A permutation-free kernel two-sample testDownload PDF

Published: 31 Oct 2022, Last Modified: 14 Jan 2023NeurIPS 2022 AcceptReaders: Everyone
Keywords: two-sample testing, kernel-mmd, permutation-free tests, U-statistics
TL;DR: A new kernel-mmd two-sample test that is permutation-free, is consistent against a fixed alternative and has minimax rate-optimal power against local alternatives.
Abstract: The kernel Maximum Mean Discrepancy~(MMD) is a popular multivariate distance metric between distributions. The usual kernel-MMD test statistic (for two-sample testing) is a degenerate U-statistic under the null, and thus it has an intractable limiting null distribution. Hence, the standard approach for designing a level-$(1-\alpha)$ two-sample test using this statistic involves selecting the rejection threshold as the $(1-\alpha)$-quantile of the permutation distribution. The resulting nonparametric test has finite-sample validity but suffers from large computational cost, since the test statistic must be recomputed for every permutation. We propose the cross-MMD, a new quadratic time MMD test statistic based on sample-splitting and studentization. We prove that under mild assumptions, it has a standard normal limiting distribution under the null. Importantly, we also show that the resulting test is consistent against any fixed alternative, and when using the Gaussian kernel, it has minimax rate-optimal power against local alternatives. For large sample-sizes, our new cross-MMD provides a significant speedup over the MMD, for only a slight loss in power.
Supplementary Material: zip
13 Replies