About

Hi, there! I’m Yifei, a second-year Ph.D. student at Berkeley AI Research (BAIR). I’m advised by prof. Sergey Levine. I am also a visiting researcher at FAIR, Meta.

These days, I am most interested in interactive decision-making/reinforcement learning with LLM/VLMs (i.e. LLM/VLM agent). In particular, I am interested in large-scale self-improvement where the agent improves itself through open-ended reinforcement learning without relying on any human supervision.