Thank you for your submission to NeurIPS. Although there is still some disagreement amongst the reviewers (specifically on how well the proposed method compares to GeoCert in practice), the general consensus is also that the proposed approach presents an interesting and valuable contribution to the class of exact verification methods for deep networks. Some additional discussion to competing approaches, mentioned the rebuttal, would help to strengthen the paper further, but overall the majority feeling is that this is a very nice contribution to the topic.