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, $p_1, p_2, łdots, p_s$, 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.
Description
[1911.07324] Testing Properties of Multiple Distributions with Few Samples
%0 Journal Article
%1 aliakbarpour2019testing
%A Aliakbarpour, Maryam
%A Silwal, Sandeep
%D 2019
%K distributed hypothesis-testing readings
%T Testing Properties of Multiple Distributions with Few Samples
%U http://arxiv.org/abs/1911.07324
%X 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, $p_1, p_2, łdots, p_s$, 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.
@article{aliakbarpour2019testing,
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.},
added-at = {2020-03-09T18:30:40.000+0100},
author = {Aliakbarpour, Maryam and Silwal, Sandeep},
biburl = {https://www.bibsonomy.org/bibtex/223ea18fd9be805717bd8ac7d6dd76602/kirk86},
description = {[1911.07324] Testing Properties of Multiple Distributions with Few Samples},
interhash = {b148a60ee99ca46e1bb49cd66ac958b7},
intrahash = {23ea18fd9be805717bd8ac7d6dd76602},
keywords = {distributed hypothesis-testing readings},
note = {cite arxiv:1911.07324Comment: ITCS 2020},
timestamp = {2020-03-09T18:30:40.000+0100},
title = {Testing Properties of Multiple Distributions with Few Samples},
url = {http://arxiv.org/abs/1911.07324},
year = 2019
}