Dist-GAN: An Improved GAN using Distance Constraints

TL;DR

Dist-GAN introduces novel distance constraints and an autoencoder-based training approach to improve GAN stability, reduce mode collapse, and enhance sample quality across multiple benchmark datasets.

Contribution

The paper proposes a new GAN framework, Dist-GAN, with two innovative distance constraints and an autoencoder coupling to address mode collapse and gradient vanishing.

Findings

Outperforms state-of-the-art methods on benchmark datasets

Effectively reduces mode collapse and gradient vanishing

Achieves better sample diversity and quality

Abstract

We introduce effective training algorithms for Generative Adversarial Networks (GAN) to alleviate mode collapse and gradient vanishing. In our system, we constrain the generator by an Autoencoder (AE). We propose a formulation to consider the reconstructed samples from AE as "real" samples for the discriminator. This couples the convergence of the AE with that of the discriminator, effectively slowing down the convergence of discriminator and reducing gradient vanishing. Importantly, we propose two novel distance constraints to improve the generator. First, we propose a latent-data distance constraint to enforce compatibility between the latent sample distances and the corresponding data sample distances. We use this constraint to explicitly prevent the generator from mode collapse. Second, we propose a discriminator-score distance constraint to align the distribution of the generated samples with that of the real samples through the discriminator score. We use this constraint to guide the generator to synthesize samples that resemble the real ones. Our proposed GAN using these distance constraints, namely Dist-GAN, can achieve better results than state-of-the-art methods across benchmark datasets: synthetic, MNIST, MNIST-1K, CelebA, CIFAR-10 and STL-10 datasets. Our code is published here (https://github.com/tntrung/gan) for research.