This seminar studies statistical representations of relational data, including Markov logic networks, relational Bayesian networks, and probabilistic databases. We will discuss their probabilistic inference problems, parameter, and structure learning algorithms, and representation power. The course also covers more general probabilistic programming languages and systems, which allow for more expressive statistical modeling.
This course requires advanced doctorate-level knowledge of artificial intelligence, machine learning, and probabilistic graphical models in particular. I will ask you to prove you have the right background to take this course before enrolling.
Grading will be based on Attendance and Participation (20%), Paper Presentation (35%) and Project (45%).
The course will consist of instructor-led lectures on the foundational techniques, followed by student presentations of research papers, and finally project presentations.
This course will have student presentations on various research directions in statistical relational learning, spanning multiple papers. Students will also be required to do a short workshop-quality project on statistical relational learning. You can work alone or in groups of two. For suggestions, see the list of research topics on CCLE