Review for NeurIPS paper: Simplify and Robustify Negative Sampling for Implicit Collaborative Filtering

The initial reviews were mixed for this paper. However, during the discussion, a certain consensus emerged regarding the value of this contribution. In particular, the reviews agree that the proposed method is simple and effective. In the proposed study, the method seems to outperform (by a small margin) others consistently. The main negative is the high computational and memory cost of the approach.