@inproceedings{refinedvaluebasedofflinerlunder, title = {Offline Minimax Soft-Q-learning Under Realizability and Partial Coverage}, year = {2023}, booktitle = {NeurIPS 2023 11}, eprint = {2302.02392}, archivePrefix = {arXiv}, url = {https://arxiv.org/abs/2302.02392v2}, }