Distributional Off-Policy Evaluation in Reinforcement Learning
强化学习中的分布型离策略评估
In the literature of reinforcement learning (RL), off-policy evaluation is mainly focused on estimating a value of a target policy given the pre-collected data generated by some behavior policy. Motivated by the recent success of distributional RL in many practical applications, we study the distributional off-policy evaluation problem in the batch setting when the reward is multi-variate. We propose an offline Wasserstein-based approach to simultaneously estimate the joint…