.2022 1:14pm - The #Bookofwhy claims that data are dumb, namely, that they tell us little about the world generating the data. That was before seeing the UNHRC data. Anyone who thinks it doesn't tell us about the moral deformity of those behind these condemnations has not seen moral deformity.

7.31.2022 12:56pm - Emory University Student: I Have Been Attacked and Defamed for Visiting Israel
7.31.2022 12:56pm - I'm retweeting this piece for the hundreds, perhaps thousands of readers facing similar situations: "I will not be stepping down, I will in fact be stepping up and speaking out for all those facing harassment and discrimination on campus because of who they are." @newzionists

7.31.2022 12:20pm - (Replying to @chroma_infinity) I'm not surprised about the Israel issue, the anatomy of which I've studied inside out. But what about econometric in industry? It's hard for me to imagine even ONE such task, where an econometrician cannot benefit immensely from the tools provided by the causal revolution.

7.31.2022 12:08pm - Retweeting: How I wish he (Sewall Wright) was alive today. Reading the #ML literature and saying: they haven't learnt a thing, have they? What an enchanting fantasy!

7.31.2022 11:56am - (Replying to @grbradsk and @YiMaTweets) That would be understandable and forgivable. Trouble starts when these script jockeys decide to take over academia to gain respectability. They then hire hordes of script jockeys to teach and review+reject conference papers and scientific proposals which they cannot understand.

7.31.2022 11:42am - (Replying to @EricJGreenberg1) Any one of the tutorials on the #Bookofwhy page would do. See 3. Scientific Backgrounds

7.31.2022 11:37am - (Replying to @roberto_limongi) I'm more inclined to believe they do not KNOW the philosophical or scientific assumptions behind #ML, else they would be thrilled to teach them.

7.31.2022 9:48am - (Replying to @NicholasMatiasz and @skdh) I havn't given it sufficient thought, since it is only last week that I saw the word "sentience" for the first time.

7.31.2022 6:11am - Unfortunately, narrow-mindedness caused by "too busy writing" also afflict reviewers, so it becomes self-enforcing, hermetically sealing the field from new ideas.

7.31.2022 5:54am - (Replying to @letonyo and @YiMaTweets) How about writing: "The 10 milestones in the history of #AI every busy #ML student should know". Same with "history of science".

7.30.2022 11:05pm - Speaking about Linda Sarsour, did you know that Wikipedia considers her "protected minority"? I tried to cite my article about her - they won't allow it -- she will for ever remain beyond criticism, like a religious icon.

7.30.2022 10:53pm - I've just watched @skdh video on "What's the difference between a brain and a computer" and, I must confess, I remain a staunch believer in AGI, seeing no theoretical impediment to mimicking a conscious human agent on a digital machine.

7.30.2022 10:14pm - Joe Biden is the only Zionist who knows how to handle the @IfNotNow crowd. When they heckled him during the 2020 campaign, he said: "Show me one Palestinian leader willing to accept Israel, and I will talk to you about the "occupation"" (paraphrased from a friendly memory).

7.30.2022 6:53pm - Whenever I see Jaffa in the horizon I am reminded that my 14 yr old father landed there in 1924, with his entire family, hoping to start a new life, back in the country they left 2000 years earlier.

7.30.2022 6:33pm - I've just registered to a lecture that I've always wanted to hear: How to be an ethical computer scientist Knowing Moshe as a top formal-logic researcher, I am sure his take on ethics will be different from the usual social media preaching. @vardi

7.30.2022 4:53pm - Education Dept. to Investigate USC’s Handling of Antisemitic Harassment Against Rose Ritch via @jewishjournal
7.30.2022 4:53pm - (1/4) In as much as I dislike litigations clouding academic affairs, I recognize that they serve a useful function. They call public attention to what's going on, they force administrators to rethink past blunders, and they entice Jewish organizations to tweet: "We're here!"
7.30.2022 4:53pm - (2/4) The next step, once litigation runs its course is for university administrators to act meaningfully in the normative sphere: (1) Do NOT appoint a showcase "task force" to deal with antisemitism. (2) Spell out the word "Zionism" openly and explicitly. (3) Tell the campus
7.30.2022 4:53pm - (3/4) why denying a people's homeland is morally deplorable, why Israeli and Zionist students and faculty are welcome on campus, and what unique cultural and nation-building experience they carry that make them valuable contributors to campus life and academic excellence.
7.30.2022 4:53pm - (4/4) Zionist students and faculty are not asking for legal protection; they ask for rightly earned equity which is within your duty to sustain, by setting norms of respectful academic discourse, while protecting everyone freedom of speech. @bandlersbanter @ADL @newzionists

7.30.2022 8:39am - (1/2) There is a similarity of music between Simpson's paradox (bad for men, bad for women, good for people) and the two-slit experiment (classical pattern from left slit, same from the right, but interference when we don't know which slit it is). However, the two-slit findings
7.30.2022 8:39am - are observed empirically, whereas Simpson's miracle drug does not exist; it only appears to reverse effect direction due to our misinterpretation of what "bad for" and "good for" mean. No reversal can be observed once we properly interpret what we mean by "good" and "bad".

7.30.2022 8:03am - A congresswoman who does not grasp how her persistent dehumanization of Israel weakens, if not negates the credibility of her arguments for alleged dehumanization of Palestinians.

7.29.2022 10:14pm - (Replying to @BabakWork) Analogies to quantum physics may have their place, but not when we are trying to capture human intuition, which is at odds with the predictions of quantum physics.

7.29.2022 4:52pm - (Replying to @artistexyz @GunnarBlohm and @TRyanGregory) A parsimonious representation, from which one can deduce a compendium of facts is still called "knowledge". This is the convention in AI, and this is the reason why we need to distinguish "explicit" from "implicit" knowledge.

7.29.2022 3:21pm - (1/2) Good question, the answer to which is elaborated in #bookofwhy (p. 200) and wikipedia, but can be seen clearly from the fact that human intuition, across all cultures and persuasions, rejects the possibility of a drug that is bad for women, bad for men, but good for people.
7.29.2022 3:21pm - Such impossibility is a theorem in causal-calculus, but not in probability calculus. In other words, you cannot prove this impossibility using the language of probability theory alone, without invoking causal assumptions and causal logic. See

7.29.2022 1:49pm - (Replying to @CresaKay @GunnarBlohm and @TRyanGregory) I've been meddling in this playground for quite some time, and it hasn't mislead me even once. I would dare call it "safe".

7.29.2022 8:20am - (Replying to @MichellePearl6 and @EpiEllie) Don't worry, IT was a male dog at birth.

7.29.2022 7:14am - (Replying to @OlaSAhmad_) The good people at @IJCAIconf promised to post it.

7.29.2022 7:12am - (Replying to @juan_pigurina) When was it that you felt uncomfortable speaking the language of causality? Simpson's paradox proves that our mind is a causal machine, not a probability machine. If it weren't for the invention of statistics (Pearson, 1890's), we would never stop speaking cause and effect.

7.29.2022 5:53am - If I were the @USAmbUN , I would simply address what "cannot go unaddressed" and call Miloon Kothari to resign disgracefully, both for what he said, and for not demanding that a UN member State that threatens the destruction of another member State be stripped of its membership.

7.29.2022 5:08am - Swayed by your recommendation, I took another look. Boy, either this book improves each time I read it, or I am becoming more addicted to commonsense!

7.28.2022 4:56pm - (Replying to @FredrikHeintz) Can't causal graphs represent temporally dependent variables, sequential planning, etc? Here is an old paper

7.28.2022 4:50pm - (Replying to @soboleffspaces) How about just commonsense? We do not need mania to be sensible.

7.28.2022 4:47pm - The reason I believe the next revolution will be "more impactful" is that it will rally all the smart data science folks behind it, all 105,980 of them, not just the heretic few who are currently working on causal inference. See

7.28.2022 1:13pm - Overwhelmed by the friendly reception of my keynote talk @IJCAIconf today, in Vienna, I am happy to share my slides for our readers: A video of the 2nd part of my talk, separately recorded, will be available on @IJCAIconf website.

7.27.2022 1:02pm - (Replying to @dylanarmbruste3) It is not really an "idea" but the "heart" of causal inference, without which causality is reduced to statistics. That's why statisticians like Rosenbaum, Imbens and Rubin love it, at the cost of being unable to solve the toy problems I present here:,

7.27.2022 1:25am - (Replying to @PHuenermund) Yes. The problem Manski complains about is "missing data", not "measurement error", but his aversion to using graphs is universal, stronger even than Imbens and Rubin.

7.27.2022 12:32am - (Replying to @PHuenermund) Manski does not like Rubin's imputation, for reasons different than ours: But when he talks about "credible assumptions" you can tell he hasn't seen "missingness graphs" yet, without which one cannot judge the credibility of one's assumption. Try it!

