Hybrid Reinforcement Learning from Offline Observation Alone

arXiv:2406.07253v1 Announce Type: new
Abstract: 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" …

abstract access action agent arxiv cs.lg data datasets hybrid information interactive observation offline reinforcement reinforcement learning research state transition type while