English

On Self Modulation for Generative Adversarial Networks

Machine Learning 2019-05-03 v2 Computer Vision and Pattern Recognition Machine Learning

Abstract

Training Generative Adversarial Networks (GANs) is notoriously challenging. We propose and study an architectural modification, self-modulation, which improves GAN performance across different data sets, architectures, losses, regularizers, and hyperparameter settings. Intuitively, self-modulation allows the intermediate feature maps of a generator to change as a function of the input noise vector. While reminiscent of other conditioning techniques, it requires no labeled data. In a large-scale empirical study we observe a relative decrease of 5%35%5\%-35\% in FID. Furthermore, all else being equal, adding this modification to the generator leads to improved performance in 124/144124/144 (86%86\%) of the studied settings. Self-modulation is a simple architectural change that requires no additional parameter tuning, which suggests that it can be applied readily to any GAN.

Keywords

Cite

@article{arxiv.1810.01365,
  title  = {On Self Modulation for Generative Adversarial Networks},
  author = {Ting Chen and Mario Lucic and Neil Houlsby and Sylvain Gelly},
  journal= {arXiv preprint arXiv:1810.01365},
  year   = {2019}
}
R2 v1 2026-06-23T04:26:11.783Z