EMaQ Expected-Max Q-Learning Operator for Simple Yet Effective Offline and Online RL

EMaQ:Expected-MaxQ-LearningOperatorforSimpleYetEffectiveOfflineandOnlineRLSeyedKamyarSeyedGhasemipour12DaleSchuurmans3ShixiangShaneGu3Abstract1.IntroductionOff-policyreinforcementlearning(RL)holdst...