Abstract
The matrix rank minimization problem arises in many engineering applications. As this problem is NP-hard, a nonconvex relaxation of matrix rank minimization, called the Schatten-p quasi-norm minimization (0 < p < 1), has been developed to approximate the rank function closely. We study the performance of projected gradient descent algorithm for solving the Schatten-p quasi-norm minimization (0 < p < 1) problem. Based on the matrix restricted isometry property (M-RIP), we give the convergence guarantee and error bound for this algorithm and show that the algorithm is robust to noise with an exponential convergence rate.
Similar content being viewed by others
References
Bahmani S, Raj B. A unifying analysis of projected gradient descent for lp-constrained least squares. Appl Comput Harmon Anal, 2013, 34: 366–378
Blumensath T, Davies M. Iterative thresholding for sparse approximations. J Fourier Anal Appl, 2008, 14: 629–654
Blumensath T, Davies M. Iterative hard thresholding for compressed sensing. Appl Comput Harmon Anal, 2009, 27: 265–274
Cai J F, Candès E J, Shen Z W. A singular value thresholding algorithm for matrix completion. SIAM J Optim, 2010, 20: 1956–1982
Cai T, Zhang A. Sharp RIP bound for sparse signal and low-rank matrix recovery. Appl Comput Harmon Anal, 2013, 35: 74–93
Cai T, Zhang A. Sparse representation of a polytope and recovery of sparse signals and low-rank matrices. IEEE Trans Inform Theory, 2014, 60: 122–132
Candès E J, Plan Y. Tight oracle inequalities for low-rank matrix recovery from a minimal number of noisy random measurements. IEEE Trans Inform Theory, 2011, 57: 2342–2359
Candès E J, Recht B. Exact matrix completion via convex optimization. Found Comput Math, 2009, 9: 717–772
Deng Y, Dai Q, Liu R, et al. Low-rank structure learning via nonconvex heuristic recovery. IEEE Trans Neural Netw Learn Syst, 2013, 24: 383–396
Fazel M. Matrix rank minimization with applications. PhD thesis. Stanford: Stanford University, 2002
Fornasier M, Rauhut H, Ward R. Low-rank matrix recovery via iteratively reweighted least squares minimization. SIAM J Optim, 2011, 21: 1614–1640
Foucart S. Sparse recovery algorithms: Sufficient conditions in terms of restricted isometry constants. Springer Proc Math, 2012, 13: 65–77
Goldfarb D, Ma S. Convergence of fixed-point continuation algorithms for matrix rank minimization. Found Comput Math, 2011, 11: 183–210
Gross D. Recovery Low-rank matrices from few coefficients in any basis. IEEE Trans Inform Theory, 2011, 57: 1548–1566
Gross D, Liu Y K, Flammia S T, et al. Quantum state tomography via compressed sensing. Phys Rev Lett, 2010, 105: 150401
Haldar J, Liang Z. Spatiotemporal imaging with partially separable functions: A matrix recovery approach. In: Proc IEEE Int Symp Biomed Imag. New York: IEEE, 2010, 716–719
Ji H, Liu C, Shen Z W, et al. Robust video denoising using low rank matrix completion. In: IEEE Conference on Computer Vision and Pattern Recognition. San Francisco, CA: IEEE Computer Society, 2010, 1791–1798
Kramer F, Ward R. New and improved Johnson-Lindenstrauss embeddings via the restricted isometry property. SIAM J Math Anal, 2011, 43: 1269–1281
Lai M, Xu Y, Yin W. Improved iteratively reweighted least square for unconstrained smoothed l q minimization. SIAM J Numer Anal, 2013, 51: 927–957
Lee K, Bresler Y. ADMiRA: Atomic decomposition for minimum rank approximation. IEEE Trans Inform Theory, 2010, 56: 4402–4416
Lewis A S. The convex analysis of unitarily invariant matrix norms. J Convex Anal, 1995, 2: 173–183
