Class-Splitting Generative Adversarial Networks

  title={Class-Splitting Generative Adversarial Networks},
  author={Guillermo L. Grinblat and Lucas C. Uzal and Pablo M. Granitto},
Generative Adversarial Networks (GANs) produce systematically better quality samples when class label information is provided., i.e. in the conditional GAN setup. This is still observed for the recently proposed Wasserstein GAN formulation which stabilized adversarial training and allows considering high capacity network architectures such as ResNet. In this work we show how to boost conditional GAN by augmenting available class labels. The new classes come from clustering in the representation… CONTINUE READING
Recent Discussions
This paper has been referenced on Twitter 86 times over the past 90 days. VIEW TWEETS


Publications referenced by this paper.
Showing 1-10 of 17 references


  • I. Goodfellow, J. Pouget-Abadie, M. Mirza, B. Xu, D. Warde-Farley, S. Ozair
  • Courville, and Y. Bengio. Generative adversarial…
  • 2014
Highly Influential
5 Excerpts

Improving generative adversarial networks with denoising feature matching

  • D. Warde-Farley, Y. Bengio
  • International Conference on Learning…
  • 2017
1 Excerpt

Similar Papers

Loading similar papers…