Offline Q-learning on Diverse Multi-Task Data Both Scales And Generalizes.

Aviral Kumar, Rishabh Agarwal, Xinyang Geng, George Tucker, Sergey Levine

Browse the full ICLR paper archive.