Improved analysis of randomized SVD for top-eigenvector approximation

02/16/2022
∙
by   Ruo-Chun Tzeng, et al.
∙
0
∙

Computing the top eigenvectors of a matrix is a problem of fundamental interest to various fields. While the majority of the literature has focused on analyzing the reconstruction error of low-rank matrices associated with the retrieved eigenvectors, in many applications one is interested in finding one vector with high Rayleigh quotient. In this paper we study the problem of approximating the top-eigenvector. Given a symmetric matrix 𝐀 with largest eigenvalue Îģ_1, our goal is to find a vector that approximates the leading eigenvector 𝐮_1 with high accuracy, as measured by the ratio R(𝐮Ė‚)=Îģ_1^-1𝐮Ė‚^T𝐀𝐮Ė‚/𝐮Ė‚^T𝐮Ė‚. We present a novel analysis of the randomized SVD algorithm of <cit.> and derive tight bounds in many cases of interest. Notably, this is the first work that provides non-trivial bounds of R(𝐮Ė‚) for randomized SVD with any number of iterations. Our theoretical analysis is complemented with a thorough experimental study that confirms the efficiency and accuracy of the method.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
∙ 07/25/2017

Computing low-rank approximations of large-scale matrices with the Tensor Network randomized SVD

We propose a new algorithm for the computation of a singular value decom...
research
∙ 09/06/2023

A multilinear NystrÃļm algorithm for low-rank approximation of tensors in Tucker format

The NystrÃļm method offers an effective way to obtain low-rank approximat...
research
∙ 06/17/2022

A general error analysis for randomized low-rank approximation methods

We propose a general error analysis related to the low-rank approximatio...
research
∙ 11/09/2022

Efficient Bounds and Estimates for Canonical Angles in Randomized Subspace Approximations

Randomized subspace approximation with "matrix sketching" is an effectiv...
research
∙ 06/21/2023

Randomized algorithms for low-rank matrix approximation: Design, analysis, and applications

This survey explores modern approaches for computing low-rank approximat...
research
∙ 05/19/2023

A randomized algorithm for the QR decomposition-based approximate SVD

Matrix decomposition is a very important mathematical tool in numerical ...
research
∙ 08/22/2019

Block Randomized Optimization for Adaptive Hypergraph Learning

The high-order relations between the content in social media sharing pla...

Please sign up or login with your details

Forgot password? Click here to reset