Near-optimal Offline Reinforcement Learning with Linear Representation: Leveraging Variance Information with Pessimism.

Ming Yin, Yaqi Duan, Mengdi Wang, Yu-Xiang Wang

Browse the full ICLR paper archive.