SelfBC: Self Behavior Cloning for Offline Reinforcement Learning

TL;DR

SelfBC introduces a dynamic self-constraint mechanism in offline reinforcement learning that enables learning more effective policies by adaptively restricting the policy based on an exponential moving average of past policies, outperforming existing methods.

Contribution

The paper proposes a novel dynamic policy constraint using an exponential moving average of policies, improving over static constraints in offline RL.

Findings

Achieves state-of-the-art performance on D4RL MuJoCo tasks.

Demonstrates nearly monotonic improvement of the reference policy.

Effectively balances policy flexibility and stability in offline RL.

Abstract

Policy constraint methods in offline reinforcement learning employ additional regularization techniques to constrain the discrepancy between the learned policy and the offline dataset. However, these methods tend to result in overly conservative policies that resemble the behavior policy, thus limiting their performance. We investigate this limitation and attribute it to the static nature of traditional constraints. In this paper, we propose a novel dynamic policy constraint that restricts the learned policy on the samples generated by the exponential moving average of previously learned policies. By integrating this self-constraint mechanism into off-policy methods, our method facilitates the learning of non-conservative policies while avoiding policy collapse in the offline setting. Theoretical results show that our approach results in a nearly monotonically improved reference policy. Extensive experiments on the D4RL MuJoCo domain demonstrate that our proposed method achieves state-of-the-art performance among the policy constraint methods.