NeurIPS 2020

Sampling-Decomposable Generative Adversarial Recommender

Meta Review

This paper received overall positive scores. All the reviewers agree that the theoretical analysis on GAN-based recommender models is the main strength of the paper. One reviewer raised some concern regarding the metrics obtained by the baseline being too low. I suggest that the authors carefully check the evaluation procedure, as well as incorporate the feedback from the reviews in the revision.