Imitation Learning: An Introduction 模仿学习在机器人学习 (Robot Learning)中扮演了比较重要的角色。这其实在之前的paper reading中已经涉及过了: · Imitation through Neural Network is not enough, while via Principle-based Optimization as safety guarantees, works… Imitation Is Not Enough: Robustifying Imitation with Reinforcement Learning for Challenging Driving Scenarios（RL解决IL的OOD问题） 3 IL和RL在策略梯度层面融合 Deep imitative reinforcement … · Diffusion Policy Policy Optimization 先用 imitation learning 去initialize一个diffusion policy，然后用rl loss 去finetune，发现有很强的structured exploration · 论文简介 论文题“Context-aware Graph Inference and Generative Adversarial Imitation Learning for Object-goal Navigation in Unfamiliar Environment”在机器人领域TOP期刊 IEEE …
