Self-Concordant Analysis of Frank-Wolfe Algorithms

被引:0
|
作者
Dvurechensky, Pavel [1 ,2 ,3 ]
Ostroukhov, Petr [4 ]
Safin, Kamil [4 ]
Shtern, Shimrit [5 ]
Staudigl, Mathias [6 ]
机构
[1] Weierstrass Inst Appl Anal & Stochast, Berlin, Germany
[2] Natl Res Univ Higher Sch Econ, Moscow, Russia
[3] Inst Informat Transmiss Problems RAS, Moscow, Russia
[4] Moscow Inst Phys & Technol, Dolgoprudnyi, Russia
[5] Technion Israel Inst Technol, Haifa, Israel
[6] Maastricht Univ, Dept Data Sci & Knowledge Engn, Maastricht, Netherlands
来源
25TH AMERICAS CONFERENCE ON INFORMATION SYSTEMS (AMCIS 2019) | 2019年
关键词
OPTIMIZATION; CONVERGENCE; COMPLEXITY;
D O I
暂无
中图分类号
TP [自动化技术、计算机技术];
学科分类号
0812 ;
摘要
Projection-free optimization via different variants of the Frank-Wolfe (FW), a.k.a. Conditional Gradient method has become one of the cornerstones in optimization for machine learning since in many cases the linear minimization oracle is much cheaper to implement than projections and some sparsity needs to be preserved. In a number of applications, e.g. Poisson inverse problems or quantum state tomography, the loss is given by a self-concordant (SC) function having unbounded curvature, implying absence of theoretical guarantees for the existing FW methods. We use the theory of SC functions to provide a new adaptive step size for FW methods and prove global convergence rate O(1/k) after k iterations. If the problem admits a stronger local linear minimization oracle, we construct a novel FW method with linear convergence rate for SC functions.
引用
收藏
页数:11
相关论文
共 50 条
  • [31] Analysis of the Frank-Wolfe method for convex composite optimization involving a logarithmically-homogeneous barrier
    Zhao, Renbo
    Freund, Robert M.
    MATHEMATICAL PROGRAMMING, 2023, 199 (1-2) : 123 - 163
  • [32] A novel Frank-Wolfe algorithm. Analysis and applications to large-scale SVM training
    Nancuef, Ricardo
    Frandi, Emanuele
    Sartori, Claudio
    Allende, Hector
    INFORMATION SCIENCES, 2014, 285 : 66 - 99
  • [33] Frank-Wolfe and friends: a journey into projection-free first-order optimization methods
    Bomze, Immanuel. M.
    Rinaldi, Francesco
    Zeffiro, Damiano
    ANNALS OF OPERATIONS RESEARCH, 2024, 343 (02) : 607 - 638
  • [34] The sliding Frank-Wolfe algorithm and its application to super-resolution microscopy
    Denoyelle, Quentin
    Duval, Vincent
    Peyre, Gabriel
    Soubies, Emmanuel
    INVERSE PROBLEMS, 2020, 36 (01)
  • [35] A Frank-Wolfe Progressive Hedging Algorithm for Improved Lower Bounds Stochastic SCUC
    Palani, Ananth M.
    Wu, Hongyu
    Morcos, Medhat M.
    IEEE ACCESS, 2019, 7 : 99398 - 99406
  • [36] Optimal Influencer Marketing Campaign Under Budget Constraints Using Frank-Wolfe
    Lopez-Dawn, Ricardo
    Giovanidis, Anastasios
    IEEE TRANSACTIONS ON NETWORK SCIENCE AND ENGINEERING, 2023, 10 (02): : 1015 - 1031
  • [37] Faster Rates for the Frank-Wolfe Method over Strongly-Convex Sets
    Garber, Dan
    Hazan, Elad
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 37, 2015, 37 : 541 - 549
  • [38] Self-concordant inclusions: a unified framework for path-following generalized Newton-type algorithms
    Quoc Tran-Dinh
    Sun, Tianxiao
    Lu, Shu
    MATHEMATICAL PROGRAMMING, 2019, 177 (1-2) : 173 - 223
  • [39] Self-Concordant Barriers for Convex Approximations of Structured Convex Sets
    Tuncel, Levent
    Nemirovski, Arkadi
    FOUNDATIONS OF COMPUTATIONAL MATHEMATICS, 2010, 10 (05) : 485 - 525
  • [40] No Self-Concordant Barrier Interior Point Method Is Strongly Polynomial
    Allamigeon, Xavier
    Gaubert, Stephane
    Vandame, Nicolas
    PROCEEDINGS OF THE 54TH ANNUAL ACM SIGACT SYMPOSIUM ON THEORY OF COMPUTING (STOC '22), 2022, : 515 - 528