English

Conic Blackwell Algorithm: Parameter-Free Convex-Concave Saddle-Point Solving

Machine Learning 2021-10-15 v3 Computer Science and Game Theory

Abstract

We develop new parameter-free and scale-free algorithms for solving convex-concave saddle-point problems. Our results are based on a new simple regret minimizer, the Conic Blackwell Algorithm+^+ (CBA+^+), which attains O(1/T)O(1/\sqrt{T}) average regret. Intuitively, our approach generalizes to other decision sets of interest ideas from the Counterfactual Regret minimization (CFR+^+) algorithm, which has very strong practical performance for solving sequential games on simplexes. We show how to implement CBA+^+ for the simplex, p\ell_{p} norm balls, and ellipsoidal confidence regions in the simplex, and we present numerical experiments for solving matrix games and distributionally robust optimization problems. Our empirical results show that CBA+^+ is a simple algorithm that outperforms state-of-the-art methods on synthetic data and real data instances, without the need for any choice of step sizes or other algorithmic parameters.

Keywords

Cite

@article{arxiv.2105.13203,
  title  = {Conic Blackwell Algorithm: Parameter-Free Convex-Concave Saddle-Point Solving},
  author = {Julien Grand-Clément and Christian Kroer},
  journal= {arXiv preprint arXiv:2105.13203},
  year   = {2021}
}
R2 v1 2026-06-24T02:31:57.817Z