Least Squares Generative Adversarial Networks

TL;DR

This paper introduces Least Squares GANs (LSGANs), which replace the traditional sigmoid cross entropy loss with a least squares loss, leading to higher quality image generation and more stable training.

Contribution

The paper proposes LSGANs that minimize Pearson's chi-squared divergence, improving image quality and training stability over traditional GANs.

Findings

LSGANs generate higher quality images than regular GANs.

LSGANs demonstrate more stable training processes.

Experimental results on five datasets confirm the advantages of LSGANs.

Abstract

Unsupervised learning with generative adversarial networks (GANs) has proven hugely successful. Regular GANs hypothesize the discriminator as a classifier with the sigmoid cross entropy loss function. However, we found that this loss function may lead to the vanishing gradients problem during the learning process. To overcome such a problem, we propose in this paper the Least Squares Generative Adversarial Networks (LSGANs) which adopt the least squares loss function for the discriminator. We show that minimizing the objective function of LSGAN yields minimizing the Pearson $\chi^2$ divergence. There are two benefits of LSGANs over regular GANs. First, LSGANs are able to generate higher quality images than regular GANs. Second, LSGANs perform more stable during the learning process. We evaluate LSGANs on five scene datasets and the experimental results show that the images generated by LSGANs are of better quality than the ones generated by regular GANs. We also conduct two comparison experiments between LSGANs and regular GANs to illustrate the stability of LSGANs.