NeurIPS 2020

A Measure-Theoretic Approach to Kernel Conditional Mean Embeddings

Meta Review

Three referees support acceptance for the contributions, finding the novel theoretical formulation of the CME to be a valuable contribution to the community. Referee R3 indicated rejection; the rebuttal disputes R3's claim that the theoretical contribution is not sufficient. I discount R3 because no concrete reason was given for their assertion that the contribution is insufficient. I agree with the remaining reviewers that the contribution is valuable and of interest to the NeurIPS community. The authors promised several changes to be made in the camera ready version to clarify the results; please do implement these changes.