Unbiased Learning with State-Conditioned Rewards in Adversarial Imitation Learning
Dong-Sig Han, Hyunseo Kim, Hyundo Lee, Je-Hwan Ryu, Byoung‐Tak Zhang · 2021
Adversarial imitation learning has emerged as a general and scalable framework for automatic reward acquisition. However, we point out that previous methods commonly exploited occupancy-dependent reward learning formulation—which hinders the reconstruction of optimal decision as an energy-based model. Despite the theoretical justification, the occupancy measures tend to cause issues in practice because of high variance and low vulnerability to domain shifts. Another reported problem is termination biases induced by provided rewarding and regularization schemes around terminal states. In order to deal with these issues, this work presents a novel algorithm called causal adversarial inverse reinforcement learning. Our formulation draws a strong connection between adversarial learning and energy-based reinforcement learning; thus, the architecture is capable of recovering a reward function that induces a multi-modal policy. In experiments, we demonstrate that our approach outperforms prior methods in challenging continuous control tasks, even under significant variation in the environments.