Review for NeurIPS paper: DisCor: Corrective Feedback in Reinforcement Learning via Distribution Correction