Robust Inverse Reinforcement Learning under Transition Dynamics Mismatch
Machine Learning
2021-12-01 v4 Machine Learning
Abstract
We study the inverse reinforcement learning (IRL) problem under a transition dynamics mismatch between the expert and the learner. Specifically, we consider the Maximum Causal Entropy (MCE) IRL learner model and provide a tight upper bound on the learner's performance degradation based on the -distance between the transition dynamics of the expert and the learner. Leveraging insights from the Robust RL literature, we propose a robust MCE IRL algorithm, which is a principled approach to help with this mismatch. Finally, we empirically demonstrate the stable performance of our algorithm compared to the standard MCE IRL algorithm under transition dynamics mismatches in both finite and continuous MDP problems.
Cite
@article{arxiv.2007.01174,
title = {Robust Inverse Reinforcement Learning under Transition Dynamics Mismatch},
author = {Luca Viano and Yu-Ting Huang and Parameswaran Kamalaruban and Adrian Weller and Volkan Cevher},
journal= {arXiv preprint arXiv:2007.01174},
year = {2021}
}