Revisiting Approximate Leverage Score Sketching for Matrix Least Squares
Abstract
We revisit the problem of sketching using approximate leverage scores for matrix least squares problems of the form where the design matrix is tall and skinny with . We derive the theoretical results from first principles and clarify the relation to previously stated bounds, improving some constants along the way. One can characterize the utility of a sketching scheme according to the number of samples it needs for an -accurate solution with high probability. Assuming is suitably small, we will show that approximate leverage score sampling requires samples, where is the failure probability and is a measure of the quality of the approximate leverage scores such that corresponds to using exact leverage scores. In cases where a few approximate leverage scores are very large (summing to ), we also show that using a hybrid deterministic and random sampling scheme reduces the required number of samples by a factor of .
Cite
@article{arxiv.2201.10638,
title = {Revisiting Approximate Leverage Score Sketching for Matrix Least Squares},
author = {Brett W. Larsen and Tamara G. Kolda},
journal= {arXiv preprint arXiv:2201.10638},
year = {2026}
}
Comments
This is detailed and standalone derivation of a result that already appears in (arXiv:2006.16438, Appendix A). arXiv admin note: substantial text overlap with arXiv:2006.16438