Yixiu Mao

offline reinforcement learning (4) out-of-distribution action (3) extrapolation error (3) policy learning (2) dynamic programming (1) policy iteration (1) action selection (1) policy improvement (1) behavior policy (1) value overestimation (1) in-sample learning (1) mild generalization (1) out-of-distribution state (1) credit assignment (1) episodic reinforcement learning (1) trust region policy optimization (1) large language model (1) reward redistribution (1) support constraint (1) latent reward (1)