Precise and Dexterous Robotic Manipulation via Human-in-the-Loop Reinforcement Learning

Jianlan Luo, Charles Xu, Jeffrey Wu, Sergey Levine
Department of EECS, University of California, Berkeley
Paper Intro Video HIL-SERL Code Join our Discord

Policy Autonomous Rollouts

Policy Robustness

System Overview

Abstract Figure

Training Timelapse

Policy Evaluation (100 Trials)