State Alignment-based Imitation Learning
This addresses a key limitation in imitation learning for robotics or control systems where dynamics mismatch is common, though it appears incremental as it builds on existing reinforcement learning frameworks.
The paper tackles imitation learning when the imitator and expert have different dynamics models by proposing a state alignment-based method that trains the imitator to follow expert state sequences, showing superiority in standard and mismatched dynamics settings.
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 imitating actions. We propose a novel state alignment-based imitation learning method to train the imitator to follow the state sequences in expert demonstrations as much as possible. The state alignment comes from both local and global perspectives and we combine them into a reinforcement learning framework by a regularized policy update objective. We show the superiority of our method on standard imitation learning settings and imitation learning settings where the expert and imitator have different dynamics models.