Finding a low-rank basis in a matrix subspace

For a given matrix subspace, how can we find a basis that consists of low-rank matrices? This is a generalization of the sparse vector problem. It turns out that when the subspace is spanned by rank-1 matrices, the matrices can be obtained by the tensor CP decomposition. For the higher rank case, the situation is not as straightforward. In this work we present an algorithm based on a greedy process applicable to higher rank problems. Our algorithm first estimates the minimum rank by applying soft singular value thresholding to a nuclear norm relaxation, and then computes a matrix with that rank using the method of alternating projections. We provide local convergence results, and compare our algorithm with several alternative approaches. Applications include data compression beyond the classical truncated SVD, computing accurate eigenvectors of a near-multiple eigenvalue, image separation and graph Laplacian eigenproblems.

[1]  Huan Wang,et al.  Exact Recovery of Sparsely-Used Dictionaries , 2012, COLT.

[2]  G. W. Stewart,et al.  Matrix algorithms , 1998 .

[3]  Marwan Mattar,et al.  Labeled Faces in the Wild: A Database forStudying Face Recognition in Unconstrained Environments , 2008 .

[4]  Joos Vandewalle,et al.  A Multilinear Singular Value Decomposition , 2000, SIAM J. Matrix Anal. Appl..

[5]  Laurent Demanet,et al.  Recovering the Sparsest Element in a Subspace , 2013, 1310.1654.

[6]  採編典藏組 Society for Industrial and Applied Mathematics(SIAM) , 2008 .

[7]  Leonid Gurvits,et al.  Classical complexity and quantum entanglement , 2004, J. Comput. Syst. Sci..

[8]  J. Chang,et al.  Analysis of individual differences in multidimensional scaling via an n-way generalization of “Eckart-Young” decomposition , 1970 .

[9]  A. Uschmajew,et al.  A new convergence proof for the higher-order power method and generalizations , 2014, 1407.4586.

[10]  James Demmel,et al.  Applied Numerical Linear Algebra , 1997 .

[11]  Christopher J. Hillar,et al.  Most Tensor Problems Are NP-Hard , 2009, JACM.

[12]  F. L. Hitchcock The Expression of a Tensor or a Polyadic as a Sum of Products , 1927 .

[13]  Sara van de Geer,et al.  Statistics for High-Dimensional Data: Methods, Theory and Applications , 2011 .

[14]  T. Coleman,et al.  The null space problem I. complexity , 1986 .

[15]  Joos Vandewalle,et al.  Computation of the Canonical Decomposition by Means of a Simultaneous Generalized Schur Decomposition , 2005, SIAM J. Matrix Anal. Appl..

[16]  Lieven De Lathauwer,et al.  Decompositions of a Higher-Order Tensor in Block Terms - Part I: Lemmas for Partitioned Matrices , 2008, SIAM J. Matrix Anal. Appl..

[17]  Stephen A. Vavasis,et al.  Nuclear norm minimization for the planted clique and biclique problems , 2009, Math. Program..

[18]  David Steurer,et al.  Rounding sum-of-squares relaxations , 2013, Electron. Colloquium Comput. Complex..

[19]  Richard A. Harshman,et al.  Foundations of the PARAFAC procedure: Models and conditions for an "explanatory" multi-model factor analysis , 1970 .

[20]  L. Demanet,et al.  Recovering the Sparsest Element in a Subspace , 2013 .

[21]  Stefan Kindermann,et al.  News Algorithms for tensor decomposition based on a reduced functional , 2014, Numer. Linear Algebra Appl..

[22]  Emmanuel J. Candès,et al.  Near-Optimal Signal Recovery From Random Projections: Universal Encoding Strategies? , 2004, IEEE Transactions on Information Theory.

[23]  Martin J. Mohlenkamp Musings on multilinear fitting , 2013 .

[24]  Aude Rondepierre,et al.  On Local Convergence of the Method of Alternating Projections , 2013, Foundations of Computational Mathematics.

[25]  Yong-Jin Liu,et al.  An implementable proximal point algorithmic framework for nuclear norm minimization , 2012, Math. Program..

[26]  Emmanuel J. Candès,et al.  The Power of Convex Relaxation: Near-Optimal Matrix Completion , 2009, IEEE Transactions on Information Theory.

[27]  Tamara G. Kolda,et al.  Tensor Decompositions and Applications , 2009, SIAM Rev..

[28]  Uwe Helmke,et al.  Critical points of matrix least squares distance functions , 1995 .

[29]  Lieven De Lathauwer,et al.  Canonical Polyadic Decomposition of Third-Order Tensors: Reduction to Generalized Eigenvalue Decomposition , 2013, SIAM J. Matrix Anal. Appl..

[30]  E. Candès The restricted isometry property and its implications for compressed sensing , 2008 .

[31]  John Wright,et al.  Finding a Sparse Vector in a Subspace: Linear Sparsity Using Alternating Directions , 2014, IEEE Transactions on Information Theory.

[32]  Lieven De Lathauwer,et al.  A Link between the Canonical Decomposition in Multilinear Algebra and Simultaneous Matrix Diagonalization , 2006, SIAM J. Matrix Anal. Appl..

