# MAGAN: Margin Adaptation for Generative Adversarial Networks

arXiv: https://arxiv.org/abs/1704.03817 · HTML: https://mlsanity.com/paper/1704.03817/

## Key facts

- Cited by 10 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 12 Apr 2017.

## Built on by (most cited first, 10 of 10)

- [Reverse Engineering of Generative Models: Inferring Model Hyperparameters from Generated Images](https://mlsanity.com/paper/2106.07873/) (2021)
- [Generative Multiplane Images: Making a 2D GAN 3D-Aware](https://mlsanity.com/paper/2207.10642/) (2022)
- [Generative Adversarial Networks (GANs) in Networking: A Comprehensive Survey & Evaluation](https://mlsanity.com/paper/2105.04184/) (2021)
- [DigGAN: Discriminator gradIent Gap Regularization for GAN Training with Limited Data](https://mlsanity.com/paper/2211.14694/) (2022)
- [Sparsity Aware Normalization for GANs](https://mlsanity.com/paper/2103.02458/) (2021)
- [MSGDD-cGAN: Multi-Scale Gradients Dual Discriminator Conditional Generative Adversarial Network](https://mlsanity.com/paper/2109.05614/) (2021)
- [3D GANs and Latent Space: A comprehensive survey](https://mlsanity.com/paper/2304.03932/) (2023)
- [HGAN: Hybrid Generative Adversarial Network](https://mlsanity.com/paper/2102.03710/) (2021)
- [Contrastive Cycle Adversarial Autoencoders for Single-cell Multi-omics Alignment and Integration](https://mlsanity.com/paper/2112.03266/) (2021)
- [HMB-GAN: Hybrid Multi-B'ezier GAN for Vector Shape Synthesis](https://mlsanity.com/paper/2609.21158/) (2026)

## Abstract (arXiv)

We propose the Margin Adaptation for Generative Adversarial Networks (MAGANs) algorithm, a novel training procedure for GANs to improve stability and performance by using an adaptive hinge loss function. We estimate the appropriate hinge loss margin with the expected energy of the target distribution, and derive principled criteria for when to update the margin. We prove that our method converges to its global optimum under certain assumptions. Evaluated on the task of unsupervised image generation, the proposed training procedure is simple yet robust on a diverse set of data, and achieves qualitative and quantitative improvements compared to the state-of-the-art.