7.27.2022 12:30am - Thanks for setting up the audio-visuals for my talk tomorrow, and for alerting me to update my slides. See you at @IJCAIconf 2022.

7.26.2022 7:05pm - (Replying to @erikbiz and @JewishJournal) Not only Arab countries are "threatened", every Palestinian person is personally threatened by the presence of a sovereign Jewish state, no matter how far, no matter how small, no matter how meek.

7.26.2022 6:21pm - Education Dept. to Investigate USC’s Handling of Antisemitic Harassment Against Rose Ritch via @jewishjournal
7.26.2022 6:21pm - An endless cycle: A university is investigated for harassing a Zionist student, and responds with "Look how great we are in talking against anti-Semitism." Moreover, Jewish leaders thank them for being invited to a "committee", instead of demanding they learn to spell "Zionism".

7.26.2022 10:04am - (Replying to @papa_cientifico) In a specific example? Or in a formal model?

7.26.2022 5:03am - (1/3) Yesterday's article by @DrMichaelOren mentions a comparison (made in @TheAtlantic ) between the death of #ShireenAbuAkle and the premeditated murder of my son, Daniel Pearl, in 2002. It compels me to reiterate my reaction to such outrageous comparisons:
7.26.2022 5:03am - (2/3) They highlight the difference between a professional journalist, who never forgets the human face behind the news, whatever the news, and a propagandist, who consistently dehumanizes one side in a conflict. I have humbly asked: In her 25 yrs of covering the Middle East,
7.26.2022 5:03am - (3) has Abu Akleh once described an Israeli child or a grandmother or any human feature of 8 million people trapped in a conflict? If the answer is NO then, regardless of the tragic circumstances of her death, describing her as a model journalist degrades my son's profession.

7.25.2022 6:28pm - This week's harvest of Causal Inference papers is listed here:, for your scrutiny, enjoyment and panoramic overview of where the field is going.

7.25.2022 10:49am - Why I am helping to start a new Jewish Studies Network? Adam Fuller speaks for me. Plus, I have seen 1st hand what happened to Jewish Studies at UCLA and neighboring campuses - our students deserve scholarly alternatives. @GillTroy @EinatWilf @blakeflayton

7.25.2022 9:41am - The celebration cannot be complete without Steve Zipperstein's in-depth analysis of the legal and historical forces at play:

7.25.2022 9:31am - Naïve me! And I was taught listening to the other side is a moral duty.

7.25.2022 6:57am - (Replying to @DKedmey) That's a tough one. Partly because I have zero experience with Times Square campaigns, and partly because passengers speak Rung 3 routinely, e.g., "Are you sure, doctor, that this surgery is worth my risk?" Nothing to be intrigued by.

7.25.2022 6:47am - (Replying to @artistexyz @maxkasy and @HarvardEcon) It's worse than geometry without figures. We can do geometry without geometrical figures because our assumption are defensible (eg. Euclide's axioms). In Potential Outcomes, the assumptions are conditional ignorability assertions, the plausibility of which humans cannot judge.

7.24.2022 10:23am - (Replying to @DavidDeutschOxf @artistexyz and @paul_snively) Confessing ignorance: Where can I find what Everett says about "Potential Outcomes"?

7.24.2022 9:22am - (Replying to @artistexyz @paul_snively and @DavidDeutschOxf) I once lamented thus: Probability function P is a powerful thing to have, it even tells us, on its own, how P ought to change when an observation X=x is received. Unfortunately, it does not have the power to tell us how P ought to change when an intervention do(X=x) is enacted.

7.24.2022 7:30am - (Replying to @NeilHudsonC_R @ToKTeacher and @DavidDeutschOxf) The notion of "intervention" is not defined in probability theory. If it were, then causality would be reducible to statistics and "correlation does not imply causation" would be false.

7.24.2022 5:40am - (Replying to @DavidDeutschOxf) A similar obstacle plagues the more modest endeavor of mimicking human thoughts (as opposed to seeking general theories). Subjective probabilities, when applied to explanatory theories, do not peak with growing evidence, so what's the point in being Bayesian? The notion of "intervention" is not defined in probability theory. If it were, then causality would be reducible to statistics and "correlation does not imply causation" would be false.

7.24.2022 5:03am - (Replying to @paul_snively and @DavidDeutschOxf) To the best of knowledge, Jeffreys/Cox/Jaynes never dealt with counterfactuals, hence, they could not distinguished Theory-1 from Theory-2 if the two gave rise to same probability. IOW, they missed the interesting part of scientific development.

7.23.2022 11:14pm - A harsh indictment of American higher education. @HusseinAboubak , how about sharing a few concrete incidents from your experience at GWU, to see if they match with mine or my students.

7.23.2022 1:23pm - How lovely are your tents, O Jacob; your encampments, O Israel! How safe is your journey, planet earth; Behold, your guardian angels from above.

7.23.2022 6:25am - @DavidDeutschOxf argues that probabilities may be harmful if our goal is to capture truth in the physical world. I am arguing that probabilities are insufficient if our goal is to capture how people think about the world. The latter defines what AI is trying to accomplish.

7.22.2022 10:23pm - (Replying to @soboleffspaces and @DavidDeutschOxf) I am referring to statements like (Pinker p.149):"the rule saying how much to revise our probabilities (change our minds) when we learn new fact." Equating "our minds" with "our probabilities" is seductive but wrong; much of our mind (eg, counterfactuals) lies beyond probability

7.22.2022 4:27pm - Have a great weekend everyone and, as the Shabbat is descending upon us, sing with me: Shalom Aleichem Malachei Hasharet. (Have peace upon you, angels of His message) Composed: Tsfat, Galilee, 16th Century

7.22.2022 11:29am - (Replying to @DeborahProudJew @AbigailShrier and @womenmarch) My heart goes for her - Saint Mother Teresa.

7.22.2022 9:57am - A youtube video of a superb talk by Steve Zipperstein (on one-hundred years to the Israel-Arab conflict) is now available here: I've vowed to memorize all the dates that Steve mentioned, before my next debate with the "new historians - BDS style".

7.21.2022 4:10am - (Replying to @stephensenn and @smueller) We do understand the problem intuitively, but when we speak about "calculus", we need to translate our understanding into formal specification. That's why I am interested in how Nedler specifies the two designs. BTW, would things change if we replace "centers" with "age-groups"?

7.21.2022 2:03am - If you are attending the ICML conference, hungry to learn about non-iid data, don't miss our paper: It will be presented by @zcccucla , at Session P15.

7.21.2022 1:52am - I haven't mentioned other gold mines, because I assumed RCT trialists are interested only in Average Treatment Effect, nothing else. For enlightened trialists interested in distinguishing "harm" from "benefit", we have a 3rd gold mine -- Rung-3 analysis:

7.21.2022 8:07pm - New Poll: Andy Levin was not very smart when he endorsed Rashida Tlaib and thus turned this election into a one-issue race: Whether she is fit to serve in the US Congress. He miscalculated how strongly voters feel against Zionophobic racists.

7.21.2022 2:15pm - (Replying to @GadSaad @JoeBiden and @KamalaHarris) As an American and an uncompromising atheist, I pray to the All-mighty! Oh God! Not @KamalaHarris . This country needs a Golda Meir!

7.21.2022 2:03pm - (Replying to @AndrewPGrieve) Well put! Carefully avoiding the phrase "the true value".

7.21.2022 1:56pm - (Replying to @exactsake) I don't know how to answer your question, because you are using probabilistic terms and, we know, they are orthogonal to causality. "No causal in, no causal out" (Cartwright)

7.21.2022 11:34am - (Replying to @mishtal) This was one land auction I always kill myself for not bidding on.

7.21.2022 7:34am - (Replying to @GedigMoritz) You start by defining what you want to "find out". Note that C is conditional dependent on A, given B. Note also, that this example illustrate the boundary between Rungs 2 and 3. (see Causality page 24)

7.21.2022 7:26am - You are right! It was July 21 2002, 20 years ago that Bernard Lewis wrote this Essay for our "I am Jewish" book. It may have been his influence that made me avoid the word "Judaism" in favor of "Jewishness" - the former is overly theological.

7.21.2022 7:34am - (Replying to @Meh3625555 and @PosenIzzy) My problems with EDI: (1) No list of "ethnic groups" is available. (2) "Jews" are not mentioned in ANY of their official documents. (3) If we demand Jewish representation, they will choose JVP, which is worse than no Jews. Campus discrimination targets Zionism, not Judaism.

7.21.2022 6:17am - (Replying to @RachelMiniK and @PosenIzzy) You will be surprised to know that ALL Diversity officials in the US think that way. I haven't met ONE (I mean ONE) that understands what "diversity" is, namely, that excelling & contributing despite ethnic discrimination is more inspiring than color-based statistics.

