NeurIPS 2020

Auto Learning Attention

Meta Review

All referees agree that the paper introduces an interesting and valuable idea but three referees have indicated reject because of missing experimental comparisons. The rebuttal includes all the experiments requested by the reviewers, providing additional insights and more compelling evidence on the benefits of the proposed approach. The referees have no further comments. After discussing with the SAC, we believe that the contribution of this paper is relevant and the rebuttal has properly addressed the concerns raised by the referees. Hence, I recommend acceptance. However, please consider revising your paper to include the results shared in the rebuttal and the reviewers' remarks to improve the clarity of the paper.