An accelerated proximal algorithm for regularized nonconvex and nonsmooth bi-level optimization

被引:1
作者
Chen, Ziyi [1 ]
Kailkhura, Bhavya [2 ]
Zhou, Yi [1 ]
机构
[1] Univ Utah, Elect & Comp Dept, 50 Cent Campus Dr 2110, Salt Lake City, UT 84112 USA
[2] Lawrence Livermore Natl Lab, 7000 East Ave, Livermore, CA 10587 USA
基金
美国国家科学基金会;
关键词
Bilevel optimization; Nesterov's momentum; Nonconvex regularization; Proximal algorithm; CONVERGENCE;
D O I
10.1007/s10994-023-06329-6
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Many important machine learning applications involve regularized nonconvex bi-level optimization. However, the existing gradient-based bi-level optimization algorithms cannot handle nonconvex or nonsmooth regularizers, and they suffer from a high computation complexity in nonconvex bi-level optimization. In this work, we study a proximal gradient-type algorithm that adopts the approximate implicit differentiation (AID) scheme for nonconvex bi-level optimization with possibly nonconvex and nonsmooth regularizers. In particular, the algorithm applies the Nesterov's momentum to accelerate the computation of the implicit gradient involved in AID. We provide a comprehensive analysis of the global convergence properties of this algorithm through identifying its intrinsic potential function. In particular, we formally establish the convergence of the model parameters to a critical point of the bi-level problem, and obtain an improved computation complexity (O) over tilde(?(3.5)e(-2))over the state-of-the-art result. Moreover, we analyze the asymptotic convergence rates of this algorithm under a class of local nonconvex geometries characterized by a Lojasiewicz-type gradient inequality. Experiment on hyper-parameter optimization demonstrates the effectiveness of our algorithm.
引用
收藏
页码:1433 / 1463
页数:31
相关论文
共 50 条
  • [21] An accelerated stochastic ADMM for nonconvex and nonsmooth finite-sum optimization
    Zeng, Yuxuan
    Wang, Zhiguo
    Bai, Jianchao
    Shen, Xiaojing
    AUTOMATICA, 2024, 163
  • [22] Convergence and rate analysis of a proximal linearized ADMM for nonconvex nonsmooth optimization
    Maryam Yashtini
    Journal of Global Optimization, 2022, 84 : 913 - 939
  • [23] A Proximal Algorithm for Optimizing Compositions of Quadratic plus Nonconvex Nonsmooth Functions
    Zhou, Yiming
    Dai, Wei
    32ND EUROPEAN SIGNAL PROCESSING CONFERENCE, EUSIPCO 2024, 2024, : 2627 - 2631
  • [24] A proximal bundle method for constrained nonsmooth nonconvex optimization with inexact information
    Lv, Jian
    Pang, Li-Ping
    Meng, Fan-Yun
    JOURNAL OF GLOBAL OPTIMIZATION, 2018, 70 (03) : 517 - 549
  • [25] Proximal Linearized Iteratively Reweighted Algorithms for Nonconvex and Nonsmooth Optimization Problem
    Yeo, Juyeb
    Kang, Myeongmin
    AXIOMS, 2022, 11 (05)
  • [26] Convergence and rate analysis of a proximal linearized ADMM for nonconvex nonsmooth optimization
    Yashtini, Maryam
    JOURNAL OF GLOBAL OPTIMIZATION, 2022, 84 (04) : 913 - 939
  • [27] Transfer Learning Based Evolutionary Algorithm for Bi-level Optimization Problems
    Chen, Lei
    Liu, Hai-Lin
    2021 IEEE CONGRESS ON EVOLUTIONARY COMPUTATION (CEC 2021), 2021, : 1643 - 1647
  • [28] Bi-level Problem and SMD Assessment Delinquent for Single Impartial Bi-level Optimization
    Vadali, Srinivas
    Deekshitulu, G. V. S. R.
    Murthy, J. V. R.
    PROCEEDINGS OF SIXTH INTERNATIONAL CONFERENCE ON SOFT COMPUTING FOR PROBLEM SOLVING (SOCPROS 2016), VOL 1, 2017, 546 : 52 - 62
  • [29] A SEQUENTIAL QUADRATIC PROGRAMMING ALGORITHM FOR NONCONVEX, NONSMOOTH CONSTRAINED OPTIMIZATION
    Curtis, Frank E.
    Overton, Michael L.
    SIAM JOURNAL ON OPTIMIZATION, 2012, 22 (02) : 474 - 500
  • [30] Evolutionary multitasking in bi-level optimization
    Gupta, Abhishek
    Mandziuk, Jacek
    Ong, Yew-Soon
    COMPLEX & INTELLIGENT SYSTEMS, 2015, 1 (1-4) : 83 - 95