Image Generation with GANs
Generative modeling is a powerful concept that provides us with immense potential to approximate or model underlying processes that generate data. In the chapters so far, we have covered concepts associated with deep learning in general and, more specifically, related to Variational Autoencoders. In this chapter, we will introduce another family of generative models called Generative Adversarial Networks, or GANs. Heavily inspired by the concepts of game theory and picking up some of the best components from previously discussed techniques, GANs provide a powerful framework to work in the generative modeling space. Since their invention in 2014 by Goodfellow et al.1, GANs have been leveraged to explore creative domains such as art auctions, fashion, and photography. The following are two amazing high-quality samples from a variant of GANs called StyleGAN (Figure 12.1). The photograph of the kid is actually a fictional person who does not exist. Similarly...
