English

Min-Max Optimization Requires Exponentially Many Queries

Data Structures and Algorithms 2026-05-14 v1 Computational Complexity Computer Science and Game Theory Machine Learning Optimization and Control

Abstract

We study the query complexity of min-max optimization of a nonconvex-nonconcave function ff over [0,1]d×[0,1]d[0,1]^d \times [0,1]^d. We show that, given oracle access to ff and to its gradient f\nabla f, any algorithm that finds an ε\varepsilon-approximate stationary point must make a number of queries that is exponential in 1/ε1/\varepsilon or dd.

Keywords

Cite

@article{arxiv.2605.13806,
  title  = {Min-Max Optimization Requires Exponentially Many Queries},
  author = {Martino Bernasconi and Matteo Castiglioni and Andrea Celli and Alexandros Hollender},
  journal= {arXiv preprint arXiv:2605.13806},
  year   = {2026}
}