Statistical Physics for Ambitious Interpretability

As part of an ongoing collaboration, this workshop brings together researchers from across disciplines to discuss physics-inspired approaches to understanding AI systems, including data structure, learning theory, and tools to interpret learned representations. Our goal is to coordinate on recent progress and seed new collaborations. Expect seminar-style talks, collaborative discussion sessions and tutorials, and plenty of time for focused discussions.

Registration is currently closed. Contact Claudine Lim (claudine AT princint DOT ai) or Lauren Greenspan (lauren AT princint DOT ai) for more information.