Low-Rank Updates of Matrix Square Roots

TL;DR

This paper develops methods for efficiently computing low-rank corrections to matrix square roots and inverse square roots, leveraging algebraic Riccati equations and eigenvalue decay bounds, with applications in data science.

Contribution

It introduces a novel approach to approximate low-rank updates of matrix square roots using Riccati equations and eigenvalue decay analysis, enabling efficient computations.

Findings

Low-rank corrections exist for (inverse) square roots of matrices with low-rank perturbations.

The proposed methods achieve controlled approximation errors with spectral and Frobenius norm bounds.

Numerical experiments demonstrate the effectiveness of the algorithms in practical applications.

Abstract

Models in which the covariance matrix has the structure of a sparse matrix plus a low rank perturbation are ubiquitous in data science applications. It is often desirable for algorithms to take advantage of such structures, avoiding costly matrix computations that often require cubic time and quadratic storage. This is often accomplished by performing operations that maintain such structures, e.g. matrix inversion via the Sherman-Morrison-Woodbury formula. In this paper we consider the matrix square root and inverse square root operations. Given a low rank perturbation to a matrix, we argue that a low-rank approximate correction to the (inverse) square root exists. We do so by establishing a geometric decay bound on the true correction's eigenvalues. We then proceed to frame the correction as the solution of an algebraic Riccati equation, and discuss how a low-rank solution to that equation can be computed. We analyze the approximation error incurred when approximately solving the algebraic Riccati equation, providing spectral and Frobenius norm forward and backward error bounds. Finally, we describe several applications of our algorithms, and demonstrate their utility in numerical experiments.