Numerical Stability of the Nyström Method

Fuente: arXiv
Guardado en:
Detalles Bibliográficos
Autores principales: Bucci, Alberto, Nakatsukasa, Yuji, Park, Taejun
Formato: Preprint
Publicado: 2025
Materias:
Acceso en línea:
Etiquetas: Agregar Etiqueta
Sin Etiquetas, Sea el primero en etiquetar este registro!
_version_ 1866915644456304640
author Bucci, Alberto
Nakatsukasa, Yuji
Park, Taejun
author_facet Bucci, Alberto
Nakatsukasa, Yuji
Park, Taejun
contents The Nyström method is a widely used technique for improving the scalability of kernel-based algorithms, including kernel ridge regression, spectral clustering, and Gaussian processes. Despite its popularity, the numerical stability of the method has remained largely an unresolved problem. In particular, the pseudo-inversion of the submatrix involved in the Nyström method may pose stability issues as the submatrix is likely to be ill-conditioned, resulting in numerically poor approximation. In this work, we establish conditions under which the Nyström method is numerically stable. We show that stability can be achieved through an appropriate choice of column subsets and a careful implementation of the pseudoinverse. Our results and experiments provide theoretical justification and practical guidance for the stable application of the Nyström method in large-scale kernel computations.
format Preprint
id arxiv_https___arxiv_org_abs_2511_15583
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Numerical Stability of the Nyström Method
Bucci, Alberto
Nakatsukasa, Yuji
Park, Taejun
Numerical Analysis
15A23, 65F55, 65G50
The Nyström method is a widely used technique for improving the scalability of kernel-based algorithms, including kernel ridge regression, spectral clustering, and Gaussian processes. Despite its popularity, the numerical stability of the method has remained largely an unresolved problem. In particular, the pseudo-inversion of the submatrix involved in the Nyström method may pose stability issues as the submatrix is likely to be ill-conditioned, resulting in numerically poor approximation. In this work, we establish conditions under which the Nyström method is numerically stable. We show that stability can be achieved through an appropriate choice of column subsets and a careful implementation of the pseudoinverse. Our results and experiments provide theoretical justification and practical guidance for the stable application of the Nyström method in large-scale kernel computations.
title Numerical Stability of the Nyström Method
topic Numerical Analysis
15A23, 65F55, 65G50
url https://arxiv.org/abs/2511.15583