NeurIPS 2020

SVGD as a kernelized Wasserstein gradient flow of the chi-squared divergence

Meta Review

The reviewers had some concerns about the empirical evaluation and the lack of discrete-time results, but agree that this paper would be a useful addition to NeurIPS. Please see the reviews (and your response) for ways to improve the final manuscript (especially in terms of clarifying the context and scope of this work).