# On the Effectiveness of Least Squares Generative Adversarial Networks

arXiv: https://arxiv.org/abs/1712.06391 · HTML: https://mlsanity.com/paper/1712.06391/

## Key facts

- Cited by 18 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 18 Dec 2017.

## Built on by (most cited first, 18 of 18)

- [UVCGAN: UNet Vision Transformer cycle-consistent GAN for unpaired image-to-image translation](https://mlsanity.com/paper/2203.02557/) (2022)
- [Privacy-preserving Spatiotemporal Scenario Generation of Renewable Energies: A Federated Deep Generative Learning Approach](https://mlsanity.com/paper/2107.07738/) (2021)
- [AutoML Two-Sample Test](https://mlsanity.com/paper/2206.08843/) (2022)
- [WAS-VTON: Warping Architecture Search for Virtual Try-on Network](https://mlsanity.com/paper/2108.00386/) (2021)
- [Activated Gradients for Deep Neural Networks](https://mlsanity.com/paper/2107.04228/) (2021)
- [Generalization Error of GAN from the Discriminator's Perspective](https://mlsanity.com/paper/2107.03633/) (2021)
- [Re-using Adversarial Mask Discriminators for Test-time Training under Distribution Shifts](https://mlsanity.com/paper/2108.11926/) (2021)
- [A Mathematical Framework for Learning Probability Distributions](https://mlsanity.com/paper/2212.11481/) (2022)
- [Multi-Discriminator Sobolev Defense-GAN Against Adversarial Attacks for End-to-End Speech Systems](https://mlsanity.com/paper/2103.08086/) (2021)
- [Stop Throwing Away Discriminators! Re-using Adversaries for Test-Time Training](https://mlsanity.com/paper/2108.12280/) (2021)
- [Generative Adversarial Networks and Adversarial Autoencoders: Tutorial and Survey](https://mlsanity.com/paper/2111.13282/) (2021)
- [PCT-CycleGAN: Paired Complementary Temporal Cycle-Consistent Adversarial Networks for Radar-Based Precipitation Nowcasting](https://mlsanity.com/paper/2211.15046/) (2022)
- [Evolving GAN Formulations for Higher Quality Image Synthesis](https://mlsanity.com/paper/2102.08578/) (2021)
- [Bi-Discriminator Class-Conditional Tabular GAN](https://mlsanity.com/paper/2111.06549/) (2021)
- [Nonlinear ill-posed problem in low-dose dental cone-beam computed tomography](https://mlsanity.com/paper/2303.01678/) (2023)
- [Generalization of GANs and overparameterized models under Lipschitz continuity](https://mlsanity.com/paper/2104.02388/) (2021)
- [Advances in Electron Microscopy with Deep Learning](https://mlsanity.com/paper/2101.01178/) (2021)
- [Deep Generative Adversarial Network for Occlusion Removal from a Single Image](https://mlsanity.com/paper/2409.13242/) (2024)

## Abstract (arXiv)

Unsupervised learning with generative adversarial networks (GANs) has proven to be 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 for both the discriminator and the generator. We show that minimizing the objective function of LSGAN yields minimizing the Pearson chi2 divergence. We also show that the derived objective function that yields minimizing the Pearson chi2 divergence performs better than the classical one of using least squares for classification. 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 stably during the learning process. For evaluating the image quality, we conduct both qualitative and quantitative experiments, and the experimental results show that LSGANs can generate higher quality images than regular GANs. Furthermore, we evaluate the stability of LSGANs in two groups. One is to compare between LSGANs and regular GANs without gradient penalty. We conduct three experiments, including Gaussian mixture distribution, difficult architectures, and a newly proposed method --- datasets with small variability, to illustrate the stability of LSGANs. The other one is to compare between LSGANs with gradient penalty (LSGANs-GP) and WGANs with gradient penalty (WGANs-GP). The experimental results show that LSGANs-GP succeed in training for all the difficult architectures used in WGANs-GP, including 101-layer ResNet.