2794f6a20ee0685f4006210f40799acd-AuthorFeedback.pdf

We thank all the reviewers for their constructive comments. We will take these suggestions in later version. Y es, we agree with you that when two finite networks are sampled from a common continuous CNN, their standard We adopt widely used statistics in hypothesis test as metrics to measure such difference. Moreover, we'd like to point out that the standard statistics mean and variance are not enough to differentiate two The proposed method just does a change of variable where the new variables hide the non-convexity of the system. The reviewer might have missed some results in our appendix.