Sublinear convergence rates of extragradient-type methods: A survey on classical and recent developments

Q Tran-Dinh - arXiv preprint arXiv:2303.17192, 2023 - arxiv.org
The extragradient (EG), introduced by GM Korpelevich in 1976, is a well-known method to
approximate solutions of saddle-point problems and their extensions such as variational …

Convergence of proximal point and extragradient-based methods beyond monotonicity: the case of negative comonotonicity

E Gorbunov, A Taylor, S Horváth… - … on Machine Learning, 2023 - proceedings.mlr.press
Algorithms for min-max optimization and variational inequalities are often studied under
monotonicity assumptions. Motivated by non-monotone machine learning applications, we …

Last-iterate convergent policy gradient primal-dual methods for constrained mdps

D Ding, CY Wei, K Zhang… - Advances in Neural …, 2024 - proceedings.neurips.cc
We study the problem of computing an optimal policy of an infinite-horizon discounted
constrained Markov decision process (constrained MDP). Despite the popularity of …

Stable nonconvex-nonconcave training via linear interpolation

T Pethick, W Xie, V Cevher - Advances in Neural …, 2024 - proceedings.neurips.cc
This paper presents a theoretical analysis of linear interpolation as a principled method for
stabilizing (large-scale) neural network training. We argue that instabilities in the …

Accelerated single-call methods for constrained min-max optimization

Y Cai, W Zheng - arXiv preprint arXiv:2210.03096, 2022 - arxiv.org
We study first-order methods for constrained min-max optimization. Existing methods either
require two gradient calls or two projections in each iteration, which may be costly in some …

Solving nonconvex-nonconcave min-max problems exhibiting weak minty solutions

A Böhm - arXiv preprint arXiv:2201.12247, 2022 - arxiv.org
We investigate a structured class of nonconvex-nonconcave min-max problems exhibiting
so-called\emph {weak Minty} solutions, a notion which was only recently introduced, but is …

Doubly optimal no-regret learning in monotone games

Y Cai, W Zheng - International Conference on Machine …, 2023 - proceedings.mlr.press
We consider online learning in multi-player smooth monotone games. Existing algorithms
have limitations such as (1) being only applicable to strongly monotone games;(2) lacking …

On the interplay between social welfare and tractability of equilibria

I Anagnostides, T Sandholm - Advances in Neural …, 2024 - proceedings.neurips.cc
Computational tractability and social welfare (aka. efficiency) of equilibria are two
fundamental but in general orthogonal considerations in algorithmic game theory …

Universal gradient descent ascent method for nonconvex-nonconcave minimax optimization

T Zheng, L Zhu, AMC So… - Advances in Neural …, 2023 - proceedings.neurips.cc
Nonconvex-nonconcave minimax optimization has received intense attention over the last
decade due to its broad applications in machine learning. Most existing algorithms rely on …

Variance reduced halpern iteration for finite-sum monotone inclusions

X Cai, A Alacaoglu, J Diakonikolas - arXiv preprint arXiv:2310.02987, 2023 - arxiv.org
Machine learning approaches relying on such criteria as adversarial robustness or multi-
agent settings have raised the need for solving game-theoretic equilibrium problems. Of …