A Learned Proximal Alternating Minimization Algorithm and Its Induced Network for a Class of Two-Block Nonconvex and Nonsmooth Optimization

被引:0
作者
Chen, Yunmei [1 ]
Liu, Lezhi [1 ]
Zhang, Lei [1 ]
机构
[1] Univ Florida, Dept Math, Gainesville, FL 32611 USA
关键词
Learned alternating minimization algorithm; Nonconvex nonsmooth optimization; Deep learning; MRI Image reconstruction; LINEARIZED MINIMIZATION; CONVERGENCE; MATRIX;
D O I
10.1007/s10915-025-02874-5
中图分类号
O29 [应用数学];
学科分类号
070104 ;
摘要
This work proposes a general learned proximal alternating minimization algorithm, LPAM, for solving learnable two-block nonsmooth and nonconvex optimization problems. We tackle the nonsmoothness by an appropriate smoothing technique with automatic diminishing smoothing effect. For smoothed nonconvex problems we modify the proximal alternating linearized minimization (PALM) scheme by incorporating the residual learning architecture, which has proven to be highly effective in deep network training, and employing the block coordinate decent (BCD) iterates as a safeguard for the convergence of the algorithm. We prove that there is a subsequence of the iterates generated by LPAM, which has at least one accumulation point and each accumulation point is a Clarke stationary point. Our method is widely applicable as one can employ various learning problems formulated as two-block optimizations, and is also easy to be extended for solving multi-block nonsmooth and nonconvex optimization problems. The network, whose architecture follows the LPAM exactly, namely LPAM-net, inherits the convergence properties of the algorithm to make the network interpretable. As an example application of LPAM-net, we present the numerical and theoretical results on the application of LPAM-net for joint multi-modal MRI reconstruction with significantly under-sampled k-space data. The experimental results indicate the proposed LPAM-net is parameter-efficient and has favourable performance in comparison with some state-of-the-art methods.
引用
收藏
页数:32
相关论文
共 34 条
[1]   A gradient-based alternating minimization approach for optimization of the measurement matrix in compressive sensing [J].
Abolghasemi, Vahid ;
Ferdowsi, Saideh ;
Sanei, Saeid .
SIGNAL PROCESSING, 2012, 92 (04) :999-1009
[2]   MoDL: Model-Based Deep Learning Architecture for Inverse Problems [J].
Aggarwal, Hemant K. ;
Mani, Merry P. ;
Jacob, Mathews .
IEEE TRANSACTIONS ON MEDICAL IMAGING, 2019, 38 (02) :394-405
[3]   Convergence of descent methods for semi-algebraic and tame problems: proximal algorithms, forward-backward splitting, and regularized Gauss-Seidel methods [J].
Attouch, Hedy ;
Bolte, Jerome ;
Svaiter, Benar Fux .
MATHEMATICAL PROGRAMMING, 2013, 137 (1-2) :91-129
[4]   Proximal Alternating Minimization and Projection Methods for Nonconvex Problems: An Approach Based on the Kurdyka-Lojasiewicz Inequality [J].
Attouch, Hedy ;
Bolte, Jerome ;
Redont, Patrick ;
Soubeyran, Antoine .
MATHEMATICS OF OPERATIONS RESEARCH, 2010, 35 (02) :438-457
[5]   ON THE CONVERGENCE OF BLOCK COORDINATE DESCENT TYPE METHODS [J].
Beck, Amir ;
Tetruashvili, Luba .
SIAM JOURNAL ON OPTIMIZATION, 2013, 23 (04) :2037-2060
[6]  
Bertsekas D., 2015, PARALLEL DISTRIBUTED
[7]   A Learnable Variational Model for Joint Multimodal MRI Reconstruction and Synthesis [J].
Bian, Wanyu ;
Zhang, Qingchao ;
Ye, Xiaojing ;
Chen, Yunmei .
MEDICAL IMAGE COMPUTING AND COMPUTER ASSISTED INTERVENTION, MICCAI 2022, PT VI, 2022, 13436 :354-364
[8]   Proximal alternating linearized minimization for nonconvex and nonsmooth problems [J].
Bolte, Jerome ;
Sabach, Shoham ;
Teboulle, Marc .
MATHEMATICAL PROGRAMMING, 2014, 146 (1-2) :459-494
[9]   Deterministic edge-preserving regularization in computed imaging [J].
Charbonnier, P ;
BlancFeraud, L ;
Aubert, G ;
Barlaud, M .
IEEE TRANSACTIONS ON IMAGE PROCESSING, 1997, 6 (02) :298-311
[10]   Learnable Descent Algorithm for Nonsmooth Nonconvex Image Reconstruction [J].
Chen, Yunmei ;
Liu, Hongcheng ;
Ye, Xiaojing ;
Zhang, Qingchao .
SIAM JOURNAL ON IMAGING SCIENCES, 2021, 14 (04) :1532-1564