A J-symmetric quasi-newton method for minimax problems

被引:1
作者
Asl, Azam [1 ]
Lu, Haihao [1 ]
Yang, Jinwen [2 ]
机构
[1] Univ Chicago Booth Sch Business, Chicago, IL 60611 USA
[2] Univ Chicago, Dept Stat, Chicago, IL USA
关键词
Minimax optimization; Quasi-Newton method; J-symmetric; Superlinear convergence; Trust region; PROXIMAL POINT ALGORITHM; SUPERLINEAR CONVERGENCE; GLOBAL CONVERGENCE; MONOTONE-OPERATORS;
D O I
10.1007/s10107-023-01957-1
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
Minimax problems have gained tremendous attentions across the optimization and machine learning community recently. In this paper, we introduce a new quasi-Newton method for the minimax problems, which we call J-symmetric quasi-Newton method. The method is obtained by exploiting the J-symmetric structure of the second-order derivative of the objective function in minimax problem. We show that the Hessian estimation (as well as its inverse) can be updated by a rank-2 operation, and it turns out that the update rule is a natural generalization of the classic Powell symmetric Broyden method from minimization problems to minimax problems. In theory, we show that our proposed quasi-Newton algorithm enjoys local Q-superlinear convergence to a desirable solution under standard regularity conditions. Furthermore, we introduce a trust-region variant of the algorithm that enjoys global R-superlinear convergence. Finally, we present numerical experiments that verify our theory and show the effectiveness of our proposed algorithms compared to Broyden's method and the extragradient method on three classes of minimax problems.
引用
收藏
页码:207 / 254
页数:48
相关论文
共 64 条
[1]   A globally convergent BFGS method for pseudo-monotone variational inequality problems [J].
Abdi, Fatemeh ;
Shakeri, Fatemeh .
OPTIMIZATION METHODS & SOFTWARE, 2019, 34 (01) :25-36
[2]  
[Anonymous], 1976, MATEKON, V12
[3]  
Applegate D., 2021, ARXIV
[4]  
Arjovsky M, 2017, PR MACH LEARN RES, V70
[5]   Analysis of the gradient method with an Armijo-Wolfe line search on a class of non-smooth convex functions [J].
Asl, Azam ;
Overton, Michael L. .
OPTIMIZATION METHODS & SOFTWARE, 2020, 35 (02) :223-242
[6]   A CUTTING PLANE ALGORITHM FOR CONVEX-PROGRAMMING THAT USES ANALYTIC CENTERS [J].
ATKINSON, DS ;
VAIDYA, PM .
MATHEMATICAL PROGRAMMING, 1995, 69 (01) :1-43
[7]  
Benzi M, 2005, ACTA NUMER, V14, P1, DOI 10.1017/S0962492904000212
[8]   A preconditioner for generalized saddle point problems [J].
Benzi, M ;
Golub, GH .
SIAM JOURNAL ON MATRIX ANALYSIS AND APPLICATIONS, 2004, 26 (01) :20-41
[9]  
Berthold T., 2017, OPERAT RES PROCEED, V2018, P159
[10]  
Bertsekas D. P., 1982, Constrained Optimization and Lagrange Multiplier Methods, DOI DOI 10.1016/C2013-0-10366-2