Energy-based Generative Adversarial Networks

Junbo Zhao, Michael Mathieu, Yann LeCun

Nov 04, 2016 (modified: Mar 06, 2017) ICLR 2017 conference submission readers: everyone
  • Abstract: We introduce the "Energy-based Generative Adversarial Network" model (EBGAN) which views the discriminator as an energy function that attributes low energies to the regions near the data manifold and higher energies to other regions. Similar to the probabilistic GANs, a generator is seen as being trained to produce contrastive samples with minimal energies, while the discriminator is trained to assign high energies to these generated samples. Viewing the discriminator as an energy function allows to use a wide variety of architectures and loss functionals in addition to the usual binary classifier with logistic output. Among them, we show one instantiation of EBGAN framework as using an auto-encoder architecture, with the energy being the reconstruction error, in place of the discriminator. We show that this form of EBGAN exhibits more stable behavior than regular GANs during training. We also show that a single-scale architecture can be trained to generate high-resolution images.
  • TL;DR: We introduce the "Energy-based Generative Adversarial Network" (EBGAN) model.
  • Conflicts:,
  • Keywords: Deep learning, Unsupervised Learning, Semi-Supervised Learning