NeurIPS 2020

Faster Wasserstein Distance Estimation with the Sinkhorn Divergence

Meta Review

Most reviewers agreed that the paper provides novel and solid theoretical results on the properties of the Sinkhorn divergence for estimating the Wasserstein distance. Some reviewers have suggested some minor points that can be improved (eg plots of error, adding computational time analysis, ...) but overall, they are happy to have this paper at Neurips.