Hybrid Transfer Reinforcement Learning: Provable Sample Efficiency from Shifted-Dynamics Data
This work addresses sample efficiency in RL for practitioners by leveraging offline data, but it is incremental as it builds on existing transfer learning methods with a focus on provable guarantees.
The paper tackles the problem of using historical data with shifted dynamics to improve sample efficiency in reinforcement learning, showing that without shift information such data does not help, but with prior knowledge, their HySRL algorithm achieves better sample complexity and outperforms online RL baselines in experiments.
Online Reinforcement learning (RL) typically requires high-stakes online interaction data to learn a policy for a target task. This prompts interest in leveraging historical data to improve sample efficiency. The historical data may come from outdated or related source environments with different dynamics. It remains unclear how to effectively use such data in the target task to provably enhance learning and sample efficiency. To address this, we propose a hybrid transfer RL (HTRL) setting, where an agent learns in a target environment while accessing offline data from a source environment with shifted dynamics. We show that -- without information on the dynamics shift -- general shifted-dynamics data, even with subtle shifts, does not reduce sample complexity in the target environment. However, with prior information on the degree of the dynamics shift, we design HySRL, a transfer algorithm that achieves problem-dependent sample complexity and outperforms pure online RL. Finally, our experimental results demonstrate that HySRL surpasses state-of-the-art online RL baseline.