logo

SCIENTIA SINICA Informationis, Volume 49, Issue 7: 900-910(2019) https://doi.org/10.1360/N112018-00085

Sparse recovery decaying signals based on $\ell^{0}$ regularization via PDASC

More info
  • ReceivedApr 16, 2018
  • AcceptedJul 23, 2018
  • PublishedApr 9, 2019

Abstract

One of the main tasks for sparse recovery is to develop and analyze computationally tractable algorithms for obtaining sparse solutions of underdetermined linear systems. Jiao et al. (2015) proposed a primal-dual active set with a continuation (PDASC) algorithm for the $\ell^0$-regularized least-squares problem and established finite step global convergence and obtained a sharp estimation error under a certain restricted isometry property (RIP) condition. In this paper, we relax the condition on the RIP constant to be independent of the sparsity level $T$ for a class of signals with a strong decay property. Furthermore, we propose a novel data-driven regularization parameter selection rule. Several numerical examples are presented to verify the efficiency and accuracy of the PDASC algorithm and the data-driven parameter choice rule.


Funded by

国家自然科学基金(11501578,11501579,11701571,11801531,11871474,41572315)

国家社科基金(17BTJ017)


References

[1] Candes E J, Romberg J, Tao T. Robust uncertainty principles: exact signal reconstruction from highly incomplete frequency information. IEEE Trans Inform Theor, 2006, 52: 489-509 CrossRef Google Scholar

[2] Donoho D L. Compressed sensing. IEEE Trans Inform Theor, 2006, 52: 1289-1306 CrossRef Google Scholar

[3] Tropp J A. Just relax: convex programming methods for identifying sparse signals in noise. IEEE Trans Inform Theor, 2006, 52: 1030-1051 CrossRef Google Scholar

[4] Zhang H, Wang Y, Chang X Y, et al. $L_{1/2}$ regularization. Sci China Inf Sci, 2010, 40: 412--422. Google Scholar

[5] Zhang H, Zhang H. Approximate message passing algorithm for $L_{1/2}$ regularization. Sci China Inf Sci, 2017, 47: 58--72. Google Scholar

[6] Chen S S, Donoho D L, Saunders M A. Atomic Decomposition by Basis Pursuit. SIAM J Sci Comput, 1998, 20: 33-61 CrossRef Google Scholar

[7] Tibshirani R. Regression shrinkage and selection via the lasso. J R Stat Soc Ser B Stat Methodol, 1996, 58: 267--288, doi: 10.2307/2346178. Google Scholar

[8] Meinshausen N, Bühlmann P. High-dimensional graphs and variable selection with the Lasso. Ann Statist, 2006, 34: 1436-1462 CrossRef Google Scholar

[9] Grasmair M, Scherzer O, Haltmeier M. Necessary and sufficient conditions for linear convergence of $\ell^1$-regularization. Comm Pure Appl Math, 2011, 64: 161-182 CrossRef Google Scholar

[10] Tropp J A, Wright S J. Computational Methods for Sparse Solution of Linear Inverse Problems. Proc IEEE, 2010, 98: 948-958 CrossRef Google Scholar

[11] Rish I, Grabarnik G Y. Sparse Modeling: Theory, Algorithms, and Applications. Boca Raton: CRC Press, 2014. Google Scholar

[12] Fan J, Li R. Variable Selection via Nonconcave Penalized Likelihood and its Oracle Properties. J Am Statistical Association, 2001, 96: 1348-1360 CrossRef Google Scholar

[13] Zhang C H, Zhang T. A General Theory of Concave Regularization for High-Dimensional Sparse Estimation Problems. Statist Sci, 2012, 27: 576-593 CrossRef Google Scholar

[14] Elsener A, Sara V. Sharp oracle inequalities for stationary points of nonconvex penalized M-estimators. 2018,. arXiv Google Scholar

[15] Zeng J S, Fang J, Xu Z B. Sparse SAR imaging based on L 1/2 regularization. Sci China Inf Sci, 2012, 55: 1755-1775 CrossRef Google Scholar

[16] Zeng J, Xu Z, Zhang B. Accelerated regularization based SAR imaging via BCR and reduced Newton skills. Signal Processing, 2013, 93: 1831-1844 CrossRef Google Scholar

[17] Jinshan Zeng , Shaobo Lin , Yao Wang . $L_{1/2}$ Regularization: Convergence of Iterative Half Thresholding Algorithm. IEEE Trans Signal Process, 2014, 62: 2317-2329 CrossRef ADS arXiv Google Scholar

[18] Huang J, Jiao Y, Jin B, et al. A unified primal dual active set algorithm for nonconvex sparse recovery. 2018,. arXiv Google Scholar

[19] Fan J Q, Lv J C. A selective overview of variable selection in high dimensional feature space. Stat Sin, 2010, 20: 101--148. Google Scholar

[20] Chang X Y, Xu Z B, Zhang H, et al. Robust regularization theory based on $L_{q}~(0Google Scholar

[21] Geman S, Geman D. Stochastic relaxation, Gibbs distributions, and the Bayesian restoration of images. IEEE Trans Pattern Anal Mach Intell, 1984, 6: 721--741. Google Scholar

[22] Leclerc Y G. Constructing simple stable descriptions for image partitioning. Int J Comput Vision, 1989, 3: 73-102 CrossRef Google Scholar

[23] Ming Yan , Yi Yang , Osher S. Robust 1-bit Compressive Sensing Using Adaptive Outlier Pursuit. IEEE Trans Signal Process, 2012, 60: 3868-3875 CrossRef ADS Google Scholar

[24] Zhang Y, Dong B, Lu Z. $\ell~_0$ Minimization for wavelet frame based image restoration. Math Comp, 2013, 82: 995-1015 CrossRef Google Scholar

