OptiDICE: Offline Policy Optimization via
             Stationary Distribution Correction Estimation

       Jongmin Lee 1 * Wonseok Jeon 2 3 * Byung-Jun Lee 4 Joelle Pineau 2 3 5 Kee-Eung Kim 1 6

            Abstract                and then to deploy the model with its parameter fixed when
                                we are satisfied with training. This offline training allows
   We consider the offline reinforcement learning
                                us to address various operationa ...