English

Tight Sampling Bounds for Eigenvalue Approximation

Data Structures and Algorithms 2024-11-06 v1

Abstract

We consider the problem of estimating the spectrum of a symmetric bounded entry (not necessarily PSD) matrix via entrywise sampling. This problem was introduced by [Bhattacharjee, Dexter, Drineas, Musco, Ray '22], where it was shown that one can obtain an ϵn\epsilon n additive approximation to all eigenvalues of AA by sampling a principal submatrix of dimension poly(logn)ϵ3\frac{\text{poly}(\log n)}{\epsilon^3}. We improve their analysis by showing that it suffices to sample a principal submatrix of dimension O~(1ϵ2)\tilde{O}(\frac{1}{\epsilon^2}) (with no dependence on nn). This matches known lower bounds and therefore resolves the sample complexity of this problem up to log1ϵ\log\frac{1}{\epsilon} factors. Using similar techniques, we give a tight O~(1ϵ2)\tilde{O}(\frac{1}{\epsilon^2}) bound for obtaining an additive ϵAF\epsilon\|A\|_F approximation to the spectrum of AA via squared row-norm sampling, improving on the previous best O~(1ϵ8)\tilde{O}(\frac{1}{\epsilon^{8}}) bound. We also address the problem of approximating the top eigenvector for a bounded entry, PSD matrix A.A. In particular, we show that sampling O(1ϵ)O(\frac{1}{\epsilon}) columns of AA suffices to produce a unit vector uu with uTAuλ1(A)ϵnu^T A u \geq \lambda_1(A) - \epsilon n. This matches what one could achieve via the sampling bound of [Musco, Musco'17] for the special case of approximating the top eigenvector, but does not require adaptivity. As additional applications, we observe that our sampling results can be used to design a faster eigenvalue estimation sketch for dense matrices resolving a question of [Swartworth, Woodruff'23], and can also be combined with [Musco, Musco'17] to achieve O(1/ϵ3)O(1/\epsilon^3) (adaptive) sample complexity for approximating the spectrum of a bounded entry PSD matrix to ϵn\epsilon n additive error.

Keywords

Cite

@article{arxiv.2411.03227,
  title  = {Tight Sampling Bounds for Eigenvalue Approximation},
  author = {William Swartworth and David P. Woodruff},
  journal= {arXiv preprint arXiv:2411.03227},
  year   = {2024}
}