Training Generative Adversarial Networks via Stochastic Nash Games

Barbara Franci, Sergio Grammatico

Research output: Contribution to journalArticleScientificpeer-review

5 Citations (Scopus)
19 Downloads (Pure)

Abstract

Generative adversarial networks (GANs) are a class of generative models with two antagonistic neural networks: a generator and a discriminator. These two neural networks compete against each other through an adversarial process that can be modeled as a stochastic Nash equilibrium problem. Since the associated training process is challenging, it is fundamental to design reliable algorithms to compute an equilibrium. In this article, we propose a stochastic relaxed forward-backward (SRFB) algorithm for GANs, and we show convergence to an exact solution when an increasing number of data is available. We also show convergence of an averaged variant of the SRFB algorithm to a neighborhood of the solution when only a few samples are available. In both cases, convergence is guaranteed when the pseudogradient mapping of the game is monotone. This assumption is among the weakest known in the literature. Moreover, we apply our algorithm to the image generation problem.

Original languageEnglish
Pages (from-to)1319-1328
JournalIEEE Transactions on Neural Networks and Learning Systems
Volume34
Issue number3
DOIs
Publication statusPublished - 2023

Bibliographical note

Green Open Access added to TU Delft Institutional Repository 'You share, we take care!' - Taverne project https://www.openaccess.nl/en/you-share-we-take-care
Otherwise as indicated in the copyright section: the publisher is the copyright holder of this work and the author uses the Dutch legislation to make this work public.

Keywords

  • Convergence
  • Games
  • Generative adversarial networks
  • Generative adversarial networks (GANs)
  • Generators
  • Neural networks
  • stochastic Nash equilibrium (SNE) problems (SNEPs)
  • Stochastic processes
  • Training
  • two-player game
  • variational inequalities.

Fingerprint

Dive into the research topics of 'Training Generative Adversarial Networks via Stochastic Nash Games'. Together they form a unique fingerprint.

Cite this