7.20.2022 9:23pm - Tomorrow morning, 9am Pacific, Steve Zipperstein will be giving a talk on the 100th anniversary of the British Mandate for Palestine (july 22, 1922). Here are the details, in case you're interested in a thorough understanding of the ME conflict

7.20.2022 7:27pm - Congratulations to @bschoelkopf on a well-deserved recognition and, more importantly, on a breakthrough 2012 paper that has built bridges between causality and machine learning.

7.20.2022 7:00pm - Engineers on the brink of extinction threaten entire tech ecosystems via @theregister
7.20.2022 7:00pm - As a PhD in electrical engineering, I feel sad to see this field on the brink of extinction. I've always attributed my love for science to my excellent EE training. But the show must go on. Are we about to see similar extinction in data-centric enterprises?

7.20.2022 6:37pm - @eliasbareinboim shares the slides of his ICML tutorial on fairness: It contains a critical appraisal of the fairness literature, and solutions based on the logic of SCMs: My Question: Is there a non-causal "fairness"? #datascience

7.20.2022 6:04pm - The instinctive commitment of Jewish leadership to the protection of religious minorities comes from centuries of history. The enigma is why it took them two decades to find out that organizations like CAIR, MAS and MPAC have nothing to do with the protection of minorities.

7.20.2022 1:57pm - Tsadik Katammar Yifrach **** The righteous blooms like a palm tree K'Errez Balvanon Yisgeh **** Like Cedar in Lebanon climbs he, LeHagid Ki Yashar Adonay **** To tell us how fair God must be Tsuri V'Lo Avlata Bo **** My rock, No wrong in him.

7.20.2022 6:21am - (Replying to @ewerlopes and @PHuenermund) Can you narrow down the "topic" to a twittable size? Else I'll have to send you all the papers I wrote since 1980. Here they are:

7.20.2022 6:15am - (Replying to @jonathanrichens @RWJE_BA and 9 others) If you take SCM to be "mechanistic," I'll agree, otherwise, I would love to see some example of "mechanistic assumptions" which are not captured in SCM.

7.20.2022 3:45am - These two gold mines could also rescue experimental economists, many of whom believe they don't need a model once they find an IV (eg, But, DAGs truly life-saving gold mines are recognized by confounding-challenged economists, in non-experimental setups.

7.19.2022 7:22pm - (Replying to @elderofziyon) They don't get it. It is so much easier to forgive an anti-Semite than a Zionophobe. The former evokes stories of unfortunate upbringing, religious prejudice, alcoholic parents etc. The latter are calculating eliminationists with no excusable circumstances -- totally despicable.

7.19.2022 4:37pm - Begging the question: Any reason they can't do the same with women's right to abortion?

7.19.2022 4:20pm - (Replying to @Dem_Bitz) Great painting. Thanks for posting.

7.19.2022 1:21pm - FYI, the harvest is actually richer than I thought. The link is the same,, but the list is now 8 instead of 5 pages long. Let us know (in our input-output language) if you think any of these papers is a breakthrough that should not be missed.

7.19.2022 9:24am - FYI, this week's harvest of causal inference papers: As you can see, it's richer than usual, presumably due to the large number of conferences this time of the year. Or, perhaps our friends in #datascience & #EconTwitter are sensing where the future lies?

7.19.2022 6:52am - The Modern Language Association, Antisemitism and Anti-Zionism –
7.19.2022 6:52am - Every academic scholar or educator should take careful notes of the techniques used by Zionophobes to take over academic institutions like MLA, and turn them into one-issue, foreign-policy mouthpieces. The main enabler: Faculty silence. Any members of MLA among our readers?

7.18.2022 11:46pm - (Replying to @Amol_Sharma) I heard they banned it in Iran.

7.18.2022 11:31pm - DAGs have two gold mines to offer RCTs trialists: (1) Cleaning up RCTs imperfections (eg, selection bias, residual confounding, placebo effects...) (2) Integrating data from multiple sources (eg. multiple RCTs, experimental and observational studies) #datascience #Bookofwhy

7.18.2022 10:21pm - (Replying to @good2epic) Strange, my biggest complaint in DS/ML in industry is almost opposite to yours; too many think probabilistically, at whatever level, and few if any think about the decisions they are about to make and how the data impact that decision.

7.18.2022 10:15pm - I've scanned the TOC and Introduction and I have strong suspicion that the course stays on Rung-1 of the Ladder, namely, limited to prediction and association. No harm in it, but students who are looking for Rung-2 and Rung-3 may be disappointed. #datascience

7.18.2022 4:52pm - And here another interesting paper on "harm" and "benefit", this time conditioned on variables that, in themselves, may be affected by the treatment: #datascience #economics #MachineLearning #Bookofwhy

7.18.2022 2:56pm - For readers interested in the epistemology of causal inference, I've revised my commentary on Dawid's "Decision Theoretic Foundations of Statistical Causality" and have posted a clean copy here

7.18.2022 2:47pm - (Replying to @tom4everitt @RWJE_BA and 8 others) Thanks for posting this paper about "harm". "Harm" and "benefit" have been the subjects of our latest papers on "personalized decisions" and Would be interesting to compare results.

7.18.2022 2:56pm -

7.18.2022 1:29pm - (Replying to @soboleffspaces and @IJCAIconf) Yes, but what is their motivation for relying ONLY on randomization and resisting any knowledge that the investigator may have acquired by experience, education and judgment.

7.18.2022 7:14am - (Replying to @GinoDelFerraro) Progressives do not count wounds, we count choices.

7.18.2022 7:05am - The ship that opened the world's eyes to British Govt treatment of Holocaust survivors, and swayed public opinion in favor the UN vote (33:13) for partition of Palestine (Nov 29, 1947) despite Arab's promises of "War of Extermination" (Azzam Pasha, Oct. 11, 1947).

7.18.2022 6:45am - (Replying to @MissDiagnosis and @IWF) Pres. Roosevelt died in 1945.

7.18.2022 6:39am - (Replying to @mishtal and @freemonotheist) Those who have no history must fake one, even at the cost of being caught.

7.18.2022 1:28am - (Replying to @maradonasattva and @BillDonovan1234) Judaism without Zionism is like Christianity without a Cross. Not entirely infeasible, yet entirely a-historical and pitifully crippled.

7.18.2022 1:17am - @POTUS main achievement: allaying fears that leaving #Afghanistan was phase one of leaving #MiddleEast. Saudi "no normalization before X" is actually hopeful, b/c X was phrased as "agreement" on 2-states, not implementation thereof.

7.17.2022 5:54pm - (Replying to @soboleffspaces and @IJCAIconf) Who are the algorithmists?

7.17.2022 10:16am - (Replying to @Wurmserscribit) I know that Biden knows who the oppressor is in the Israel-Palestinian conflict. I assume therefore that he chose this rhetoric as the only common denominator he could find with his audience, who, btw, also knew what he thinks of them.

7.17.2022 10:10am - Extremely interesting paper: BTW, what is the current status of the statistical marginal problem?

7.17.2022 7:40am - That's why we call it "data science" - a two-body enterprise - the data "suggest" and the science "reasons".

7.17.2022 7:05am - (Replying to @KPMDas) Many tell me: "OK, so the NYT has an obsession with Israel but, by and large, it is still "good journalism"". I say: I can only judge them in areas that I've some competency, Israel and AI, and if they are glaringly so anti-professional in the former, I can't trust them anywhere.

7.17.2022 6:41am - (Replying to @HillCitizen) "Saying" and "documenting" quantitatively are too different things. I have some fairly rational friends telling me: "I read NYT and haven't seen anti-Israel bias". Seeing the word statistics may convince them.

7.17.2022 6:37am - (Replying to @EvanHoffman) Unfortunately, terrorism has been normalized in "freedom fighters" land, so it carries zero weight in public discourse.

7.17.2022 6:33am - (Replying to @adamalmo) These two papers summarize the main thrust of my talk:,

7.17.2022 4:15am - Comparing Irish and Palestine? I don't recall Irish Catholics shouting "London is a stolen land!" or "Dismantle Great Britain". Have you?

7.17.2022 3:17am - To restore your reputation, you must remind the public what your values are, and to state explicitly that they include dignity and safety for Anna's brothers and sisters who survived the camps and have found hope and new life in Israel. Vagueness on values projects empty values.

7.17.2022 2:41am - I am delighted to hear that Michael Waldmann is hiring a postdoc for the project “Mechanisms, Capacities, and Dependencies: A New Theory of Causal Reasoning”, funded by the German Research Foundation (Koselleck program). See for more info.

