Weizmann Logo
ECCC
Electronic Colloquium on Computational Complexity

Under the auspices of the Computational Complexity Foundation (CCF)

Login | Register | Classic Style



REPORTS > DETAIL:

Paper:

TR18-002 | 31st December 2017 23:31

Which Distribution Distances are Sublinearly Testable?

RSS-Feed




TR18-002
Authors: Constantinos Daskalakis, Gautam Kamath, John Wright
Publication: 2nd January 2018 18:46
Downloads: 965
Keywords: 


Abstract:

Given samples from an unknown distribution $p$ and a description of a distribution $q$, are $p$ and $q$ close or far? This question of "identity testing" has received significant attention in the case of testing whether $p$ and $q$ are equal or far in total variation distance. However, in recent work, the following questions have been been critical to solving problems at the frontiers of distribution testing:
-Alternative Distances: Can we test whether $p$ and $q$ are far in other distances, say Hellinger?
-Tolerance: Can we test when $p$ and $q$ are close, rather than equal? And if so, close in which distances?
Motivated by these questions, we characterize the complexity of distribution testing under a variety of distances, including total variation, $\ell_2$, Hellinger, Kullback-Leibler, and $\chi^2$. For each pair of distances $d_1$ and $d_2$, we study the complexity of testing if $p$ and $q$ are close in $d_1$ versus far in $d_2$, with a focus on identifying which problems allow strongly sublinear testers (i.e., those with complexity $O(n^{1 - \gamma})$ for some $\gamma > 0$ where $n$ is the size of the support of the distributions $p$ and $q$). We provide matching upper and lower bounds for each case. We also study these questions in the case where we only have samples from $q$ (equivalence testing), showing qualitative differences from identity testing in terms of when tolerance can be achieved. Our algorithms fall into the classical paradigm of $\chi^2$-statistics, but require crucial changes to handle the challenges introduced by each distance we consider. Finally, we survey other recent results in an attempt to serve as a reference for the complexity of various distribution testing problems.



ISSN 1433-8092 | Imprint