Fetching the paper…

Offline Reinforcement Learning Under Value and Density-Ratio Realizability: The Power of Gaps · Around