Lin J, Li S. Convergence of projected Landweber iteration for matrix rank minimization. Appl Comput Harmon Anal, 2014, 36: 316–325
Liu L, Huang W, Chen D R. Exact minimum rank approximation via Schatten p-norm minimization. J Comput Appl Math, 2014, 267: 218–227
Liu Z, Vandenberghe L. Interior-point method for nuclear norm approximation with application to system identification. SIAM J Matrix Anal Appl, 2008, 31: 1235–1256
Ma S, Goldfarb D, Chen L. Fixed point and Bregman iterative methods for matrix rank minimization. Math Program, 2011, 128: 321–353
Majumdar A, Ward R K. An algorithm for sparse MRI reconstruction by Schatten p-norm minimization. Magnetic Resonance Imaging, 2011, 29: 408–417
Marjanovic G, Solo V. On l q Optimization and Matrix Completion. IEEE Trans Signal Process, 2012, 60: 5714–5724
Mazumder R, Hastie T, Tibshirani R. Spectral regularization algorithms for learning large incomplete matrices. J Mach Learn Res, 2010, 11: 2287–2322
Meka R, Jain P, Dhillon L S. Guaranteed rank minimization via singular value projection. ArXiv:0909.5457, 2009
Mohan K, Fazel M. Reweighted nuclear norm minimization with application to system identification. In: Proc of the American Control Conference. Piscataway, NJ: IEEE, 2010: 2953–2959
Mohan K, Fazel M. Iterative reweighted least squares for matrix rank minimization. In: 48th Annual Allerton Conference on Communication, Control and Computing (Allerton). Piscataway, NJ: IEEE, 2010: 653–661
Needell D, Tropp J. CoSaMP: Iterative signal recovery from incomplete and inaccurate samples. Appl Comput Harmon Anal, 2009, 26: 301–321
Netflix Prize [Online]. Available: http://www.netflixprize.com/.
Oymak S, Mohan K, Fazel M, et al. A simplified approach to recovery conditions for low rank matrices. In: IEEE International Symposium on Information Theory Proceedings. Piscataway, NJ: IEEE, 2011: 2318–2322
Recht B, Fazel M, Parrilo P. Guaranteed minimum-rank solutions of linear matrix equations via nuclear norm minimization. SIAM Rev, 2010, 52: 471–501
Singer A. A remark on global positioning from local distances. Proc Natl Acad Sci USA, 2008, 105: 9507
Srebro N, Rennie J F M, Jakola T S. Maximum-margin matrix factorization. In: Advance in Neural Information Processing Systems. Cambridge-Massachusetts: MIT press, 2005
Toh K C, Yun S W. An accelerated proximal gradient algorithm for nuclear norm regularized least squares problems. Pac J Optim, 2010, 6: 615–640
Wang H, Li S. The bounds of restricted isometry constants for low rank matrices recovery. Sci China Math, 2013, 56: 1117–1127
Zhang H, Cheng L Z. Projected Landweber iteration for matrix completion. J Comput Appl Math, 2010, 235: 593–601
Zhang M, Huang Z, Zhang Y. Restricted p-Isometry Properties of Nonconvex matrix recovery. IEEE Trans Inform Theory, 2013, 59: 4316–4323
Zhao B, Haldar J, Brinegar C, et al. Low rank matrix recovery for real-time cardiac MRI. In: Proc IEEE Int Symp Biomed Imag. New York: IEEE, 2010, 996–999
Author information
Authors and Affiliations
Corresponding author
Rights and permissions
About this article
Cite this article
Cai, Y., Li, S. Convergence analysis of projected gradient descent for Schatten-p nonconvex matrix recovery. Sci. China Math. 58, 845–858 (2015). https://doi.org/10.1007/s11425-014-4949-1
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s11425-014-4949-1
Keywords
- low rank matrix recovery
- nonconvex matrix recovery
- projected gradient descent
- restricted isometry property