7.17.2022 12:15am - Top Israeli Daily’s Exposé Paints Troubling Picture of New York Times’ Israel Coverage
7.17.2022 12:15am - The poetic power of statistics. Here are irritating facts that we (readers of NYT) encounter on a daily basis, but when quantified and compiled into statistical patterns, unveil a purposeful design.

7.16.2022 11:48pm - (Replying to @soboleffspaces PS/IPW has also been used in #MachineLearning as a disguise for doing statistical estimation and calling it "causal inference". #Bookofwhy, #DataScience

7.16.2022 5:21pm - Philosophers, linguists, and XAI practitioners would find this paper of interest. It deals with evidential and causal conditionals and proposes a system that handles both: I need to understand why my proposal in is inferior.

7.16.2022 12:40pm - The Daniel Pearl Foundation is not as powerful as the @AnneFrankTrust but, believe me, we screen our speakers very carefully. See

7.16.2022 12:12pm - This is precisely what fascinates me about Diplomacy. What you say or do does not represent your beliefs; it represents who you wish to please. "Peace" pleases Israelis, "UNRWA" pleases AOC, and both sound so noble.

7.16.2022 3:57pm - (1/3) The habit of assuming away the identification problem and treating matching and PS methods as if they reduce confounding bias comes from surveys like this one: (from ACM Trans. on Knowledge Discovery from Data) which is wedded to Potential Outcomes
7.16.2022 3:57pm - (2/3) and mentions SCM as an "equivalent approach". Missing is the warning that logical equivalence is not conceptual equivalence: In particular, that the PO approach "forces investigators to articulate assumptions whose meaning they do not comprehend,
7.16.2022 3:57pm - (3) whose plausibility they cannot judge, and whose statistical implications they cannot predict." See: One explanation for the prevalence of this habit is that scientists fear transparency of assumptions they must defend. Habits are preferred.#datascience

7.16.2022 11:55am - It's true, with one exception: The official title of my talk reads: “What is Causal Inference and Where is Data Science Going?" See you there. #Bookofwhy, #datascience, @IJCAIconf

7.16.2022 7:42am - @julius_adebayo sharply dissects the XAI enterprise to its core, but his conclusion: "explainable ML = model debugging" still leaves an elephant in the room: A model of what?" of the ML program? Or a model of reality?. Examples abound where the former doesn't match the latter.

7.16.2022 5:30am - (Replying to @yudapearl @roydanroy and @nickchk) To answer your question, Yes. David assumes away the identification problem (using the magical un-confoundedness assumption), which reduces causal inference tasks to statistical estimation, and gives ML folks the impression that "it's all in the data". Not advisable.

7.16.2022 5:13am - (Replying to @roydanroy) I am pointing to backdoor admissibility because it is necessary for deciding what features to match on, and which ones to exclude from matching. #Bookofwhy has example where matching on all features leads to wrong conclusion. @nickchk makes this point nicely in "The Effect".

7.15.2022 6:33pm - (Replying to @jeffreywatumull and @DavidDeutschOxf) My first reaction: Bacon, Hume and Popper tried to capture the way science works, this paper tries to capture the way machine learning works. I do not see the connection given how far machine learning is from emulating scientific thinking. (e.g., counterfactual thinking)

7.15.2022 1:54pm - (Replying to @GadSaad) Satirical or not, this outspoken protects the honor and credentials of a tenured UC professor. Hats off, outspoken!!!

7.15.2022 9:27am - In Bethlehem, Biden pledges to keep up efforts for Israel Palestine peace
7.15.2022 9:27am - Biden: "Even if the ground is not right at this moment to restart negotiations" It seems that I would never be a diplomat. I would never be able to control myself from telling Abbas "WHY not restart now?", and what he can do about it. Diplomats can ignore giant elephants.

7.15.2022 4:40am - Try reading some of the graphical causal inference literature and I bet your enthusiasm will increase 10 fold. E.g.,, Specifically, introducing matching and propensity score w/o backdoor admissibility is not advisable. #datascience

7.14.2022 4:01pm - (Replying to @Eve_Barlow) Here is the irony. Jews get their strength from history, and their history has culminated in Israel. The poor souls that reject Israel have lost the strength to stand up to their bullies -- a vicious circle. I empathize with their dilemma here:

7.14.2022 1:41pm - (Replying to @artistexyz and @soboleffspaces0 I tried to follow it, but I couldn't. Do you have a gentle tutorial to the method?

7.14.2022 12:04pm - (Replying to @DKedmey) I think it worked; the smearing guy stopped. I believe these guys simply do not see the difference between talking and smearing; typical of Al-Jazeera news coverage, where the facts/accusations ratio is below 10%.

7.14.2022 9:46am - (Replying to @WelshBrendan2) I understand. The question is whether a peace-seeker like yourself would be willing to accept the harsh facts as learned by another peace-seeker (myself) who has been searching for a moderate voice the past 75 years -- there isn't any, and it ain't "climate." Hard to swallow!

7.14.2022 9:33am - (Replying to @WelshBrendan2 and @CotlerWunsh) So you, personally, believe Israel "operates under an apartheid model". Am I right? You do not buy the possibility that the "apartheid model" was attached to her out of pure malice?" Am I right?

7.14.2022 9:12am - (Replying to @soboleffspaces @JohanDH2O and 2 others) For me, getting meaningful definitions of NIE and NDE was more important that preserving the additivity property, which I thought to be a disposable feature of linear systems.

7.14.2022 7:56am - (1/2) Certain breed of Jews, mainly those who believe that Israel is a colonial apartheid state, get extremely upset when called "Ex-Jewish". Some even call me "anti-Semite" for suggesting their exclusion from the Jewish fold. I'm willing to reconsider and call them:
7.14.2022 7:56am - "colonial apartheid Jews", except they want to be treated as mainstream. I once proposed "The New Marranos", and explained why they should not be given a podium at the Bimma - our universities. Today, I much prefer "Ex-Jewish", or "Yester-Jews".

7.14.2022 6:50am - Finally, the Jerusalem Declaration signed! The key sentence: "The United States stresses that integral to this pledge is the commitment NEVER TO ALLOW Iran to acquire a nuclear weapon, and that it is prepared to use ALL ELEMENTS OF ITS NATIONAL POWER to ensure that outcome."

7.14.2022 4:03am - Confession, I could not hold my tears when President Biden listened to these two ladies and, ignoring all the "time is up" hints from his entourage, including Blinken, continued to listen to what his visit meant to people who survived Auschwitz and have made their home in Israel.

7.14.2022 3:44am - (Replying to @tomgoldsteincs) This is a legitimate question for every acronym in machine learning. I've asked it on Stanford's "Foundation model" with no answer in sight. The problem is that acronyms are attached to successful programs, not to scientific principles. #datascience, #Bookofwhy

7.14.2022 3:30am - (Replying to @hari_hari) I am searching for the fundamental ideas, and all I can find are different labels of the same problem, coined by different groups in DARPA .

7.14.2022 3:18am - I don't know much about pregnancy, borders or obesity, but have had some fun in science, till yesterday, when I read that it's more racist than Trump, w/ all its equations named after white male adults; Galileo, Newton, Maxwell, Einstein -- incurably oppressive to most students!

7.13.2022 8:40pm - (Replying to @shagunsodhani) To those versed in causal inference, what you are describing seem to be minute differences of emphasis in several versions of "Data-Fusion" theory I wonder if lifelong folks looked into the solutions offered by Data Fusion.

7.13.2022 7:23pm - President Biden has given Jewish students their first line: "You don't need to be a Jew to be a Zionist." Their next line should be: "Dear Chancellor, You can't bring yourself to state that Zionist students are more than welcome on campus? Aren't you the one who sets norms here?"

7.13.2022 6:34pm - I've always wondered, but was afraid to ask: How does "lifelong learning" differ from “transfer learning,” or “domain adaptation." In other words, would there be a problem of "lifelong learning" if domains stayed constant? Or, won't LLL be solved if we solved "transfer learning?

7.13.2022 6:10pm - Congratulations to @tgjelten & @MomentMagazine on winning two first place AJPA awards, and for the excellent investigative story on the plight of the Uyghurs in China; a topic that ceased to interest social media activists since the start of the war in Ukraine. @NadineEpsteinDC

7.13.2022 5:31pm - (Replying to @malickheydar and @Harvard) Me too. I just gave a copy of #Bookofwhy to my dentist, dedicated with: "Easy on my wisdom tooth, doctor, it ain't dead."

7.13.2022 5:18pm - (1/2) Thanks for re-twitting this timely post. As more economists enter the world of DAGs, I sense their genuine appreciation of DAGs as a communication language, but insufficient appreciation of DAG's computational power. This post shows that even in
7.13.2022 5:18pm - simple IV models, you can't go very far without DAGs. Imbens's claim that economists indulgence in simple IV's absolves them of needing DAGs is grossly misleading. Same with Heckman's disposal of d-separation @causalinf , @nickchk

7.13.2022 3:54pm - (Replying to @RitchieTorres) I salute you @RitchieTorres for articulating so clearly what the nature of BDS is, and what your vision is of ME peace, based on "co-existence", a word that no BDS crony has ever uttered. You have my back.

