Testing Positive Semi-Definiteness via Random Submatrices
Abstract
We study the problem of testing whether a matrix with bounded entries () is positive semi-definite (PSD), or -far in Euclidean distance from the PSD cone, meaning that , where denotes that is PSD. Our main algorithmic contribution is a non-adaptive tester which distinguishes between these cases using only queries to the entries of . If instead of the Euclidean norm we considered the distance in spectral norm, we obtain the "-gap problem", where is either PSD or satisfies . For this related problem, we give a query tester, which we show is optimal up to factors. Our testers randomly sample a collection of principal submatrices and check whether these submatrices are PSD. Consequentially, our algorithms achieve one-sided error: whenever they output that is not PSD, they return a certificate that has negative eigenvalues. We complement our upper bound for PSD testing with Euclidean norm distance by giving a lower bound for any non-adaptive algorithm. Our lower bound construction is general, and can be used to derive lower bounds for a number of spectral testing problems. As an example of the applicability of our construction, we obtain a new sampling lower bound for testing the Schatten- norm with a gap, extending a result of Balcan, Li, Woodruff, and Zhang [SODA'19]. In addition, it yields new sampling lower bounds for estimating the Ky-Fan Norm, and the cost of the best rank- approximation.
Cite
@article{arxiv.2005.06441,
title = {Testing Positive Semi-Definiteness via Random Submatrices},
author = {Ainesh Bakshi and Nadiia Chepurko and Rajesh Jayaram},
journal= {arXiv preprint arXiv:2005.06441},
year = {2020}
}
Comments
Minor Edits, highlighted connection between \ell_\infty gap and spectral norm gap