Quantification before Selection: Active Dynamics Preference for Robust Reinforcement Learning. (arXiv:2209.11596v2 [cs.LG] UPDATED)

Training a robust policy is critical for policy deployment in real-world
systems or dealing with unknown dynamics mismatch in different dynamic systems.
Domain Randomization~(DR) is a simple and elegant approach that trains a
conservative policy to counter different dynamic systems without expert
knowledge about the target system parameters. However, existing works reveal
that the policy trained through DR tends to be over-conservative and performs
poorly in target domains. Our key insight is that dynamic systems with
different parameters provide different …

arxiv dynamics quantification reinforcement reinforcement learning