Embrace rejection: Kernel matrix approximation by accelerated randomly pivoted Cholesky

Fuente: arXiv
Enregistré dans:
Détails bibliographiques
Auteurs principaux: Epperly, Ethan N., Tropp, Joel A., Webber, Robert J.
Format: Preprint
Publié: 2024
Sujets:
Accès en ligne:
Tags: Ajouter un tag
Pas de tags, Soyez le premier à ajouter un tag!
_version_ 1866913779392970752
author Epperly, Ethan N.
Tropp, Joel A.
Webber, Robert J.
author_facet Epperly, Ethan N.
Tropp, Joel A.
Webber, Robert J.
contents Randomly pivoted Cholesky (RPCholesky) is an algorithm for constructing a low-rank approximation of a positive-semidefinite matrix using a small number of columns. This paper develops an accelerated version of RPCholesky that employs block matrix computations and rejection sampling to efficiently simulate the execution of the original algorithm. For the task of approximating a kernel matrix, the accelerated algorithm can run over $40\times$ faster. The paper contains implementation details, theoretical guarantees, experiments on benchmark data sets, and an application to computational chemistry.
format Preprint
id arxiv_https___arxiv_org_abs_2410_03969
institution arXiv
publishDate 2024
record_format arxiv
spellingShingle Embrace rejection: Kernel matrix approximation by accelerated randomly pivoted Cholesky
Epperly, Ethan N.
Tropp, Joel A.
Webber, Robert J.
Numerical Analysis
Computation
Machine Learning
65F55, 65C99, 68T05
Randomly pivoted Cholesky (RPCholesky) is an algorithm for constructing a low-rank approximation of a positive-semidefinite matrix using a small number of columns. This paper develops an accelerated version of RPCholesky that employs block matrix computations and rejection sampling to efficiently simulate the execution of the original algorithm. For the task of approximating a kernel matrix, the accelerated algorithm can run over $40\times$ faster. The paper contains implementation details, theoretical guarantees, experiments on benchmark data sets, and an application to computational chemistry.
title Embrace rejection: Kernel matrix approximation by accelerated randomly pivoted Cholesky
topic Numerical Analysis
Computation
Machine Learning
65F55, 65C99, 68T05
url https://arxiv.org/abs/2410.03969