Introduction

In this chapter, we will introduce GANs. Just as in autoencoder networks, GANs have a generator and a discriminator network. However, GANs are fundamentally different. They represent an unsupervised learning problem, where the two networks compete, and cooperate with each other at the same time. It is important that the generator and discriminator don't overpower each other. The idea behind GANs is to generate new examples based on training data. Applications can range from generating new handwritten MNIST images to generating music. GANs have received a lot of attention lately because the results of using them are fascinating.