Google, IBM & Meta Certificates — All 10,000+ Courses at 40% Off
One annual plan covers every course and certificate on Coursera. 40% off for a limited time.
Get Full Access
This seminar presentation explores how AI agents can learn through dynamic interactions rather than static datasets. Delivered by Dr. Hao Zhu from Stanford University on April 17, 2025, at USC Information Sciences Institute, discover three foundational approaches for building more human-like learning systems: exploring the internet for open-ended learning experiences, internalizing social norms through crowd-sourced annotation, and developing automated evaluation metrics based on human feedback. Learn how these complementary methods could help create AI agents that more effectively adapt to and integrate into our social world. Dr. Zhu, a postdoctoral researcher at Stanford who completed his PhD at CMU, brings expertise in AI agents, human-agent interaction, robotics, and embodied AI to this discussion of how artificial agents might better understand human social and embodied cognition.
