Reinforcement Learning from Human Feedback: Progress and Challenges
EECS Colloquium
Wednesday, April 19, 2023
310 Sutardja Dai Hall (Banatao Auditorium)
5:00 – 6:00 pm
John Schulman
Co-founder, OpenAI
Biography
John Schulman received a Ph.D. from Berkeley EECS in 2016, advised by Pieter Abbeel. He now leads a team working on ChatGPT and RL from Human Feedback at OpenAI, where he was a co-founder. His previous work includes foundational algorithms of deep RL (TRPO, PPO), generalization in RL (ProcGen), mathematical reasoning by language models (GSM8K), combining RL with retrieval (WebGPT) and studying scaling laws of RL and alignment. In his free time, he enjoys running, jazz piano, and raising chickens.