Tight Oracle Inequalities for Low-Rank Matrix Recovery From a Minimal Number of Noisy Random Measurements

• Published in: IEEE transactions on information theory Vol. 57; no. 4; pp. 2342 - 2359
• Main Authors: Candès, Emmanuel J, Plan, Y
• Format: Journal Article
• Language: English
• Published: New York, NY IEEE 01.04.2011; Institute of Electrical and Electronics Engineers; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Access methods and protocols, osi model; Applied sciences; Compressed sensing; Convex optimization; Dantzig selector; Error correction & detection; Exact sciences and technology; Information processing; Information theory; Information, signal and communications theory; Linear matrix inequalities; Matrix; matrix completion; Measurement uncertainty; Minimization; Noise; Noise measurement; norm of random matrices; oracle inequalities and semidefinite programming; Sparse matrices; Telecommunications; Telecommunications and information theory; Teleprocessing networks. Isdn; Singularity; Isometry; oracle inequalities and semidefinite programming; matrix completion; Transmission protocol; Theoretical study; norm of random matrices; Dantzig selector; Constrained optimization; Linear combination; Convex optimization; Minimax method; Routing protocols; Oracle
• ISSN: 0018-9448; 1557-9654
• DOI: 10.1109/TIT.2011.2111771

This paper presents several novel theoretical results regarding the recovery of a low-rank matrix from just a few measurements consisting of linear combinations of the matrix entries. We show that properly constrained nuclear-norm minimization stably recovers a low-rank matrix from a constant number of noisy measurements per degree of freedom; this seems to be the first result of this nature. Further, with high probability, the recovery error from noisy data is within a constant of three targets: (1) the minimax risk, (2) an "oracle" error that would be available if the column space of the matrix were known, and (3) a more adaptive "oracle" error which would be available with the knowledge of the column space corresponding to the part of the matrix that stands above the noise. Lastly, the error bounds regarding low-rank matrices are extended to provide an error bound when the matrix has full rank with decaying singular values. The analysis in this paper is based on the restricted isometry property (RIP).