Solving strongly convex-concave composite saddle point problems with a small dimension of one of the variables
- 作者: Alkousa M.S.1, Gasnikov A.V.1,2,3,4, Gladin E.L.5, Kuruzov I.A.1, Pasechnyuk D.A.1, Stonyakin F.S.1,6
-
隶属关系:
- Moscow Institute of Physics and Technology (National Research University)
- Ivannikov Institute for System Programming of the RAS
- Institute for Information Transmission Problems of the Russian Academy of Sciences (Kharkevich Institute)
- Caucasus Mathematical Center, Adyghe State University
- Humboldt University
- V. I. Vernadsky Crimean Federal University
- 期: 卷 214, 编号 3 (2023)
- 页面: 3-53
- 栏目: Articles
- URL: https://bakhtiniada.ru/0368-8666/article/view/133505
- DOI: https://doi.org/10.4213/sm9700
- ID: 133505
如何引用文章
详细
Algorithmic methods are developed that guarantee efficient complexity estimates for strongly convex-concave saddle-point problems in the case when one group of variables has a high dimension, while another has a rather low dimension (up to 100). These methods are based on reducing problems of this type to the minimization (maximization) problem for a convex (concave) functional with respect to one of the variables such that an approximate value of the gradient at an arbitrary point can be obtained with the required accuracy using an auxiliary optimization subproblem with respect to the other variable. It is proposed to use the ellipsoid method and Vaidya's method for low-dimensional problems and accelerated gradient methods with inexact information about the gradient or subgradient for high-dimensional problems. In the case when one group of variables, ranging over a hypercube, has a very low dimension (up to five), another proposed approach to strongly convex-concave saddle-point problems is rather efficient. This approach is based on a new version of a multidimensional analogue of Nesterov's method on a square (the multidimensional dichotomy method) with the possibility to use inexact values of the gradient of the objective functional.
作者简介
Mohammad Alkousa
Moscow Institute of Physics and Technology (National Research University)
Email: mohammad.alkousa@phystech.edu
Alexander Gasnikov
Moscow Institute of Physics and Technology (National Research University); Ivannikov Institute for System Programming of the RAS; Institute for Information Transmission Problems of the Russian Academy of Sciences (Kharkevich Institute); Caucasus Mathematical Center, Adyghe State University
Email: gasnikov@yandex.ru
Doctor of physico-mathematical sciences, Associate professor
Egor Gladin
Humboldt University
Email: mohammad.alkousa@phystech.edu
Ilya Kuruzov
Moscow Institute of Physics and Technology (National Research University)
Email: mohammad.alkousa@phystech.edu
Dmitry Pasechnyuk
Moscow Institute of Physics and Technology (National Research University)
Email: pasechniuk.da@phystech.edu
without scientific degree, no status
Fedor Stonyakin
Moscow Institute of Physics and Technology (National Research University); V. I. Vernadsky Crimean Federal University
编辑信件的主要联系方式.
Email: fedyor@mail.ru
Candidate of physico-mathematical sciences, Associate professor
参考
- М. С. Алкуса, А. В. Гасников, Д. М. Двинских, Д. А. Ковалев, Ф. С. Стонякин, “Ускоренные методы для седловых задач”, Ж. вычисл. матем. и матем. физ., 60:11 (2020), 1843–1866
- W. Azizian, I. Mitliagkas, S. Lacoste-Julien, G. Gidel, “A tight and unified analysis of gradient-based methods for a whole spectrum of differentiable games”, Proceedings of the twenty third international conference on artificial intelligence and statistics, Proceedings of Machine Learning Research (PMLR), 108, 2020, 2863–2873
- А. В. Гасников, П. Е. Двуреченский, Ю. Е. Нестеров, “Стохастические градиентные методы с неточным оракулом”, Труды МФТИ, 8:1 (2016), 41–91
- А. В. Гасников, Д. М. Двинских, П. Е. Двуреченский, Д. И. Камзолов, В. В. Матюхин, Д. А. Пасечнюк, Н. К. Тупица, А. В. Чернов, “Ускоренный метаалгоритм для задач выпуклой оптимизации”, Ж. вычисл. матем. и матем. физ., 61:1 (2021), 20–31
- Le Thi Khanh Hien, Renbo Zhao, W. B. Haskell, An inexact primal-dual framework for large-scale non-bilinear saddle point problem
- Guanghui Lan, First-order and stochastic optimization methods for machine learning, Springer Ser. Data Sci., Springer, Cham, 2020, xiii+582 pp.
