Anca Dragan

Anca Dragan: Human-Robot Interaction and Reward Engineering | Lex Fridman Podcast #81

Anca Dragan, Lex Fridman

Berkeley professor Anka Jagan explores the complexities of human-robot interaction by challenging standard prediction models with a bounded rationality framework that infers human intent through intuitive physics and active information gathering. Her research applies these principles to autonomous driving, where robots must negotiate ambiguous human behaviors and adapt to "leaked" environmental reward signals rather than relying on fixed objectives. Jagan advocates for treating robots as collaborative partners in a dynamic, game-theoretic system that continuously learns from human corrections and physical interactions to navigate unpredictable edge cases safely.