Generative Adversarial Parallelization

TL;DR

This paper introduces Generative Adversarial Parallelization, a novel training framework for GANs that enhances convergence and mode coverage by training multiple GANs simultaneously with shared discriminators.

Contribution

It proposes a new parallel training framework for GANs that improves stability and mode coverage, along with an enhanced metric for evaluating GAN collections.

Findings

Improved convergence in GAN training.

Enhanced mode coverage in generated samples.

Effective scoring of GAN collections using the new metric.

Abstract

Generative Adversarial Networks have become one of the most studied frameworks for unsupervised learning due to their intuitive formulation. They have also been shown to be capable of generating convincing examples in limited domains, such as low-resolution images. However, they still prove difficult to train in practice and tend to ignore modes of the data generating distribution. Quantitatively capturing effects such as mode coverage and more generally the quality of the generative model still remain elusive. We propose Generative Adversarial Parallelization, a framework in which many GANs or their variants are trained simultaneously, exchanging their discriminators. This eliminates the tight coupling between a generator and discriminator, leading to improved convergence and improved coverage of modes. We also propose an improved variant of the recently proposed Generative Adversarial Metric and show how it can score individual GANs or their collections under the GAP model.