Perturbation Bounds for Low-Rank Inverse Approximations under Noise
Phuc Tran, Nisheeth K. Vishnoi · 2025
Read Perturbation Bounds for Low-Rank Inverse Approximations under Noise by Phuc Tran, Nisheeth K. Vishnoi (2025) 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-2510-25571.txt.
Description
Phuc Tran and Nisheeth K Vishnoi bound perturbations for low-rank inverses. Their 2025 paper studies matrix inverse approximations under noise. Low-rank pseudoinverses approximate inverses in scalable computing. Machine learning and tuning use them widely. Real-world matrices are often observed with noise. Sampling, sketching, and quantization all present noise. Spectral-norm robustness remains poorly understood in literature.
They study spectral-norm error for symmetric matrices systematically. A noisy reading adds an error matrix to original. Best rank-p inverse approximations are compared against noisy versions. Mild noise assumptions support their theoretical study. Sharp non-asymptotic perturbation bounds are derived. Error scales with eigengap and spectral decay, they reveal. Noise matching with low-curvature directions also matters. Contour integral techniques are applied novelly here. Bounds improve over classic full-inverse adaptations largely. Empirical tests track true perturbation error closely. Classic values greatly overpredict error in tests.
Catalog record
| CREATOR | Phuc Tran, Nisheeth K. Vishnoi |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2510-25571