On the Sample Complexity of Vanilla Model-Based Offline Reinforcement Learning with Dependent Samples

Open in new window