Talk Title: “AI Agents That Do What We Want: Progress and Open Challenges”
Speaker: Anca Dragan, Associate Professor of Electrical Engineering and Computer Sciences, UC Berkeley
Register To Attend | Watch Livestream on YouTube
Abstract: Researchers used to define objectives for artificial intelligence (AI) agents by hand, but with progress in optimization and reinforcement learning, it became obvious that it’s too difficult to think of everything ahead of time and write it down. Instead, these days the objective is viewed as a hidden part of the state on which researchers can receive feedback or observations from humans — how they act and react, how they compare options, what they say. In this talk, Dragan will discuss what this transition has achieved, what open challenges researchers still face and ideas for mitigating them. This lecture will go over applications in robotics and how the lessons there apply to virtual agents like large language models.
Speaker Bio: Anca Dragan is an associate professor in the Department of Electrical Engineering and Computer Sciences at UC Berkeley. She runs the InterACT Lab, where her team focuses on algorithms for human-robot interaction that move beyond the robot’s function in isolation, generate robot behavior that coordinates well with people and align with what they actually want the robot to do. The lab works across different applications, from assistive arms to quadrotors to autonomous cars, and draws from optimal control, game theory, reinforcement learning, Bayesian inference and cognitive science. Dragan also helped found and serves on the steering committee for the Berkeley AI Research (BAIR) Lab, and she is a co-principal investigator of the Center for Human-Compatible AI. She has been honored with the Sloan Fellowship, MIT Innovators Under 35 (TR35), the Okawa Prize, an NSF CAREER award and the Presidential Early Career Awards for Scientists and Engineers (PECASE).
About the Talk: Co-hosted with the UC Berkeley College of Computing, Data Science, and Society and the UC Berkeley Artificial Intelligence Research (BAIR) Lab.
About the Series: CITRIS Research Exchange delivers fresh perspectives on information technology and society from distinguished academic, industry and civic leaders. Free and open to the public, these seminars feature leading voices on societal-scale research issues. Presentations take place on Wednesdays from noon to 1 p.m. PT. Have an idea for a great talk? Please feel free to suggest potential speakers for our series.
Sign up to receive the latest news and updates from CITRIS.