No-regret algorithms in on-line learning, games and convex optimization

被引:1
作者
Sorin, Sylvain [1 ]
机构
[1] Sorbonne Univ, Inst Math Jussieu, CNRS UMR 7586, PRG, Campus Pierre Marie Curie, Paris, France
关键词
68T05; 68W27; 68W50; 90C25; 91A26; PROJECTED SUBGRADIENT METHODS; DYNAMICAL-SYSTEMS; 1ST-ORDER METHODS; VARIATIONAL-INEQUALITIES; MONOTONE-OPERATORS; GRADIENT; CONVERGENCE; DESCENT; APPROXIMATIONS; MINIMIZATION;
D O I
10.1007/s10107-023-01927-7
中图分类号
TP31 [计算机软件];
学科分类号
081202 ; 0835 ;
摘要
The purpose of this article is to underline the links between some no-regret algorithms used in on-line learning, games and convex optimization and to compare the continuous and discrete time versions.
引用
收藏
页码:645 / 686
页数:42
相关论文
共 107 条
[1]  
Abernethy J. D., 2011, P MACHINE LEARNING R, V19, P27
[2]   Hessian Riemannian gradient flows in convex programming [J].
Alvarez, F ;
Bolte, J ;
Brahic, O .
SIAM JOURNAL ON CONTROL AND OPTIMIZATION, 2004, 43 (02) :477-501
[3]   Convex Optimization: Algorithms and Complexity [J].
不详 .
FOUNDATIONS AND TRENDS IN MACHINE LEARNING, 2015, 8 (3-4) :232-+
[4]  
[Anonymous], 1954, Proceedings of the International Congress of Mathematicians
[5]  
[Anonymous], 1966, USSR Comput Math Math Phys, DOI DOI 10.1016/0041-5553(66)90114-5
[6]   Singular Riemannian Barrier methods and gradient-projection dynamical systems for constrained optimization [J].
Attouch, H ;
Bolte, J ;
Redont, P ;
Teboulle, M .
OPTIMIZATION, 2004, 53 (5-6) :435-454
[7]   Regularized Lotka-Volterra dynamical system as continuous proximal-like method in optimization [J].
Attouch, H ;
Teboulle, M .
JOURNAL OF OPTIMIZATION THEORY AND APPLICATIONS, 2004, 121 (03) :541-570
[8]   Fast convergence of inertial dynamics and algorithms with asymptotic vanishing viscosity [J].
Attouch, Hedy ;
Chbani, Zaki ;
Peypouquet, Juan ;
Redont, Patrick .
MATHEMATICAL PROGRAMMING, 2018, 168 (1-2) :123-175
[9]  
Aubin Jean-Pierre., 1984, Set-valued maps and viability theory, V264, pxiii+342
[10]  
Auer P, 2003, SIAM J COMPUT, V32, P48, DOI 10.1137/S0097539701398375