State Alignment-based Imitation Learning

Fangchen Liu, Zhan Ling, Tongzhou Mu +1

Consider an imitation learning problem that the imitator and the expert have different dynamics models. Most of the current imitation learning methods fail because they focus on im…