Our paper “Towards Optimal Offline Reinforcement Learning” is available online.