Saved in:
Bibliographic Details
Main Authors: Cai, Jian-Feng, Xu, Zhiqiang, Xu, Zili
Format: Preprint
Published: 2023
Subjects:
Online Access:https://arxiv.org/abs/2303.07984
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866914632611921920
author Cai, Jian-Feng
Xu, Zhiqiang
Xu, Zili
author_facet Cai, Jian-Feng
Xu, Zhiqiang
Xu, Zili
contents This paper investigates the spectral norm version of the column subset selection problem. Given a matrix $\mathbf{A}\in\mathbb{R}^{n\times d}$ and a positive integer $k\leq\text{rank}(\mathbf{A})$, the objective is to select exactly $k$ columns of $\mathbf{A}$ that minimize the spectral norm of the residual matrix after projecting $\mathbf{A}$ onto the space spanned by the selected columns. We use the method of interlacing polynomials introduced by Marcus-Spielman-Srivastava to derive a new upper bound on the minimal approximation error. This new bound is asymptotically sharp when the matrix $\mathbf{A}\in\mathbb{R}^{n\times d}$ obeys a spectral power-law decay. The relevant expected characteristic polynomials can be written as an extension of the expected polynomial for the restricted invertibility problem, incorporating two extra variable substitution operators. Finally, we propose a deterministic polynomial-time algorithm that achieves this error bound up to a computational error.
format Preprint
id arxiv_https___arxiv_org_abs_2303_07984
institution arXiv
publishDate 2023
record_format arxiv
spellingShingle Interlacing Polynomial Method for the Column Subset Selection Problem
Cai, Jian-Feng
Xu, Zhiqiang
Xu, Zili
Data Structures and Algorithms
Functional Analysis
15A60, 90C27
This paper investigates the spectral norm version of the column subset selection problem. Given a matrix $\mathbf{A}\in\mathbb{R}^{n\times d}$ and a positive integer $k\leq\text{rank}(\mathbf{A})$, the objective is to select exactly $k$ columns of $\mathbf{A}$ that minimize the spectral norm of the residual matrix after projecting $\mathbf{A}$ onto the space spanned by the selected columns. We use the method of interlacing polynomials introduced by Marcus-Spielman-Srivastava to derive a new upper bound on the minimal approximation error. This new bound is asymptotically sharp when the matrix $\mathbf{A}\in\mathbb{R}^{n\times d}$ obeys a spectral power-law decay. The relevant expected characteristic polynomials can be written as an extension of the expected polynomial for the restricted invertibility problem, incorporating two extra variable substitution operators. Finally, we propose a deterministic polynomial-time algorithm that achieves this error bound up to a computational error.
title Interlacing Polynomial Method for the Column Subset Selection Problem
topic Data Structures and Algorithms
Functional Analysis
15A60, 90C27
url https://arxiv.org/abs/2303.07984