View a PDF of the paper titled Variance Reduction for Matrix Computations with Applications to Gaussian Processes, by Anant Mathur and 1 other authors
Abstract:In addition to recent developments in computing speed and memory, methodological advances have contributed to significant gains in the performance of stochastic simulation. In this paper we focus on variance reduction for matrix computations via matrix factorization. We provide insights into existing variance reduction methods for estimating the entries of large matrices. Popular methods do not exploit the reduction in variance that is possible when the matrix is factorized. We show how computing the square root factorization of the matrix can achieve in some important cases arbitrarily better stochastic performance. In addition, we detail a factorized estimator for the trace of a product of matrices and numerically demonstrate that the estimator can be up to 1,000 times more efficient on certain problems of estimating the log-likelihood of a Gaussian process. Additionally, we provide a new estimator of the log-determinant of a symmetric positive-definite matrix where the log-determinant is treated as a normalizing constant of a probability density.
Submission history
From: Anant Mathur [view email]
[v1]
Mon, 28 Jun 2021 10:41:22 UTC (529 KB)
[v2]
Tue, 6 Jul 2021 08:19:40 UTC (370 KB)
[v3]
Sun, 26 Mar 2023 08:32:01 UTC (1 KB) (withdrawn)
[v4]
Thu, 19 Nov 2026 02:08:45 UTC (65 KB)
![[2106.14565] Variance Reduction for Matrix Computations with Applications to Gaussian Processes [2106.14565] Variance Reduction for Matrix Computations with Applications to Gaussian Processes](https://techwavedispatch.com/wp-content/uploads/2026/11/210614565-Variance-Reduction-for-Matrix-Computations-with-Applications-to-Gaussian-1024x597.png)