Generative Multi-Adversarial Networks

Venue

arXiv:1611.01673 [cs], vol.

Publication Year

2016

Keywords

Computer Science - Machine Learning,Computer Science - Multiagent Systems,Computer Science - Neural and Evolutionary Computing

Authors

  • Ishan Durugkar
  • Ian Gemp
  • Sridhar Mahadevan

Abstract

Generative adversarial networks (GANs) are a framework for producing a generative model by way of a two-player minimax game. In this paper, we propose the \emph{Generative Multi-Adversarial Network} (GMAN), a framework that extends GANs to multiple discriminators. In previous work, the successful training of GANs requires modifying the minimax objective to accelerate training early on. In contrast, GMAN can be reliably trained with the original, untampered objective. We explore a number of design perspectives with the discriminator role ranging from formidable adversary to forgiving teacher. Image generation tasks comparing the proposed framework to standard GANs demonstrate GMAN produces higher quality samples in a fraction of the iterations when measured by a pairwise GAM-type metric.