7.13.2022 3:42pm - (Replying to @mynameisjro and @RitchieTorres) I salute @RitchieTorres for speaking for our people, because those EX-Jewish New Yorkers who believe that Israel is an apartheid state, and that BDS is not eliminationist, have abandoned us, in need of a non-Jewish Congressman to counter our predators in the US Congress.

7.13.2022 7:19am - I once put it this way: "The landscape has become a scripture". Lord Balfour put it even nicer: ".. intimately bound up with a petty territory wedged in between States more powerful far than it could ever be". See

7.13.2022 12:50am - (Replying to @GadSaad and @AyannaPressley) I do have the capacity of answering this question, but I'd rather keep mouth shut for fear that "my line of answering" would render me vulnerable to accusations of transphobia (or truthphobia) hence indirectly responsible for suicides among bald white men, a protected minority.

7.12.2022 4:25pm - (Replying to @adamalmo) These two papers summarize the main thrust of my talk:,

7.12.2022 4:20pm - You caught me off guard. I was going to jump out of my skin and shout: "Peace! finally!" before I saw the humble #WishfulHeadlines. Back to the real Mahmoud Abbas.

7.12.2022 4:15pm - This red carpet is longer than River Jordan It's purpose to generate media focus on the real issues of ME, keeping @AJEnglish and other noise makers out of the conversation. My question: Will Biden treat Palestinians as responsible agents, or just reinforce their passive posture.

7.12.2022 8:57am - Here are some facts that even Israel's worst enemies know to be true, and which they find to be just nasty obstacles in their heroic mission to prove the opposite.

7.12.2022 7:52am - (Replying to @richard_landes) A Palestinian never says or "I recognize Israel", it's always: "Arafat did, in 1990", forgetting that Israelis have learned something in the past 30 years, especially to demand a fresh Arabic copy of the PLO charter, w/ the words "right to exist" spelled out for children to read.

7.12.2022 6:56am - Glad the Academy of Management is about to enter the age of causation, through the very practical and omnipresent question of Good and Bad controls, a question that graph-averse economists have chosen to avoid rather than address.,

7.12.2022 6:31am - This paper, though written in 2014, should be read today in other fields undergoing methodological transformations: especially those laden with paradoxes or controversies (who isn't?)

7.12.2022 4:09am - The market is waiting to see how innovation economists circumvent the infinite inertia of mainstream economics.

7.12.2022 3:30am - Welcome to peace-dodging land! Here is a billboard that would be torn down immediately by our neighbors, yet is put up repeatedly as if we do not know; as if they never told us; as if pretending we are deft would make them change their dream.

7.12.2022 1:27am - How can one tell when Causal Inference (CI) goes mainstream? (1) When universities dedicate buildings as CI-Centers. (2) When recruiters demand CI-qualifications. (3) When US patent office grants patents to CI-ideas. Beware! The latter just occurred:

7.11.2022 6:45pm - For an academic organization (Association for Israel Studies (AIS)) that counts Zionophobes like @Ian_Lustick among its former presidents, breaking ties with MESA must be a traumatic catharsis. MESA can safely change its name now to IMESA - the Irrelevant side of ME Studies.

7.11.2022 1:21pm - FYI, this week's harvest of causal inference papers:

7.11.2022 9:35am - This is a fair comparison. Biden should meet the parents of Malki Roth in Jerusalem. They have been waiting 21 years for justice, and Malki, their slain daughter, was a symbol of goodness and joy.

7.11.2022 9:22am - Informative narrative of the history of Czech-Israel relations. Thanks

7.11.2022 7:29am - (Replying to @DerriereLaLune and @rcguo) Interesting! First time I see it. But I don't buy the argument: "The derivation does not use an intervention on the mediator, which in many situations would not be possible in reality." Derivational step need not be "possible in reality". Best example: Complex numbers in math.

7.11.2022 7:05am - This is all true. I'm honored to be speaking again at IJCAI (first time was in 1983 and to challenge the direction "big data" is taking. No, I'm not flying to Vienna, but will make my Zoom screen as big as possible. See you there.

7.10.2022 5:58pm - (Replying to @elderofziyon @BenWeinthal and @ShuratHaDin) I'm glad nevertheless that a serious complaint has been filed with the International Criminal Court against the PLO; people just don't get the idea that the evidence is not clear cut, until legal complaints are filed, each supporting one side of the story.

7.10.2022 2:12pm - (Replying to @HaimHabulai) Thanks for the correction, I didn't know it. Kol Halomed MeChavero Davar ....Rabbi

7.10.2022 8:15am - Israeli president to meet with Czech counterpart in Prague via @JNS_org
7.10.2022 8:15am - Whenever I see a Czech flag, I can't but remember Jan Masaryk, the Foreign Minister of Czechoslovakia, who agreed to send Israel 4,000 rifles, thus saving us from a genocidal attack in 1948. My basic training (1953) was still with Czech rifles. Masaryk Square now crowns Tel-Aviv.

7.10.2022 7:35am - Jewish father files lawsuit over discrimination at Los Angeles private school via @JNS_org
7.10.2022 7:35am - As much as I hate law suits, this one is absolutely necessary to encourage parents everywhere to standup to recent takeovers of our high schools by EDI priests, who do not value Jewish students as a distinct minority that can contribute positively to other minorities on campus.

7.10.2022 5:58am - (Replying to @elderofziyon @BenWeinthal and @ShuratHaDin) I'm glad nevertheless that a serious complaint has been filed with the International Criminal Court against the PLO; people just don't get the idea that the evidence is not clear cut, until legal complaints are filed, each supporting one side of the story.

7.10.2022 5:42am - Refreshing sight, even to a Californian like me. My mother used to take us (me and my younger brother) to this very beach every week. You see, she was Polish, and believed religiously that sun rays and sea waters is good for your health. Wow she was right!

7.10.2022 5:28am - (Replying to @kareem_carr) What about for the kind of data scientist that's more like a data-scientists?

7.10.2022 5:16am - Most economists use potential outcomes in conjunction with instrumental variables, not realizing that both are derivatives of SCM, Some begin to use DAGs but, to make them look home-grown, strip them of their main powers.

7.10.2022 4:53am - (Replying to @blakeflayton) We got here (partially) because even Zionist writers use the term "anti-Zionism" as if it was some sort of a legitimate political opinion, instead of using "Zionophobia", which highlights its racist, genocidal and pathological character.

7.10.2022 4:42am - (1/2) Our next heroic effort is (again) in economics. My attention is captivated by a BA thesis on the mediation fallacy and the analysis of police violence: Considering the graph-averse culture dominating economics, and the risk-averse, sheepish style of
7.10.2022 4:42am - most theses (including PhDs) it's both mature and heroic to assert: "DAGs’ main strengths are their accessibility, clarity, transparency, and testability, and the current skepticism by economists against their use is largely unfounded." Prediction: Johanna Barop is going places.

7.9.2022 11:44pm - (Replying to @EricJGreenberg1 @mishtal and @Ixar_Bargains) That was easy. I asked him to sing me a Philistine song, or tell me a Philistine story, or name a Philistine hero he would name his son after, or a Philistine holiday he enjoys celebrating, or, ... or.. It was easy.

7.9.2022 10:13pm - (Replying to @rickwsint and @skdh) Got it, thanks!

7.9.2022 10:11pm - (Replying to @rickwsint and @skdh) I didn't get the better source.

7.9.2022 8:42am - It's called science-centric paradigm:

7.9.2022 8:12am - (Replying to @Ostrov_A @Israel and 10 others0 The one on the bottom left reminds me of Shuck Hapishpeshim (the bed-bugs market) where I bought my first acoustic guitar, in 1976. Glad the place is still striving.

7.9.2022 7:57am - (Replying to @roydanroy) "Swallow", of course, by claim of ownership (called "inductive bias"), but not before ML folks learn to articulate what it is they claim to own.

7.9.2022 7:34am - (Replying to @rasbt) Jobless - not yet, but outdated, on the mark.

7.9.2022 7:31am - Surprise! Ali Khamenei Slams Western Values: A perfect candidate for Dean of Social Science at one of our finest US Universities. The salary and faculty enthusiasm would surely match his current employment benefits.

7.9.2022 6:40am - When ‘Racial Justice’ Means Antisemitism: A Takeover of the Yale Postdoctoral Association
7.9.2022 6:40am - Delighted to see faculty members at @Yale (my semi-alma) speaking out against the Zionophobic takeover of their Postdoctoral Association. Like USC before them, when faculty say "enough is enough", it means: "enough will prevail."

