A stochastic first-order trust-region method with inexact restoration for finite-sum minimization

被引:0
|
作者
Stefania Bellavia
Nataša Krejić
Benedetta Morini
Simone Rebegoldi
机构
[1] Università degli Studi di Firenze,Dipartimento di Ingegneria Industriale
[2] University of Novi Sad,Department of Mathematics and Informatics, Faculty of Sciences
关键词
Finite-sum minimization; Inexact restoration; Trust-region methods; Subsampling; Worst-case iteration complexity;
D O I
暂无
中图分类号
学科分类号
摘要
We propose a stochastic first-order trust-region method with inexact function and gradient evaluations for solving finite-sum minimization problems. Using a suitable reformulation of the given problem, our method combines the inexact restoration approach for constrained optimization with the trust-region procedure and random models. Differently from other recent stochastic trust-region schemes, our proposed algorithm improves feasibility and optimality in a modular way. We provide the expected number of iterations for reaching a near-stationary point by imposing some probability accuracy requirements on random functions and gradients which are, in general, less stringent than the corresponding ones in literature. We validate the proposed algorithm on some nonconvex optimization problems arising in binary classification and regression, showing that it performs well in terms of cost and accuracy, and allows to reduce the burdensome tuning of the hyper-parameters involved.
引用
收藏
页码:53 / 84
页数:31
相关论文
共 50 条
  • [31] A trust-region method for stochastic variational inference with applications to streaming data
    Theis, Lucas
    Hoffman, Matthew D.
    INTERNATIONAL CONFERENCE ON MACHINE LEARNING, VOL 37, 2015, 37 : 2503 - 2511
  • [32] Efficient Decentralized Stochastic Gradient Descent Method for Nonconvex Finite-Sum Optimization Problems
    Zhan, Wenkang
    Wu, Gang
    Gao, Hongchang
    THIRTY-SIXTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE / THIRTY-FOURTH CONFERENCE ON INNOVATIVE APPLICATIONS OF ARTIFICIAL INTELLIGENCE / TWELVETH SYMPOSIUM ON EDUCATIONAL ADVANCES IN ARTIFICIAL INTELLIGENCE, 2022, : 9006 - 9013
  • [33] Method with batching for stochastic finite-sum variational inequalities in non-Euclidean setting
    Pichugin, Alexander
    Pechin, Maksim
    Beznosikov, Aleksandr
    Novitskii, Vasilii
    Gasnikov, Alexander
    CHAOS SOLITONS & FRACTALS, 2024, 187
  • [34] Higher order weighted integral stochastic finite element method and simplified first-order application
    Noh, Hyuk-Chun
    Lee, Phill-Seung
    INTERNATIONAL JOURNAL OF SOLIDS AND STRUCTURES, 2007, 44 (11-12) : 4120 - 4144
  • [35] A Nonlinear Conjugate Gradient Method Using Inexact First-Order Information
    Tiantian Zhao
    Wei Hong Yang
    Journal of Optimization Theory and Applications, 2023, 198 : 502 - 530
  • [36] A Nonlinear Conjugate Gradient Method Using Inexact First-Order Information
    Zhao, Tiantian
    Yang, Wei Hong
    JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS, 2023, 198 (02) : 502 - 530
  • [37] Minimization of the Ginzburg-Landau energy functional by a Sobolev gradient trust-region method
    Kazemi, P.
    Renka, R. J.
    APPLIED MATHEMATICS AND COMPUTATION, 2013, 219 (11) : 5936 - 5942
  • [38] A method for convex minimization based on translated first-order approximations
    A. Astorino
    M. Gaudioso
    E. Gorgone
    Numerical Algorithms, 2017, 76 : 745 - 760
  • [39] A method for convex minimization based on translated first-order approximations
    Astorino, A.
    Gaudioso, M.
    Gorgone, E.
    NUMERICAL ALGORITHMS, 2017, 76 (03) : 745 - 760
  • [40] Cubic-regularization counterpart of a variable-norm trust-region method for unconstrained minimization
    Martinez, J. M.
    Raydan, M.
    JOURNAL OF GLOBAL OPTIMIZATION, 2017, 68 (02) : 367 - 385