- Tianyi Lin, Chi Jin, M. I. Jordan, “Near-optimal algorithms for minimax optimization”, Proceedings of thirty third conference on learning theory, Proceedings of Machine Learning Research (PMLR), 125, 2020, 2738–2779
- I. Usmanova, M. Kamgarpour, A. Krause, K. Levy, “Fast projection onto convex smooth constraints”, Proceedings of the 38th international conference on machine learning, Proceedings of Machine Learning Research (PMLR), 139, 2021, 10476–10486
- А. В. Гасников, Современные численные методы оптимизации. Метод универсального градиентного спуска, 2-е испр. изд., МЦНМО, М., 2021, 272 с.
- B. Cox, A. Juditsky, A. Nemirovski, “Decomposition techniques for bilinear saddle point problems and variational inequalities with affine monotone operators”, J. Optim. Theory Appl., 172:2 (2017), 402–435
- Yu. Nesterov, “Excessive gap technique in nonsmooth convex minimization”, SIAM J. Optim., 16:1 (2005), 235–249
- Junyu Zhang, Mingyi Hong, Shuzhong Zhang, “On lower iteration complexity bounds for the convex concave saddle point problems”, Math. Program., 194:1-2, Ser. A (2022), 901–935
- Yuanhao Wang, Jian Li, “Improved algorithms for convex-concave minimax optimization”, NIPS 2020, Adv. Neural Inf. Process. Syst., 33, MIT Press, Cambridge, MA, 2020, 11 pp.
- A. Nemirovski, S. Onn, U. G. Rothblum, “Accuracy certificates for computational problems with convex structure”, Math. Oper. Res., 35:1 (2010), 52–78
- A. Nemirovski, Information-based complexity of convex programming, Lecture notes, 1995, 268 pp.
- Д. А. Пасечнюк, Ф. С. Стонякин, “Об одном методе минимизации выпуклой липшицевой функции двух переменных на квадрате”, Компьютерные исследования и моделирование, 11:3 (2019), 379–395
- Б. Т. Поляк, Введение в оптимизацию, Наука, М., 1983, 384 с.
- O. Devolder, F. Glineur, Yu. Nesterov, “First-order methods of smooth convex optimization with inexact oracle”, Math. Program., 146:1-2, Ser. A (2014), 37–75
- P. M. Vaidya, “A new algorithm for minimizing convex functions over convex sets”, Math. Program., 73:3, Ser. A (1996), 291–341
- S. Bubeck, “Convex optimization: algorithms and complexity”, Found. Trends Mach. Learn., 8:3-4 (2015), 231–357
- E. Gladin, A. Sadiev, A. Gasnikov, P. Dvurechensky, A. Beznosikov, M. Alkousa, “Solving smooth min-min and min-max problems by mixed oracle algorithms”, MOTOR 2021: Mathematical optimization theory and operations research–recent trends, Commun. Comput. Inf. Sci., 1476, Springer, Cham, 2021, 19–40
- А. В. Гасников, Ю. Е. Нестеров, “Универсальный метод для задач стохастической композитной оптимизации”, Ж. вычисл. матем. и матем. физ., 58:1 (2018), 52–69
- J. M. Danskin, “The theory of Max-Min, with applications”, SIAM J. Appl. Math., 14:4 (1966), 641–664
- O. Devolder, Exactness, inexactness and stochasticity in first-order methods for large-scale convex optimization, PhD thesis, UCL, CORE, Louvain-la-Neuve, 2013, vii+309 pp.
- O. Devolder, F. Glineur, Yu. Nesterov, First-order methods with inexact oracle: the strongly convex case, CORE Discussion Papers, no. 2013/16, 2013, 35 pp.
- Н. З. Шор, Методы минимизации недифференцируемых функций и их приложения, Наукова думка, Киев, 1979, 199 с.
- Исходный код экспериментов на GitHub
- P. Bernhard, A. Rapaport, “On a theorem of Danskin with an application to a theorem of von Neumann–Sion”, Nonlinear Anal., 24:8 (1995), 1163–1181
补充文件
