Generative models are the most promising push toward enabling computers to have an understanding of the world. They are true unsupervised models, and are able to perform those tasks that many today consider to be at the cutting edge of artificial intelligence (AI). Generative models are different for precisely the reason as it sounds: they generate data. Centered mostly around computer vision tasks, this class of network has the power to create new faces, new handwriting, or even paintings.
In this section, we'll introduce generative models and their foundations, focusing specifically on the two most popular types of model, the variational autoencoder (VAE), and the generative adversarial network (GAN), where you'll learn how to generate pictures and faces with these networks. We'll also touch upon other common generative models, and finish up...
