Abstract:We study the matrix completion problem: an underlying matrix P is low rank, with incoherent singular vectors, and a random matrix A is equal to P on a (uniformly) random subset of entries of size dn . All other entries of A are equal to zero. The goal is to retrieve information on P from the observation of A . Let be the random matrix where each entry of A is multiplied by an independent -Bernoulli random variable with parameter 1/2. This paper is about when, how and why the non-Hermitian eigen-spectra of the matrices and captures more of the relevant information about the principal component structure of A than the eigen-spectra of and . We show that the eigenvalues of the asymmetric matrices and with modulus greater than a detection threshold are asymptotically equal to the eigenvalues of and and that the associated eigenvectors are aligned as well. The central surprise is that by intentionally inducing asymmetry and additional randomness via the matrix, we can extract more information than if we had worked with the singular value decomposition (SVD) of A . The associated detection threshold is asymptotically exact and is non-universal since it explicitly depends on the element-wise distribution of the underlying matrix P . We show that reliable, statistically optimal but not perfect matrix recovery, via a universal data-driven algorithm, is possible above this detection threshold using the information extracted from the asymmetric eigen-decompositions. Averaging the left and right eigenvectors provably improves estimation accuracy but not the detection threshold. Our results encompass the very sparse regime where d is of order 1 where matrix completion via the SVD of A fails or produces unreliable recovery. We define another variant of this asymmetric principal component analysis procedure that bypasses the randomization step and has a detection threshold that is smaller by a constant factor but with a computational cost that is larger by a polynomial factor of the number of observed entries. Both detection thresholds allow to go beyond the barrier due to the well-known information theoretical limit for exact matrix completion found in the literature.

On the circulant matrix MDS testing and the search for circulant MDS matrices

Construction of a set of circulant matrix submatrices for faster MDS matrix verification

On the Counting of Involutory MDS Matrices

A note on MDS Property of Circulant Matrices

On the Construction of Near-MDS Matrices

On MDS Property of g-Circulant Matrices

On the Direct Construction of MDS and Near-MDS Matrices

A Systematic Construction Approach for All $4\times 4$ Involutory MDS Matrices

A systematic construction approach for all involutory MDS matrices

Direct Construction of Recursive MDS Diffusion Layers using Shortened BCH Codes

On the construction of ultra-light MDS matrices

A Characterization of Semi-Involutory MDS Matrices

Construction of all MDS and involutory MDS matrices

Efficient implementation of the linear layer of block ciphers with large MDS matrices based on a new lookup table technique

Private and Secure Distributed Matrix Multiplication Schemes for Replicated or MDS-Coded Servers

Information-Theoretically Private Matrix Multiplication From MDS-Coded Storage

Preconditioning Technique for Symmetric M-matrices

Detection Thresholds in Very Sparse Matrix Completion

Matrix Multiplication Verification Using Coding Theory

Flexible Field Sizes in Secure Distributed Matrix Multiplication via Efficient Interference Cancellation

A Matrix Hyperbolic Cosine Algorithm and Applications