NeurIPS 2020

A Simple Language Model for Task-Oriented Dialogue

Meta Review

All reviewers find this work quite strong both in terms of approach and results, and reviewers applaud that the work has proved robustly reproducible. One important point is that several contemporaneous papers share some commonalities with this submission. We agree that they were published less than a month before the deadline and should therefore be considered contemporaneous; however it would have been much better scientific practice to include the discussion of these works in the submission, if the authors were aware of them -- regardless on when the authors put their initial submission on arxiv. The discussion that situates the submission in the context of these other works in the authors response is enlightening and interesting, and should definitely be in the final version. Conditioned on this being the case, we are happy to accept the paper. Authors should also use the additional page in the final version to incorporate the many excellent suggestions from reviewers to improve the paper.