Adversarial symmetric GANs: bridging adversarial samples and adversarial networks

TL;DR

This paper introduces adversarial symmetric GANs (AS-GANs), which enhance training stability and sample quality by symmetrically applying adversarial training to both real and fake samples, addressing a key vulnerability in vanilla GANs.

Contribution

The paper proposes AS-GANs that incorporate adversarial training on real samples, making the discriminator more robust and improving training stability and sample quality.

Findings

Training stability is improved with AS-GANs.

FID scores are significantly better with AS-GANs.

Discriminator robustness is enhanced through symmetric adversarial training.

Abstract

Generative adversarial networks have achieved remarkable performance on various tasks but suffer from training instability. Despite many training strategies proposed to improve training stability, this issue remains as a challenge. In this paper, we investigate the training instability from the perspective of adversarial samples and reveal that adversarial training on fake samples is implemented in vanilla GANs, but adversarial training on real samples has long been overlooked. Consequently, the discriminator is extremely vulnerable to adversarial perturbation and the gradient given by the discriminator contains non-informative adversarial noises, which hinders the generator from catching the pattern of real samples. Here, we develop adversarial symmetric GANs (AS-GANs) that incorporate adversarial training of the discriminator on real samples into vanilla GANs, making adversarial training symmetrical. The discriminator is therefore more robust and provides more informative gradient with less adversarial noise, thereby stabilizing training and accelerating convergence. The effectiveness of the AS-GANs is verified on image generation on CIFAR-10 , CelebA, and LSUN with varied network architectures. Not only the training is more stabilized, but the FID scores of generated samples are consistently improved by a large margin compared to the baseline. The bridging of adversarial samples and adversarial networks provides a new approach to further develop adversarial networks.