Robot Learning
How robots learn skills from data and experience, covering imitation learning, reinforcement learning, Sim2Real, and diffusion policy.
Contents:
- Robot Learning Overview — Learning paradigms: IL, RL, self-supervised
- Imitation Learning — BC, DAgger, ACT, GAIL
- RL for Robotics — Reward engineering, massively parallel training
- Sim2Real — Domain randomization, adaptation, teacher-student
- Teleoperation & Data Collection — ALOHA, UMI, GELLO, data scaling
- Diffusion Policy — Diffusion Policy, DP3, Consistency Policy
- Multi-task & Generalization — Cross-task generalization, few-shot adaptation