7.9.2022 6:07am - (Replying to @thinkyparts) "Reasoning" entails decision making under murky waters.

7.9.2022 6:04am - This paper by Grosz etal, "The Taboo Against Explicit Causal Inference in NonExperimental Psychology," explains much of the research environment in that field, and why I consider the works of @dingding_peng & associates to be so critical to psychotherapy.

7.9.2022 4:36am - Every country on the right side of reason is a fortress on the right side of humanity

7.9.2022 4:04am - (Replying to @blakeflayton) You are giving readers the impression that Zionism just "does not say they dont have a right". I truth, Zionism explicitly says they DO have this right "even if it compromises Zionism". Why not quote the sources, Ben Gurion, 1931

7.9.2022 3:51am - (Replying to @eIumax) Problematic is the mind that sees the sun, the moon, and living organisms as "problematic"

7.9.2022 3:18am - (Replying to @blakeflayton) Why? Has anyone mistaken it with the civil rights movement of another nation? French? Egypt?

7.8.2022 9:21am - (Replying to @JohanDH2O) Excuse my ignorance, but what ideology does Jordan Peterson represent??

7.8.2022 9:08am - (Replying to @tkaiser_science) Very encouraging, but what is the NAME of this journal? We need to publicize it and commend the editors on being forward looking. (unlike some fields I won't mention).

7.8.2022 8:53am - (1/2) Our next hero welcome goes to an economist (who would ever imagine). Nick Huntington-Klein has a new book which, according to its TOC, promises to deliver economics students from the graph-less torture that their professors went through, and which
7.8.2022 8:53am - (2/2) Heckman calls "Econometric Causality." It starts with research questions and identification, and then goes to the tool box, armed with backdoors, front-doors, perhaps even with do-calculus and the First Law. I've ordered a kindle copy and will let you know more.

7.8.2022 7:23am - (Replying to @MiriamElman @GeraldNGOM and 2 others) Miriam, I need no FYI's. Here is what I posted two days ago: and, you are right, Joe was on the UCLA committee that fought the Zionophobic racism of the Asian-American Studies department, and testified before the Regents. I know we will prevail.

7.8.2022 6:00am - (1/2) The next heroic effort is in psychology. Based on a 2018 paper by Julia Rohrer @dingding_peng My first reaction was: Who is this community of Psychologists who were not convinced by Julia's article? It's hard to believe
7.8.2022 6:00am - (2/2) that they can continue to function in 21st century without causal inference tools. The progress of a field invariably depends on one/two enlightened leaders, so, don't they have any? Kudos to Li Qian and Julia for trying to lift this field out of the stone age.

7.8.2022 5:10am - (Replying to @IllyseusDasein @SamHarrisOrg and 2 others) I do not understand what you are trying to insinuate, but I've known Joe Manson for 20 years. He ain't an oppressor, nor a "self-styled victim", nor any of the adjectives you associate with people you don't like. Please do not use this smear technique on this platform.

7.8.2022 4:48am - (1/2) I would like to dedicate the next few posts to heroic readers who are struggling to introduce causal clarity to fields where such clarity has been lacking. I will start with which focus on neuroscience. My only gripe: they push me into "statistics" and
7.8.2022 4:48am - (2/2) fail to mention that many (if not all) of the diverse types of causes they discuss are now unified through SCM. I never understood those who prefer pluralism on monotheism. See

7.8.2022 3:59am - This film taught me quire a few things I didn't know. But most of all, I liked the dedication to feminists like Linda Sarsour and other fake leaders of the Women March, who continue to peddle their dishonesty to the young and unsuspecting. @ZionessMovement @bariweiss

7.8.2022 3:10am - (Replying to @SAFECUNY) Great! As the first step. The next steps are: 1. CUNY's administrators learn to spell "Zionism" 2. They learn to pronounce it, loud, w/o watching if anyone is listening. 3. They tell the campus what it contributes to diversity and education.

7.8.2022 2:22am - Populist slogans are not supposed to mirror facts or figures; they are made to feed the ignorant and move the mob.

7.7.2022 11:30pm - The next take belongs to the @latimes , whom Californians trust to spot such stories and turn them into a public outcry: "Are we losing our universities?" They used to listen to me, before the woke takeover. Perhaps they would listen to @bariwies.

7.7.2022 11:05pm - (Replying to @KabirCreates @ProfLaurenRoss and 2 others) The beauty of Hume's counterfactual definition of "cause" is that, for 2 centuries, people thought it's final, and no one asked: "So how do you define counterfactuals?". Why? Because counterfactuals are so entrenched in our mind, we thought they're irreducible primitives. Wrong!

7.7.2022 3:34pm - (Replying to @ProfLaurenRoss @KordingLab and @shansiddiqi) I am still trying to understand what people want when they crave for "definition of cause". If they crave for a trick to help them think when they construct a model, then I have a good one: "who listens to whom?". If they seek reduction to probability, we know it's impossible.

7.7.2022 2:57pm - (Replying to @mishtal) Now you see why I play it safe, and call him a "dangerous Zionophobe." He can't sue me for what he is so proud of, though it is ten times more dangerous and certainly more despicable.

7.7.2022 8:33am - As Abdurrahman Wahid said (former President of Indonesia): "Countries that deny the Holocaust also tend to be backward and repressive." Why? "Because those who are dishonest when it comes to the past are hardly in a position to reckon honestly with the problems of the present."

7.7.2022 5:25am - Johnson's resignation speech starts 9:20 into this video Beautiful words. Gloomy future. Hail to Britannia -- God Save the Queen

7.7.2022 12:15am - I can't dive too well, but if anyone is swimming already in the Indian Ocean, I would be interested to examine the earliest counterfactual expression in ancient Indian literature. Similar to the one spoken by Abraham: "What if there were 50 righteous people in the wicked city?"

7.6.2022 7:34pm - A confession of a repentant Zionophobe.

7.6.2022 5:45pm - The New Zionist Congress - 2022 @blakeflayton @newzionists

7.6.2022 5:05pm - Photo of the leaders of Palestine and Hamas, very symbolic, but this will not improve the situation via @ePrimefeed
7.6.2022 5:05pm - "The New Zionist Congress - 2022"

7.6.2022 8:23am - An important page of history. The event that led to Israel's "defensive shield" campaign, which some regard (including me) as the only effective "war on terror" by any country, though still continuing.

7.6.2022 7:36am - (Replying to @EpiEllie @KordingLab and 2 others) Beautiful introduction to a Short Introduction, and very useful, for every CI scholar, to collect this arsenal of examples to communicate nuances and variants of causal thoughts. Thanks.

7.6.2022 3:26am - (Replying to @MechncOfMeaning) I'm only concerned about the last line. do(c) always appears under a probability expression. e.g., P(r|do(c)) = P(r)

7.6.2022 3:19am - (Replying to @intrinsic_motiv and @ylecun) What do you expect to get by implementing this strategy? I am asking, because a theory is now available that can tell you what the strategy can and cannot do. I would recommend that use the theory to guide your implementation. e.g.,, #Bookofwhy

7.6.2022 2:45am - (Replying to @JGreenblattADL @almog_doron and @JewishAgency) Congratulations, Almog, on assuming this leadership. You can count on our help.

7.6.2022 12:31am - Yes, you can get the Solution Manual if you write to my assistant

7.6.2022 12:03am - Can you help me climb this paywall? It tells me: "Access to this article via University of California Los Angeles is not available." Why? I think I've lived a clean life!

7.5.2022 11:52pm - Agree full-heartedly. Unfortunately ML folks are busy building Gothic Cathedrals for a handcuffed God, funding them, and convincing investors that Gods should know how to liberate themselves.

7.5.2022 11:37pm - I would recommend Primer which is available free (chapter by chapter) and logically coherent, starting with the First Law and ending with Tool Kits for Mediation and Attribution.

7.5.2022 11:30pm - (Replying to @giulioprisco @SMcfarnell and 12 others) Interesting conjecture which has skipped my #FreeWill optics. I am inclined to surmise that the philosophical debate is orthogonal to the political, with the latter misunderstanding the former.

7.5.2022 11:14pm - I was right then in predicting that Imbens writings and prizes will entice people to study causal inference the scientific way, starting with the First Law: @causalinf #Bookofwhy @VC31415

7.5.2022 10:46pm - (Replying to @animesh1977) Viewed as a threat, yes, but I don't recall the "clowns" era.

7.5.2022 10:41pm - For readers and teachers of Causality (2009), viewgraphs and exercises are available here: Alternatively, Primer is loaded with exercises:, & a solution manual is available. Please RT, to let others know, especially #Bookofwhy folks

7.5.2022 10:17pm - (Replying to @texan_marshall) I suspect if such a system turns successful, it would spawn a movement to "defund AI," which no AI researcher wishes to see happening.