[25] Blumensath T, Davies M E. Iterative Thresholding for Sparse Approximations. J Fourier Anal Appl, 2008, 14: 629-654 CrossRef Google Scholar

[26] Lu Z, Zhang Y. Sparse Approximation via Penalty Decomposition Methods. SIAM J Optim, 2013, 23: 2448-2478 CrossRef Google Scholar

[27] Tseng P. Convergence of a Block Coordinate Descent Method for Nondifferentiable Minimization. J Optimization Theor Appl, 2001, 109: 475-494 CrossRef Google Scholar

[28] Ito K, Kunisch K. A variational approach to sparsity optimization based on Lagrange multiplier theory. Inverse Problems, 2014, 30: 015001 CrossRef ADS Google Scholar

[29] Jiao Y, Jin B, Lu X. A primal dual active set algorithm for a class of nonconvex sparsity optimization. 2013,. arXiv Google Scholar

[30] Hale E T, Yin W, Zhang Y. Fixed-Point Continuation for $\ell_1$-Minimization: Methodology and Convergence. SIAM J Optim, 2008, 19: 1107-1130 CrossRef Google Scholar

[31] Fan Q, Jiao Y, Lu X. A Primal Dual Active Set Algorithm With Continuation for Compressed Sensing. IEEE Trans Signal Process, 2014, 62: 6276-6285 CrossRef ADS arXiv Google Scholar

[32] Jiao Y, Jin B, Lu X. A primal dual active set with continuation algorithm for the $\ell^0$-regularized optimization problem. Appl Comput Harmonic Anal, 2015, 39: 400-426 CrossRef Google Scholar

[33] Proakis, J G, Manolakis, D G. Digital Signal Processing: Principles Algorithms and Applications. New Jersey: Prentice Hall, 1996. Google Scholar

[34] Candes E J, Tao T. Decoding by Linear Programming. IEEE Trans Inform Theor, 2005, 51: 4203-4215 CrossRef Google Scholar

[35] Vershynin R. Introduction to the non-asymptotic analysis of random matrices. 2010,. arXiv Google Scholar

[36] Huang S, Zhu J. Recovery of sparse signals using OMP and its variants: convergence analysis based on RIP. Inverse Problems, 2011, 27: 035003 CrossRef ADS Google Scholar

[37] Mo Q, Shen Y. A Remark on the Restricted Isometry Property in Orthogonal Matching Pursuit. IEEE Trans Inform Theor, 2012, 58: 3654-3656 CrossRef Google Scholar

[38] Ito K, Jin B. Inverse Problems: Tikhonov Theory and Algorithms. Singapore: World Scientific, 2014. Google Scholar

[39] Schwarz G. Estimating the Dimension of a Model. Ann Statist, 1978, 6: 461-464 CrossRef Google Scholar

[40] Wang H, Li R, Tsai C L. Tuning parameter selectors for the smoothly clipped absolute deviation method.. Biometrika, 2007, 94: 553-568 CrossRef PubMed Google Scholar

[41] Chen J, Chen Z. Extended Bayesian information criteria for model selection with large model spaces. Biometrika, 2008, 95: 759-771 CrossRef Google Scholar

[42] Wang H, Li B, Leng C. Shrinkage tuning parameter selection with a diverging number of parameters. J R Statistical Soc-Ser B (Statistical Methodology), 2009, 71: 671-683 CrossRef Google Scholar

[43] Wang L, Kim Y, Li R. Calibrating non-convex penalized regression in ultra-high dimension. Ann Statist, 2013, 41: 2505-2536 CrossRef PubMed Google Scholar

[44] Shi Y Y, Jiao Y L, Yan L, et al. A modified BIC tuning parameter selector for SICA-penalized Cox regression models with diverging dimensionality. J Math, 2017, 37: 723--730 [石跃勇, 焦雨领, 严良, 等. 发散维数 SICA 惩罚 Cox 回归模型的一种修正 BIC 调节参数选择器. 数学杂志, 2017, 37: 723--730]. Google Scholar

[45] Zou H, Hastie T, Tibshirani R. On the "degrees of freedom" of the lasso. Ann Statist, 2007, 35: 2173-2192 CrossRef Google Scholar

[46] Becker S, Bobin J, Candès E J. NESTA: A Fast and Accurate First-Order Method for Sparse Recovery. SIAM J Imag Sci, 2011, 4: 1-39 CrossRef Google Scholar

[47] Blumensath T, Davies M E. Stagewise Weak Gradient Pursuits. IEEE Trans Signal Process, 2009, 57: 4333-4346 CrossRef ADS Google Scholar

[48] Blumensath T. Accelerated iterative hard thresholding. Signal Processing, 2012, 92: 752-756 CrossRef Google Scholar

  •   

    Algorithm 1 PDASC

    Require:$\lambda_0~=~\frac{1}{2}\|\Psi^t~y\|^2_{\infty}$, $\mathcal{A}(\lambda_0)~=~\emptyset$, $x(\lambda_0)~=0$,

    $d(\lambda_0)=~\Psi^t~y$,

    下降因子 $\mu~\in(0,1)$.

    for $k=1,2,\ldots$

    令 $\lambda_k~=~\mu\lambda_{k-1}$, $\mathcal{A}_0~=~\mathcal{A}(\lambda_{k-1})$, $(x^0,d^0)~=~(x(\lambda_{k-1}),~d(\lambda_{k-1}))$;

    for $j=1,2,\ldots$

    定义积极集 $\mathcal{A}_j$ 和非积极集 $\mathcal{I}_j$分别为

Copyright 2019 Science China Press Co., Ltd. 《中国科学》杂志社有限责任公司 版权所有

京ICP备18024590号-1