Continuous Conditional Generative Adversarial Networks (cGAN) with Generator Regularization

TL;DR

This paper introduces a Lipschitz regularization for continuous conditional GANs to improve training stability and sample quality by leveraging neighboring condition information.

Contribution

It proposes a novel generator regularization term based on Lipschitz penalty to enhance continuous conditional GAN training.

Findings

Improved sample consistency for neighboring conditions

Robust performance on synthetic and real-world tasks

Enhanced training stability

Abstract

Conditional Generative Adversarial Networks are known to be difficult to train, especially when the conditions are continuous and high-dimensional. To partially alleviate this difficulty, we propose a simple generator regularization term on the GAN generator loss in the form of Lipschitz penalty. Thus, when the generator is fed with neighboring conditions in the continuous space, the regularization term will leverage the neighbor information and push the generator to generate samples that have similar conditional distributions for each neighboring condition. We analyze the effect of the proposed regularization term and demonstrate its robust performance on a range of synthetic and real-world tasks.