Debiasing Random Oblique Projections for Subsampled OLS and Fast CUR in High Dimensions

Chengmei Niu, Sachin Garg, Michał Dereziński, Zhenyu Liao · 2026

Read Debiasing Random Oblique Projections for Subsampled OLS and Fast CUR in High Dimensions by Chengmei Niu, Sachin Garg, Michał Dereziński, Zhenyu Liao (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2605-24955.txt.

Description

Chengmei Niu, Sachin Garg, Michał Dereziński, and Zhenyu Liao expose a hidden bias in random sampling methods for large matrix problems. Because inverting a sketched matrix is nonlinear, even an unbiased sketch can yield systematically biased projections. That skews subsampled least squares and fast low-rank approximations. The authors build a unified theory for these random oblique projections in high dimensions and a principled correction, sharpening bias-variance analysis for least squares and improving accuracy for CUR decomposition.

Catalog record

CREATORChengmei Niu, Sachin Garg, Michał Dereziński, Zhenyu Liao
YEAR2026
COLLECTIONPAPERS
FORMATTXT
FILE SIZE258832
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2605.24955

Browse the full library · About the Tower of Babel