Exponential Convergence of Gradient Methods in Concave Network Zero-Sum Games

• Published in: Machine Learning and Knowledge Discovery in Databases Vol. 12458; pp. 19 - 34
• Main Authors: Kadan, Amit, Fu, Hu
• Format: Book Chapter
• Language: English
• Published: Switzerland Springer International Publishing AG 2021; Springer International Publishing
• Series: Lecture Notes in Computer Science
• Subjects: Convergence of gradient ascent; Generative adversarial networks; Last iterate convergence; Network zero-sum games
• ISBN: 3030676609; 9783030676605
• ISSN: 0302-9743; 1611-3349
• DOI: 10.1007/978-3-030-67661-2_2

Motivated by Generative Adverserial Networks, we study the computation of Nash equilibrium in concave network zero-sum games (NZSGs), a multiplayer generalization of two-player zero-sum games first proposed with linear payoffs. Extending previous results, we show that various game theoretic properties of convex-concave two-player zero-sum games are preserved in this generalization. We then generalize last iterate convergence results obtained previously in two-player zero-sum games. We analyze convergence rates when players update their strategies using Gradient Ascent, and its variant, Optimistic Gradient Ascent, showing last iterate convergence in three settings—when the payoffs of players are linear, strongly concave and Lipschitz, and strongly concave and smooth. We provide experimental results that support these theoretical findings.