Small singular values can increase in lower precision
Numerical Analysis
2024-02-22 v3 Numerical Analysis
Abstract
We perturb a real matrix of full column rank, and derive lower bounds for the smallest singular values of the perturbed matrix, in terms of normwise absolute perturbations. Our bounds, which extend existing lower-order expressions, demonstrate the potential increase in the smallest singular values, and represent a qualitative model for the increase in the small singular values after a matrix has been downcast to a lower arithmetic precision. Numerical experiments confirm the qualitative validity of this model and its ability to predict singular values changes in the presence of decreased arithmetic precision.
Cite
@article{arxiv.2303.03547,
title = {Small singular values can increase in lower precision},
author = {Christos Boutsikas and Petros Drineas and Ilse C. F. Ipsen},
journal= {arXiv preprint arXiv:2303.03547},
year = {2024}
}