Bayesian GAN

Yunus Saatchi, Andrew Gordon Wilson

Generative adversarial networks (GANs) can implicitly learn rich distributions over images, audio, and data which are hard to model with an explicit likelihood. We present a practi…