DuelGAN: A Duel Between Two Discriminators Stabilizes the GAN Training

19 Jan 2021  ·  Jiaheng Wei, Minghao Liu, Jiahao Luo, Andrew Zhu, James Davis, Yang Liu ·

In this paper, we introduce DuelGAN, a generative adversarial network (GAN) solution to improve the stability of the generated samples and to mitigate mode collapse. Built upon the Vanilla GAN's two-player game between the discriminator $D_1$ and the generator $G$, we introduce a peer discriminator $D_2$ to the min-max game. Similar to previous work using two discriminators, the first role of both $D_1$, $D_2$ is to distinguish between generated samples and real ones, while the generator tries to generate high-quality samples which are able to fool both discriminators. Different from existing methods, we introduce another game between $D_1$ and $D_2$ to discourage their agreement and therefore increase the level of diversity of the generated samples. This property alleviates the issue of early mode collapse by preventing $D_1$ and $D_2$ from converging too fast. We provide theoretical analysis for the equilibrium of the min-max game formed among $G, D_1, D_2$. We offer convergence behavior of DuelGAN as well as stability of the min-max game. It's worth mentioning that DuelGAN operates in the unsupervised setting, and the duel between $D_1$ and $D_2$ does not need any label supervision. Experiments results on a synthetic dataset and on real-world image datasets (MNIST, Fashion MNIST, CIFAR-10, STL-10, CelebA, VGG, and FFHQ) demonstrate that DuelGAN outperforms competitive baseline work in generating diverse and high-quality samples, while only introduces negligible computation cost.

PDF Abstract

Results from the Paper


Task Dataset Model Metric Name Metric Value Global Rank Result Benchmark
Image Generation CelebA 64x64 PeerGAN FID 13.95 # 27
Image Generation CIFAR-10 PeerGAN Inception score 7.45 # 64
FID 21.55 # 135
Image Generation Fashion-MNIST PeerGAN FID 21.73 # 3
Image Generation STL-10 PeerGAN FID 51.37 # 27

Methods


No methods listed for this paper. Add relevant methods here