Poster
On Self Modulation for Generative Adversarial Networks
Ting Chen · Mario Lucic · Neil Houlsby · Sylvain Gelly
Great Hall BC #39
Keywords: [ generative adversarial networks ] [ unsupervised learning ] [ deep generative modelling ]
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% in FID. Furthermore, all else being equal, adding this modification to the generator leads to improved performance in 124/144 (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.
Live content is unavailable. Log in and register to view live content