7.5.2022 4:42pm - When an activist quotes @AlJazeera and says: "It truly appears to be XYZ" you can bet on two things: (1) He knows nothing about what happened, and (2) what happened is definitely not XYZ.

7.5.2022 4:09pm - Defeat Overture INT o2 PCUSA - Sign the Petition! via @Change
7.5.2022 4:09pm - I've signed this petition, because it hurts me to see my friends in the Presbyterian Church bullied by hard core Zionophobes who took over the PCUSA administration. Please join me in defeating overture INT 02.

7.5.2022 3:13pm - (Replying to @equilibriumcow) Yes, this is the page. I'll tell readers to go there. A more gentile technical introduction is Primer It has exercises and solutions.

7.5.2022 5:57am - (Replying to @equilibriumcow) Thanks for the comments, but which "causality" do you have in mind, and which website are you referring to?

7.5.2022 4:02am - I'll second your recommendation. I truly have no idea how many ML folks read #Bookofwhy. There are days when I think: "zero". And then, when I see them struggling to emulate causal thoughts, I think: "99%". Hard to tell.

7.5.2022 1:11AM - To complete my story about the Kielce Pogrom below is a picture of my aunt, Mania (an Auschwitz survivor), during the funeral, consoling families of the victims. (Courtesy of her son, now living in Israel, a proud graduate of my own college, The Technion.)

7.4.2022 9:19pm - (Replying to @Wetassprior and @KeurigBad) No, Z is a son of W, so, conditioning on Z opens the backdoor path X<--U2--->W<--U1-->Y.

7.4.2022 9:12pm - Dan Rather: "The U.S.A. is and always has been an idea more than a place" " Made me wonder: How many societies think like that about their countries? I know Israelis do. But I don't get it from my Canadian granddaughters. Switzerland? Egypt? Anyone knows?

7.4.2022 2:07pm - I've recently took the liberty of calling it "Ladder of Knowledge" rather than "Ladder of causation"; I've seen it govern areas that are not immediately associated with causation.

7.4.2022 1:43pm - On July 4th we realize that the words "defending freedom and democracy" are more than a broken record. It hits you that certain freedoms and certain democracies do need defending.

7.4.2022 1:23pm - This hits close to home. My mother's family is from Kielce. And her sister witnessed the pogrom in her own eyes, miraculously survived, and later eulogized the victims in the cemetery, as was recorded on movie camera. Only one year after liberation.

7.4.2022 11:12am - What a coincidence for July 4th holiday. An important addendum, for American Jews to remember. In 1997, Herzl said:“Zionism is a homecoming to the Jewish fold even before it becomes a homecoming to the Jewish land.” We can see it being incarnated today, in the US.

7.4.2022 10:30am - I'm still surprised that no one in the media raises the question of motive: Why would Israel aim to kill a journalist when she has everything to lose and nothing to gain from such act?

7.4.2022 9:27am - Happy 4th of July! On this day, it is meaningful and relevant to read aloud Longfellow's poem: "Thou, too, sail on, O Ship of State! Sail on, O Union, strong and great! Humanity with all its fears, With all the hopes of future years, Is hanging breathless on thy fate!" How true!

7.4.2022 12:44am - (Replying to @arbutuspointBC) You can tell from his logic that he ain't likely to be a fan of Israel or of other islands of commonsense. @EnglerYves

7.4.2022 12:34am - I have elaborated on these Eight Myths in "The Causal Foundations of Structural Equation Modeling", an updated Chapter for the upcoming R. H. Hoyle (Ed.), Handbook of Structural Equation Modeling. New York: Guilford Press,

7.4.2022 12:22am - (1/2) Toronto logic at its glory: "Irrespective of the veracity of accusations, rallying for the accuser is reasonable!" Imagine rallying against Ukrainian borscht at TO store called “Taste of Ukraine” because Putin called Zelensky Nazi and deemed Ukraine "stolen land."
7.4.2022 12:22am - (2/2) This is the new logic Toronto shoppers are encouraged to follow at a food store called "Taste of Israel". Thanks to @EnglerYves for unveiling this logic, and to @MelissaLantsman for exposing its dangers.

7.4.2022 4:10pm - (Replying to @robertclab) I doubt people who shun DAGs even for simple problems would fall in love with SWIGs when problems get harder.

7.3.2022 11:45pm - (Replying to @arbutuspointBC) Toronto logic at its glory: "Irrespective of the veracity of accusations, rallying for the accuser is reasonable!" Imagine rallying against Ukrainian borscht at TO store called “Taste of Ukraine” because Putin called Zelensky Nazi and deemed Ukraine stolen land. Grand-Logic!

7.3.2022 7:42am - (Replying to @anon61845948 @jrootham and @MelissaLantsman) Beg to differ. An ethnic group that equates its liberation, loudly and unequivably, with the elimination of another ethnic group is not equally deserving of support and respect. Please listen to how THEY define liberation, not how Westerners wish them to define it.

7.3.2022 1:37am - PM Lapid spoke for me when he said:" It was here before us, and be here long after us". I will vote for him, if I'm so lucky as to be in Israel November 1st.

7.2.2022 12:20pm - Every time a reader retweets this interview I am tempted to re-listen to it and then wonder how the interviewer, Guido Imbens, could possibly continue to work with Potential Outcomes instead of switching to graphs. An resolved mystery of scientific evolution.

7.2.2022 11:41am - Israel is the only country whose future depends on its neighbors textbooks more than it depends on their guns.

7.2.2022 7:42am - (Replying to @hdjnicklin) Your logic is right, but missing one fact: Palestinian riots of 1936-1939 forced the British Government to stop immigration from inferno Europe, thus sealing the fate of 6 million Jews who sought refuge in Palestine, the only country where they were welcome as brethren.

7.2.2022 7:30am - (Replying to @dizzdangles and @MelissaLantsman) You got a point, if it weren't for the death-threat encoded by the "Free Palestine" slogan, these "protester were shouting.

7.2.2022 7:17am - (Replying to @jrootham and @MelissaLantsman) Not sure if you seriously do not know what "Free Palestine" means to those who shout it.

7.2.2022 7:11am - @MelissaLantsman , you are talking as if anti-Zionism is an excuse for this form of protest. Is it? Suppose it was purely anti-Zionist, soaked with angelic love for non-Zionist Jews. Would it be less racist? Less morally deranged? Less despicable?

7.2.2022 6:46am - COVID-19 hero: Pfizer CEO Albert Bourla reflects on the past, future

7.2.2022 12:53am - Only a Zionophobe like @KenRoth would retweet this misleading statement. It was not the Presbyterian Church that voted, but its lopsided Committee on International Engagement. The overture will be voted on by the entire membership next week. Hoping Presbyterians defeat INT 02.

7.2.2022 12:25am - I wish left extremist would be as generous and accuse us of things we are proud of.

7.2.2022 12:15am - Jewish Groups Criticize CUNY Chancellor for Not Appearing at NYC City Council Antisemitism Hearing via @jewishjournal
7.2.2022 12:15am - Glad the conversation at CUNY is converging toward the real issue: Welcoming Zionist students as a valuable minority group. Sadly, CUNY adm. has not passed the litmus test for open-mindedness: Spelling the word Zionism.

7.1.2022 1:18pm - (Replying to @stephensenn) By "asymptotic" I mean "as the number of samples in each center increases indefinitely". Isn't ACE an asymptotic concept? (assuming the 'A' in ACE stands for 'average', not sample 'mean.'

7.1.2022 1:09pm - I welcome the emergence of the new Jewish Studies Zionist Networks. From now on, anti-Zionist professors will not be able to say "We represent THE Association for Jewish Studies" without specifying WHICH ONE; the old and irrelevant, or the new, vibrant and organic? I am joining!

7.1.2022 12:31pm - (Replying to @kareem_carr) Great! But didn't they rule last year that machine learning is statistics on steroids?

7.1.2022 2:16am - (Replying to @elderofziyon) We call it "The Tectonic shift of John Kerry."

7.1.2022 1:06am - Replying to @elderofziyon We call it "The Tectonic shift of John Kerry."

6.30.2022 4:51pm - As a holder of a dual citizenship, Israeli+American, I am proud to introduce readers to the new Prime Minister of Israel, Yair Lapid, a man of vision and integrity, who just took office today, and is being congratulated by my President Joe Biden. To their partnership!- LeChayim!

6.30.2022 4:22pm - (Replying to @AvrahamAdler @mishtal and @benandjerry) The malice comes from Anuradha Mittal. But @benandjerry remind me of so many Jewish American uncles, who just want to please everyone (including Mittal) as long as you dont ask them to study more carefully what's going on in ME, or what the Palestinians say about Israel's future.

