NeurIPS 2020

Natural Graph Networks

Meta Review

While I think all reviewers and myself included appreciate the value of the work, there are a few lingering concerns around the clarity of the presentation and potentially slightly weak empirical evidence. In particular there are questions about comparison with additional work (as pointed out by L6). I think the value outweighs the negatives. However I urge the authors to incorporate as much as they can from the feedback that they got, and the details mention in the rebuttal, in order to maximize the impact this work can have within the community!