Kai-Wei Chang

Professor, UCLA Computer Science

Co-director, UCLA DataX AI Technology Center

My group studies how AI systems can perceive the world through multimodal models, capture and organize knowledge through language, reason over evidence, follow constraints, and collaborate with people in high-stakes settings. We aim to build models and agents that are capable, reliable, scientifically ambitious, and grounded in human values. I have long-standing experience across academia and industry, connecting foundational research with real-world AI systems.

Google Scholar Publications Group CV

News

Apr 2026

Ctrl-R was selected as a spotlight at ICML 2026.

Feb 2026

I will give a tutorial on Knowledge Control at ACL 2026.

Research at a glance

Multimodal Foundation Models

We developed VisualBERT, one of the first vision-language models, and follow-ups, including GLIP and DesCo that can recognize objects through detailed language descriptions. Our recent work extends this line to robust video-language alignment and multimodal reasoning; for example, our 7B multimodal reasoning model outperforms GPT-4o on a visual math benchmark.

VisualBERT GLIP OpenVLThinker VideoCon VideoPhy2

AI Agents for Science

We build agentic systems that synthesize literature, reason over evidence, use tools, and support scientific discovery with human oversight. This includes compositional tool-use systems for scientific question answering and benchmarks such as MathVista that are widely used by frontier labs to evaluate visual mathematical reasoning.

Chameleon MathVista Embodied web agents expMath news

Reasoning in LLMs

We developed methods enabling LLMs to adhere to specified constraints and study commonsense, mathematical, logical, and structured reasoning capabilities. Recent projects include controllable generation, open reasoning data recipes, and trajectory-control methods for stronger reasoning models.

Constrained generation HoneyBee Ctrl-R

Trustworthy AI

We published pioneering work on aligning AI systems with human values and safety, focusing on fairness in generative AI, robustness, unlearning, and safety in LLMs and multimodal systems. Recent work includes selective unlearning for removing sensitive content and agentic safety reasoning. I also organize Trustworthy NLP workshops at ACL* since 2020.

Gender bias amplification SafeWorld Customized guardrails Unlearning

Highlighted papers and projects

Ctrl-R ICML 2026 Spotlight; learning structured reasoning via tractable trajectory control. MathVista Widely used benchmark for mathematical reasoning of foundation models in visual contexts. AI4Math Moonshot Laude Institute Moonshots grant to build AI systems as research partners for mathematics. Grounded Language-Image Pre-training (GLIP) CVPR 2022 Best Paper Finalist; grounded language-image pre-training for open-vocabulary recognition. CleanCLIP Best Paper at ICLR Trustworthy ML Workshop; mitigating data poisoning attacks in multimodal contrastive learning. Chameleon Plug-and-play compositional reasoning with LLMs for ScienceQA and TabMWP. Large Linear Classification KDD 2010 Best Paper; scalable learning when data cannot fit in memory. Men Also Like Shopping EMNLP 2017 Best Long Paper; corpus-level constraints for reducing gender bias amplification. Man is to Computer Programmer as Woman is to Homemaker? NeurIPS 2016; first paper showing and mitigating gender bias in word embeddings.