Training Generative Adversarial Networks Via Turing Test

10/25/2018

∙

by   Jianlin Su, et al.

∙

2

∙

share

In this article, we introduce a new mode for training Generative Adversarial Networks (GANs). Rather than minimizing the distance of evidence distribution p̃(x) and the generative distribution q(x), we minimize the distance of p̃(x_r)q(x_f) and p̃(x_f)q(x_r). This adversarial pattern can be interpreted as a Turing test in GANs. It allows us to use information of real samples during training generator and accelerates the whole training procedure. We even find that just proportionally increasing the size of discriminator and generator, it succeeds on 256x256 resolution without adjusting hyperparameters carefully.