[33]  Terrence J. Sejnowski,et al.  An Information-Maximization Approach to Blind Separation and Blind Deconvolution , 1995, Neural Computation.

[34]  Emmanuel J. Candès,et al.  Exact Matrix Completion via Convex Optimization , 2009, Found. Comput. Math..

[35]  Andrzej Cichocki,et al.  Tensor Decompositions for Signal Processing Applications: From two-way to multiway component analysis , 2014, IEEE Signal Processing Magazine.

[36]  André Uschmajew,et al.  Local Convergence of the Alternating Least Squares Algorithm for Canonical Tensor Approximation , 2012, SIAM J. Matrix Anal. Appl..

[37]  Adrian S. Lewis,et al.  Alternating Projections on Manifolds , 2008, Math. Oper. Res..

[38]  F. Andersson,et al.  Alternating Projections on Nontangential Manifolds , 2013 .

[39]  Dmitriy Drusvyatskiy,et al.  Transversality and Alternating Projections for Nonconvex Sets , 2014, Found. Comput. Math..

[40]  Marek Karpinski,et al.  Generalized Wong sequences and their applications to Edmonds' problems , 2015, J. Comput. Syst. Sci..

[41]  S. Leurgans,et al.  A Decomposition for Three-Way Arrays , 1993, SIAM J. Matrix Anal. Appl..

[42]  Adrian S. Lewis,et al.  Local Linear Convergence for Alternating and Averaged Nonconvex Projections , 2009, Found. Comput. Math..

[43]  Emmanuel J. Candès,et al.  A Singular Value Thresholding Algorithm for Matrix Completion , 2008, SIAM J. Optim..

[44]  John Wright,et al.  Complete Dictionary Recovery Over the Sphere II: Recovery by Riemannian Trust-Region Method , 2015, IEEE Transactions on Information Theory.

[45]  Liqi Wang,et al.  On the Global Convergence of the Alternating Least Squares Method for Rank-One Approximation to Generic Tensors , 2014, SIAM J. Matrix Anal. Appl..

[46]  J. Edmonds Systems of distinct representatives and linear algebra , 1967 .

[47]  David R. Karger,et al.  Deterministic network coding by matrix completion , 2005, SODA '05.

[48]  Sara van de Geer,et al.  Statistics for High-Dimensional Data , 2011 .

[49]  Na Li,et al.  Some Convergence Results on the Regularized Alternating Least-Squares Method for Tensor Decomposition , 2011, 1109.3831.

[50]  James Oxley Matroid Applications: Infinite Matroids , 1992 .

[51]  G. Stewart,et al.  Matrix Perturbation Theory , 1990 .

[52]  Lieven Vandenberghe,et al.  Interior-Point Method for Nuclear Norm Approximation with Application to System Identification , 2009, SIAM J. Matrix Anal. Appl..

[53]  Wotao Yin,et al.  A Block Coordinate Descent Method for Regularized Multiconvex Optimization with Applications to Nonnegative Tensor Factorization and Completion , 2013, SIAM J. Imaging Sci..

[54]  Benjamin Recht,et al.  A Simpler Approach to Matrix Completion , 2009, J. Mach. Learn. Res..

[55]  John Wright,et al.  Complete Dictionary Recovery Over the Sphere I: Overview and the Geometric Picture , 2015, IEEE Transactions on Information Theory.

[56]  D. Russell Luke Prox-Regularity of Rank Constraint Sets and Implications for Algorithms , 2012, Journal of Mathematical Imaging and Vision.

[57]  Pablo A. Parrilo,et al.  Guaranteed Minimum-Rank Solutions of Linear Matrix Equations via Nuclear Norm Minimization , 2007, SIAM Rev..

[58]  László Lovász,et al.  Singular spaces of matrices and their application in combinatorics , 1989 .

[59]  David R. Karger,et al.  The complexity of matrix completion , 2006, SODA '06.

[60]  Jack Dongarra,et al.  Templates for the Solution of Algebraic Eigenvalue Problems , 2000, Software, environments, tools.

[61]  Saeid Sanei,et al.  Blind Separation of Image Sources via Adaptive Dictionary Learning , 2012, IEEE Transactions on Image Processing.

[62]  Rajeev Motwani,et al.  Randomized Algorithms , 1995, SIGA.

[63]  Johan Håstad,et al.  Tensor Rank is NP-Complete , 1989, ICALP.

[64]  Stephen P. Boyd,et al.  A rank minimization heuristic with application to minimum order system approximation , 2001, Proceedings of the 2001 American Control Conference. (Cat. No.01CH37148).

[65]  P. Wedin Perturbation bounds in connection with singular value decomposition , 1972 .

[66]  Gene H. Golub,et al.  Matrix computations , 1983 .

[67]  Lieven De Lathauwer,et al.  Decompositions of a Higher-Order Tensor in Block Terms - Part II: Definitions and Uniqueness , 2008, SIAM J. Matrix Anal. Appl..

[68]  Wei Dai,et al.  Joint image separation and dictionary learning , 2013, 2013 18th International Conference on Digital Signal Processing (DSP).