Iterative methods for solving factorized linear systems

Aaditya Ramdas Optimization

Stochastic iterative algorithms such as the Kaczmarz and Gauss-Seidel methods have gained recent attention because of their speed, simplicity, and the ability to approximately solve large-scale linear systems of equations without needing to access the entire matrix. In this work, we consider the setting where we wish to solve a linear system in a large matrix X that is stored in a factorized form, X = UV; this setting either arises naturally in many applications or may be imposed when working with large low-rank datasets for reasons of space required for storage. We propose a variant of the randomized Kaczmarz method for such systems that takes advantage of the factored form, and avoids computing X. We prove an exponential convergence rate and supplement our theoretical guarantees with experimental evidence demonstrating that the factored variant yields significant acceleration in convergence.

Published On: January 25, 2017

Link: https://arxiv.org/abs/1701.07453

Authors: Aaditya Ramdas, Anna Ma, Deanna Needell