GANGs: Generative Adversarial Network Games

Abstract

Generative Adversarial Networks (GAN) have become one of the most successful frameworks for unsupervised generative modeling. As GANs are difficult to train much research has focused on this. However, very little of this research has directly exploited gametheoretic techniques. We introduce Generative Adversarial Network Games (GANGs), which explicitly model a finite zero-sum game between a generator (G) and classifier (C) that use mixed strategies. The size of these games precludes exact solution methods, therefore we define resource-bounded best responses (RBBRs), and a resourcebounded Nash Equilibrium (RB-NE) as a pair of mixed strategies such that neither G or C can find a better RBBR. The RB-NE solution concept is richer than the notion of 'local Nash equilibria' in that it captures not only failures of escaping local optima of gradient descent, but applies to any approximate best response computations, including methods with random restarts. To validate our approach, we solve GANGs with the Parallel Nash Memory algorithm, which provably monotonically converges to an RB-NE. We compare our results to standard GAN setups, and demonstrate that our method deals well with typical GAN problems such as mode collapse, partial mode coverage and forgetting.

Generative Adversarial Networks and Adversarial Autoencoders: Tutorial and Survey

“…Moreover, some papers have specifically combined GAN with game theory. Interested readers can refer to (Oliehoek et al, 2017;Arora et al, 2017;Unterthiner et al, 2018;Tembine, 2019).…”

Semi-supervised Conditional GAN for Simultaneous Generation and Detection of Phishing URLs: A Game theoretic Perspective

“…A cyber-defense scenario almost always depends on game theory to understand the attacker's motives and perspective for maximizing the defender's reward. Many similar games have been designed for phishing URL detection with humans in the loop [29,43,14] and adversarial games for generative networks [22,31,17]. However, no work has been proposed, combining adversarial components of GAN and gametheoretic perspectives of attacker-defender for detecting phishing URLs.…”

A game-theoretic approach for Generative Adversarial Networks

“…The generator is represented by a differentiable function g, that is, a neural network class with parameter vector x g ∈ Ω g ⊆ R ng . The (fake) output of the generator is denoted with g(z, x g ) ∈ R q where the input z is a random noise drawn from the model prior distribution, z ∼ p z , that the generator uses to create the fake output g(z, x g ) [6]. The actual strategies of the generator are the parameters x g that allows g to produce the fake output.…”

“…It can be proven that the two-player game with cost functions (2) and ( 4) and the zero-sum game with cost function (2) and relation (1) have the same equilibria [6,Theorem 10].…”

“…Given their connection with robust optimization and game theory, GANs have received theoretical attention as well, both for modelling as Nash equilibrium problems [6], [7] and for designing algorithms that improve the training process [8], [7].…”