machinelearningmedium.com
Generative Adversarial Networks
GAN originally presented by Goodfellow et al is a novel technique that uses a minmax two-player game to learn latent data distributions.