Improving GANs Using Optimal Transport

TL;DR

This paper introduces OT-GAN, a new GAN variant that uses a novel mini-batch energy distance metric based on optimal transport, leading to improved stability and state-of-the-art image generation results.

Contribution

The paper proposes a new metric combining optimal transport and energy distance, enhancing GAN training stability and performance.

Findings

OT-GAN is highly stable with large mini-batches.

Achieves state-of-the-art results on benchmark image generation tasks.

The mini-batch energy distance provides a highly discriminative and unbiased gradient.

Abstract

We present Optimal Transport GAN (OT-GAN), a variant of generative adversarial nets minimizing a new metric measuring the distance between the generator distribution and the data distribution. This metric, which we call mini-batch energy distance, combines optimal transport in primal form with an energy distance defined in an adversarially learned feature space, resulting in a highly discriminative distance function with unbiased mini-batch gradients. Experimentally we show OT-GAN to be highly stable when trained with large mini-batches, and we present state-of-the-art results on several popular benchmark problems for image generation.