English

Efficient error and variance estimation for randomized matrix computations

Numerical Analysis 2024-10-03 v5 Numerical Analysis Machine Learning

Abstract

Randomized matrix algorithms have become workhorse tools in scientific computing and machine learning. To use these algorithms safely in applications, they should be coupled with posterior error estimates to assess the quality of the output. To meet this need, this paper proposes two diagnostics: a leave-one-out error estimator for randomized low-rank approximations and a jackknife resampling method to estimate the variance of the output of a randomized matrix computation. Both of these diagnostics are rapid to compute for randomized low-rank approximation algorithms such as the randomized SVD and randomized Nystr\"om approximation, and they provide useful information that can be used to assess the quality of the computed output and guide algorithmic parameter choices.

Keywords

Cite

@article{arxiv.2207.06342,
  title  = {Efficient error and variance estimation for randomized matrix computations},
  author = {Ethan N. Epperly and Joel A. Tropp},
  journal= {arXiv preprint arXiv:2207.06342},
  year   = {2024}
}

Comments

22 pages, 10 figures, 13 pages of supplementary material. v5: added derivation of leave-one-out error estimate to supplementary material