Does block size matter in randomized block Krylov low-rank approximation?

Fuente: arXiv
Salvato in:
Dettagli Bibliografici
Autori principali: Chen, Tyler, Epperly, Ethan N., Meyer, Raphael A., Musco, Christopher, Rao, Akash
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866909859837902848
author Chen, Tyler
Epperly, Ethan N.
Meyer, Raphael A.
Musco, Christopher
Rao, Akash
author_facet Chen, Tyler
Epperly, Ethan N.
Meyer, Raphael A.
Musco, Christopher
Rao, Akash
contents We study the problem of computing a rank-$k$ approximation of a matrix using randomized block Krylov iteration. Prior work has shown that, for block size $b = 1$ or $b = k$, a $(1 + \varepsilon)$-factor approximation to the best rank-$k$ approximation can be obtained after $\tilde O(k/\sqrt{\varepsilon})$ matrix-vector products with the target matrix. On the other hand, when $b$ is between $1$ and $k$, the best known bound on the number of matrix-vector products scales with $b(k-b)$, which could be as large as $O(k^2)$. Nevertheless, in practice, the performance of block Krylov methods is often optimized by choosing a block size $1 \ll b \ll k$. We resolve this theory-practice gap by proving that randomized block Krylov iteration produces a $(1 + \varepsilon)$-factor approximate rank-$k$ approximation using $\tilde O(k/\sqrt{\varepsilon})$ matrix-vector products for any block size $1\le b\le k$. Our analysis relies on new bounds for the minimum singular value of a random block Krylov matrix, which may be of independent interest. Similar bounds are central to recent breakthroughs on faster algorithms for sparse linear systems [Peng & Vempala, SODA 2021; Nie, STOC 2022].
format Preprint
id arxiv_https___arxiv_org_abs_2508_06486
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Does block size matter in randomized block Krylov low-rank approximation?
Chen, Tyler
Epperly, Ethan N.
Meyer, Raphael A.
Musco, Christopher
Rao, Akash
Data Structures and Algorithms
Numerical Analysis
65F55, 65F15
G.1.3; F.2.1
We study the problem of computing a rank-$k$ approximation of a matrix using randomized block Krylov iteration. Prior work has shown that, for block size $b = 1$ or $b = k$, a $(1 + \varepsilon)$-factor approximation to the best rank-$k$ approximation can be obtained after $\tilde O(k/\sqrt{\varepsilon})$ matrix-vector products with the target matrix. On the other hand, when $b$ is between $1$ and $k$, the best known bound on the number of matrix-vector products scales with $b(k-b)$, which could be as large as $O(k^2)$. Nevertheless, in practice, the performance of block Krylov methods is often optimized by choosing a block size $1 \ll b \ll k$. We resolve this theory-practice gap by proving that randomized block Krylov iteration produces a $(1 + \varepsilon)$-factor approximate rank-$k$ approximation using $\tilde O(k/\sqrt{\varepsilon})$ matrix-vector products for any block size $1\le b\le k$. Our analysis relies on new bounds for the minimum singular value of a random block Krylov matrix, which may be of independent interest. Similar bounds are central to recent breakthroughs on faster algorithms for sparse linear systems [Peng & Vempala, SODA 2021; Nie, STOC 2022].
title Does block size matter in randomized block Krylov low-rank approximation?
topic Data Structures and Algorithms
Numerical Analysis
65F55, 65F15
G.1.3; F.2.1
url https://arxiv.org/abs/2508.06486