arXiv Analytics

Sign in

arXiv:2311.07035 [math.NA]AbstractReferencesReviewsResources

ContHutch++: Stochastic trace estimation for implicit integral operators

Jennifer Zvonek, Andrew Horning, Alex Townsend

Published 2023-11-13Version 1

Hutchinson's estimator is a randomized algorithm that computes an $\epsilon$-approximation to the trace of any positive semidefinite matrix using $\mathcal{O}(1/\epsilon^2)$ matrix-vector products. An improvement of Hutchinson's estimator, known as Hutch++, only requires $\mathcal{O}(1/\epsilon)$ matrix-vector products. In this paper, we propose a generalization of Hutch++, which we call ContHutch++, that uses operator-function products to efficiently estimate the trace of any trace-class integral operator. Our ContHutch++ estimates avoid spectral artifacts introduced by discretization and are accompanied by rigorous high-probability error bounds. We use ContHutch++ to derive a new high-order accurate algorithm for quantum density-of-states and also show how it can estimate electromagnetic fields induced by incoherent sources.

Related articles: Most relevant | Search more
arXiv:2109.10659 [math.NA] (Published 2021-09-22)
Improved variants of the Hutch++ algorithm for trace estimation
arXiv:2301.07825 [math.NA] (Published 2023-01-19)
XTrace: Making the most of every sample in stochastic trace estimation
arXiv:2212.09841 [math.NA] (Published 2022-12-19)
Matrix recovery from matrix-vector products