Efficient Distance Approximation for Structured High-Dimensional Distributions via Learning

Meta Review

While the technical delta from prior works in the paper is not very big, there are nice conceptual insights that should of broader interest to the community and generic enough so that they may be applicable to future work.