Hybrid Reinforcement Learning from Offline Observation Alone
This addresses a practical challenge for reinforcement learning practitioners by enabling the use of more abundant observation-only offline data, though it is incremental as it builds on existing hybrid RL frameworks with specific assumptions.
The paper tackles the problem of hybrid reinforcement learning using offline datasets that contain only state observations, without action or reward information, and proposes the first algorithm that matches the performance of methods requiring stronger reset models under admissibility assumptions, with proof-of-concept experiments suggesting its effectiveness.
We consider the hybrid reinforcement learning setting where the agent has access to both offline data and online interactive access. While Reinforcement Learning (RL) research typically assumes offline data contains complete action, reward and transition information, datasets with only state information (also known as observation-only datasets) are more general, abundant and practical. This motivates our study of the hybrid RL with observation-only offline dataset framework. While the task of competing with the best policy "covered" by the offline data can be solved if a reset model of the environment is provided (i.e., one that can be reset to any state), we show evidence of hardness when only given the weaker trace model (i.e., one can only reset to the initial states and must produce full traces through the environment), without further assumption of admissibility of the offline data. Under the admissibility assumptions -- that the offline data could actually be produced by the policy class we consider -- we propose the first algorithm in the trace model setting that provably matches the performance of algorithms that leverage a reset model. We also perform proof-of-concept experiments that suggest the effectiveness of our algorithm in practice.