6.30.2022 1:44p - (Replying to @mishtal) You are too harsh on innocent boycotters like @benandjerrys who simply wish to signal virtue, peace, love and ice-cream and, not knowing any better, picked up the most populist slogan on the internet, "occupation of Palestinian land", and made it their banner of thoughtlessness.

6.30.2022 5:40am - (Replying to @LekhtNaya) And I want to be an "oppressor" - teach me!

6.30.2022 2:16am - (Replying to @PMinervini and @kerstingAIML) I would ask it more provocatively: I have a large model that someone found on the pavement outside. We know nothing about it except that it is "large", in fact huge, humongous. Is it a "Foundation Model" in Stanford sense"? In anyone's sense?

6.30.2022 1:26am - (Replying to @kerstingAIML) A calculus of intuitive causation would be great to have. Any aspirations as to its eventual structure? Here is an intuitive causal inference: If a drug helps men and helps women it must help people. Can "Foundation Models" reach this conclusion?

6.30.2022 1:14am - (Replying to @kerstingAIML) What is LLM? Couldn't find it in your paper, nor in Wikipedia.

6.30.2022 12:58am - Orwellian. But I don't believe President Pollack, a no-nonsense lady, would have approved of such stupidity. It tells more of how Wokism works: Some looney complains, a clerk gets scared of being accused of insensitivity, another hurries to his protection -- soon Lincoln is gone.

6.29.2022 11:39pm - My list of recent articles on causal inference: One paper catching my attention is "Can Foundation Models Talk Causality" a topic discussed on our platform which made me wonder: How can one ask "Can X do Y?" when X is undefined?

6.29.2022 10:56pm - @MESA_1966 decision to boycott #Israel academic institutions amounts to stripping the Middle East from the one country that has ushered the region to modernity-- in art, science, business and democracy -- thus committing the region to another century of hopelessness.

6.29.2022 4:10pm - This betrayal thrives on disinformation: Donors to "Israel Studies Centers" are kept ignorant of what leaders of those Centers teach or advocate. I therefore link here to the original statement, to help current & potential donors identify hijacked Centers.

6.29.2022 1:00pm - An amazing building! They mushroom on so many campuses. Naughty me, I once called them "Gothic Cathedrals for a handcuffed God." I still do, while hoping that the ratio of "science" to "data" would justify the hype.

6.29.2022 8:14am - (1/2) I have signed this petition, and I urge other readers to do so. However, adopting the IHRA definition of antisemitism is no replacement for pro-actively nourishing a Zionist-friendly culture on CUNY's campus. First, by explicitly welcoming Zionist students and faculty for
6.29.2022 8:14am - their unique contributions to campus life and, second, by declaring anti-Zionism incompatible with the educational goals, the principles of diversity and the norms of discourse the administration is striving to establish.

6.29.2022 4:27am - (Replying to @stephensenn and @snoble) Agree, and promise to reciprocate. Moreover, no need to compress "everything", just TWO different ways of doing RCT would do, and if it takes 3 x 280 character, so be it.

6.29.2022 4:09am - Congratulations to our colleague @Adi_Schwartz , co-author of "The War of Return" (with @EinatWilf), on receiving a doctorate degree from Bar Illan University.

6.29.2022 4:03am - (Replying to @Ostrov_A) What is there for a journalist to cover in Israel when you know in advance what you are about to see?

6.29.2022 3:34am - (Replying to @artistexyz and @AleksanderMolak) I see a "formulae" there, but unless I see the assumptions that lead to that formulae I cannot call it a "definition". I understand the assumptions that go into the sequential backdoor (as in Pearl and Robins 1995), but not the ones in WhatIf page 261, my weakness.

6.29.2022 3:27am - (Replying to @DrBobGoldberg) When and how did she make it clear? Has anyone listened to her? Is she on Twitter?

6.29.2022 3:08am - A clear video description of the Zionophobic culture permitted to incubate @CUNY . Hoping the City Hall protest tomorrow ends with actionable items, not merely protest. First item: apology for inviting a speaker who criminalized the identity of thousands of students and faculty.

6.29.2022 1:44am - More on dynamic systems. The best I've seen so far is: Bongers, Stephan, Tineke Blom, and Joris M. Mooij (2022). “Causal Modeling of Dynamical Systems”. Preprint. Available at arXiv:1803.08784v3 [http://cs.AI]. More in Bongers PhD dissertation, forthcoming.

6.29.2022 1:06am - (Replying to @artistexyz and @AleksanderMolak) I confess that I've never understood what g-formula is. Is it any truncated product of conditional probabilities? Or one that emanates from a specific graph (as in the do-operator). If the latter, how can you define it w/o a graph or an SCM? This I havn't been able to figure out.

6.28.2022 11:32pm - Losing your mind? Why? Fatima Salah and her linguistic fertility is just another example of Western scholars refusing to treat Palestinian mentality as an epiphenomenon deserving its own logic and psychology. An elephant in the room of rejectionism:

6.28.2022 10:22pm - (1/3) This is a gross misunderstanding of Palestinian rejectionism. Not a single refugee existed in 1936 (when I was born), no "occupation" nor "ethnic cleansing", yet rejectionism was at its highest. "Not even the size of a postage stamp" screamed Mufti Amin al-Husseini, when
6.28.2022 10:22pm - (2/3) 9 million Jews (including my grandparents) were trapped in Europe, seeking refuge. Not a single Arab refugee existed in 1945-7, when Holocaust survivors lingered in DP camp, seeking a home, and Arab rejectionism said NO to partition. Arab rejectionism is DNA deep, as
6.28.2022 10:22pm - (3/3) articulated clearly by every Arab historian (eg Rashid Khalidi) but ignored by Western analysts, who refuse to accept the absence of a quick "fix" they can engineer after taking a class in "conflict resolution". It remains the longest ignored elephant in the room of history.

6.28.2022 5:36pm - (Replying to @razingarizona @Vandalay_Inc and 3 others) I see nothing wrong with a 'Two-State Solution Act" if only it correctly identifies the main obstacle to such solution, and conditions every policy item on removing that obstacle - Palestinian rejectionism.

6.28.2022 5:23pm - (Replying to @blakeflayton) They blocked you because you represent the thing they fear most: A secured "Pro-Israel Pro-Peace" ideology rooted in the understanding that the latter follows organically from the former, and that it is wrong therefore to trumpet the latter as if it doesn't.

6.28.2022 2:56pm - (Replying to @BrettWilkinsSF @impsardo and 2 others) An interesting new logic. Israel is guiltier because she is hated more viciously, by more creative gang of predators, able to stockpile a richer arsenal of populist slogans, fabricated more craftily for Western minds to fall for. The more battered, the guiltier.

6.28.2022 9:00am - (Replying to @StaceyEBurke @Vandalay_Inc and 2 others) This may be the reason why JStreet, as well as @NewIsraelFund , spend their time attacking other Jewish organizations - impressing their donors with some action and uniqueness. I once wrote an oped about the Jews of Discomfort: Still relevant today.

6.28.2022 7:37am - Samantha, those who harassed you expect you to beg for protection. Try to address them as "Zionophobic racists" and see how they lose balance. Example: Zionophobes are shocked when their ideology is examined as a moral deformity, by facts & principles.

6.28.2022 6:57am - (Replying to @AleksanderMolak) The closest I got to dynamic systems was in a paper with James Robins ". "Probabilistic Evaluation of Sequential Plans" which computes the effect of time varying actions, in discrete times, each depending on the outcomes of its predecessors.

6.28.2022 4:33am - Thoughtful analysis of the Boston Mapping Project, especially on why Jewish groups have a hard time naming the ideological roots of progressive antisemitism.

6.28.2022 1:30am - (Replying to @GadSaad and @KamalaHarris) Finally, someone surpassing @AOC understanding of the Middle East.

6.28.2022 1:05am - (Replying to @stephensenn and @snoble) It's not a general rule. It's specific to this particular problem which we have cycled dozens of times, with zero progress, despite my honest attempts to penetrate the verbal curtains of RCT's discourse. It should also be helpful for trialists to see their words cast in 2 Eqs.

6.28.2022 12:53am - (Replying to @Martin_Kramer @MESA_1966 and @DovWaxman) Hard to believe my letter pushed UCLA to rescind its membership. I assume there were stronger oppressive and colonial forces behind the process, and no trumpets, to avoid backlash. That should only encourage Israel Studies Centers at other universities to try.

6.28.2022 12:14am - Illuminating article and fascinating demonstration

6.27.2022 11:50am - I became friendly with the Presbyterian Church when they asked me to go to Alabama to debate BDS cronies. Today they sent me an interesting video which I woul