Testing Properties of Multiple Distributions with Few Samples

TL;DR

This paper introduces sample-efficient algorithms for testing properties of multiple distributions simultaneously, such as uniformity, identity, and closeness, even with limited samples per distribution.

Contribution

It presents the first sample-optimal testers for multiple distribution property testing under a new setting with few samples per distribution.

Findings

Sample-optimal testers for uniformity, identity, and closeness testing.

Effective testing with limited samples per distribution.

Assumption of a natural condition enables optimal sample complexity.

Abstract

We propose a new setting for testing properties of distributions while receiving samples from several distributions, but few samples per distribution. Given samples from $s$ distributions, $p_1, p_2, \ldots, p_s$, we design testers for the following problems: (1) Uniformity Testing: Testing whether all the $p_i$'s are uniform or $\epsilon$-far from being uniform in $\ell_1$-distance (2) Identity Testing: Testing whether all the $p_i$'s are equal to an explicitly given distribution $q$ or $\epsilon$-far from $q$ in $\ell_1$-distance, and (3) Closeness Testing: Testing whether all the $p_i$'s are equal to a distribution $q$ which we have sample access to, or $\epsilon$-far from $q$ in $\ell_1$-distance. By assuming an additional natural condition about the source distributions, we provide sample optimal testers for all of these problems.