Optimal Condition Numbers in Low-Rank Positive Semidefinite Matrix Sensing

DOI10.48550/arXiv.2608.22418MaRDI QIDQ7285012arXivFDO

Authors Mingxuan Sun, Zhiqiang Xu

Publication date 23 August 2026

Published in arXiv

Tldr We provide deterministic universal bounds on the global condition numbers of positive semidefinite matrix sensing maps and show that random rank-one Gaussian measurements achieve these lower limits asymptotically, hence being optimal in terms of stability guarantees for related low-rank matrix recovery methods.

Full work available at URL https://doi.org/10.48550/arXiv.2608.22418

arXiv classification

cs.ITmath.FAmath.NA

Cites work

• Complex phase retrieval from subgaussian measurements
• Exact and Stable Covariance Estimation From Quadratic Sampling via Convex Programming
• Inequalities: theory of majorization and its applications
• Introduction to probability models
• Invertibility and robustness of phaseless reconstruction
• Learning without concentration
• Locality and stability for phase retrieval
• Low-Rank Positive Semidefinite Matrix Recovery From Corrupted Rank-One Measurements
• Minimax estimation in sparse canonical correlation analysis
• On Lipschitz analysis and Lipschitz synthesis for the phase retrieval problem
• Quantum state tomography for matrix product density operators
• Saving phase: injectivity and stability for phase retrieval
• Solving quadratic equations via phaselift when there are about as many equations as unknowns
• Stability in phase retrieval: characterizing condition numbers and the optimal vector set
• Stable low-rank matrix recovery via null space properties
• Tail bounds via generic chaining
• The Condition Number in Phase Retrieval from Intensity Measurements

This page was built for publication: Optimal Condition Numbers in Low-Rank Positive Semidefinite Matrix Sensing

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q7285012)