Linearly convergent bilevel optimization with single-step inner methods

被引:1
作者
Suonpera, Ensio [1 ]
Valkonen, Tuomo [1 ,2 ]
机构
[1] Univ Helsinki, Dept Math & Stat, Helsinki, Finland
[2] Escuela Politec Nacl, ModeMat, Quito, Ecuador
基金
芬兰科学院;
关键词
Bilevel optimization; Nonsmooth; Inverse problems; Forward-backward;
D O I
10.1007/s10589-023-00527-7
中图分类号
C93 [管理学]; O22 [运筹学];
学科分类号
070105 ; 12 ; 1201 ; 1202 ; 120202 ;
摘要
We propose a new approach to solving bilevel optimization problems, intermediate between solving full-system optimality conditions with a Newton-type approach, and treating the inner problem as an implicit function. The overall idea is to solve the full-system optimality conditions, but to precondition them to alternate between taking steps of simple conventional methods for the inner problem, the adjoint equation, and the outer problem. While the inner objective has to be smooth, the outer objective may be nonsmooth subject to a prox-contractivity condition. We prove linear convergence of the approach for combinations of gradient descent and forward-backward splitting with exact and inexact solution of the adjoint equation. We demonstrate good performance on learning the regularization parameter for anisotropic total variation image denoising, and the convolution kernel for image deconvolution.
引用
收藏
页码:571 / 610
页数:40
相关论文
共 55 条
  • [41] Luo Z.-Q., 1996, Mathematical Programs with Equilibrium Constraints, DOI [DOI 10.1017/CBO9780511983658, 10.1017/CBO9780511983658]
  • [42] Sufficient Optimality Conditions in Bilevel Programming
    Mehlitz, Patrick
    Zemkoho, Alain B.
    [J]. MATHEMATICS OF OPERATIONS RESEARCH, 2021, 46 (04) : 1573 - 1598
  • [43] Techniques for Gradient-Based Bilevel Optimization with Non-smooth Lower Level Problems
    Ochs, Peter
    Ranftl, Rene
    Brox, Thomas
    Pock, Thomas
    [J]. JOURNAL OF MATHEMATICAL IMAGING AND VISION, 2016, 56 (02) : 175 - 194
  • [44] A FIRST ORDER METHOD FOR SOLVING CONVEX BILEVEL OPTIMIZATION PROBLEMS
    Sabach, Shoham
    Shtern, Shimrit
    [J]. SIAM JOURNAL ON OPTIMIZATION, 2017, 27 (02) : 640 - 660
  • [45] An inertial extrapolation method for convex simple bilevel optimization
    Shehu, Yekini
    Phan Tu Vuong
    Zemkoho, Alain
    [J]. OPTIMIZATION METHODS & SOFTWARE, 2021, 36 (01) : 1 - 19
  • [46] Learning the Sampling Pattern for MRI
    Sherry, Ferdia
    Benning, Martin
    De los Reyes, Juan Carlos
    Graves, Martin J.
    Maierhofer, Georg
    Williams, Guy
    Schonlieb, Carola-Bibiane
    Ehrhardt, Matthias J.
    [J]. IEEE TRANSACTIONS ON MEDICAL IMAGING, 2020, 39 (12) : 4310 - 4321
  • [47] StephanDempe A.Z., 2020, BILEVEL OPTIMIZATION
  • [48] Suonpera E., 2023, CODES LINEARLY CONVE, DOI DOI 10.5281/ZENODO.7974062
  • [49] Testing and Non-linear Preconditioning of the Proximal Point Method
    Valkonen, Tuomo
    [J]. APPLIED MATHEMATICS AND OPTIMIZATION, 2020, 82 (02) : 591 - 636
  • [50] A primal-dual hybrid gradient method for nonlinear operators with applications to MRI
    Valkonen, Tuomo
    [J]. INVERSE PROBLEMS, 2014, 30 (05)