On Pathologies in KL-Regularized Reinforcement Learning from Expert Demonstrations
This addresses stability and efficiency issues in reinforcement learning for robotics, though it is incremental as it builds on existing KL-regularized methods.
The paper identifies pathological training dynamics in KL-regularized reinforcement learning with expert demonstrations, leading to slow and suboptimal learning, and proposes a remedy using non-parametric policies that outperforms state-of-the-art methods on locomotion and manipulation tasks.
KL-regularized reinforcement learning from expert demonstrations has proved successful in improving the sample efficiency of deep reinforcement learning algorithms, allowing them to be applied to challenging physical real-world tasks. However, we show that KL-regularized reinforcement learning with behavioral reference policies derived from expert demonstrations can suffer from pathological training dynamics that can lead to slow, unstable, and suboptimal online learning. We show empirically that the pathology occurs for commonly chosen behavioral policy classes and demonstrate its impact on sample efficiency and online policy performance. Finally, we show that the pathology can be remedied by non-parametric behavioral reference policies and that this allows KL-regularized reinforcement learning to significantly outperform state-of-the-art approaches on a variety of challenging locomotion and dexterous hand manipulation tasks.