"The desire to create is one of the deepest yearnings of the human soul."
-- Dieter F. Uchtdorf
Learning goals from this chapter:
- What Are Generative Adversarial Networks?
- Understand the basic components of GANs: Generative and Discriminative models.
- Learn different techniques to evaluate generative models
- Learn how to build a GAN model
Generative Adversarial Networks (GANs) are a new type of neural architectures introduced by Ian Goodfellow and other researchers at the University of Montreal, including Yoshua Bengio, in June 2014, in their paper “Generative Adversarial Nets” in 2014. GANs have been called “the most interesting idea in the last 10 years in ML” by Yann LeCun, Facebook’s AI research director. The excitement is well justified. The most notable features of GANs are their capacity to create hyperrealistic images, videos, music, and text. For example, none of the faces in Figure 8.1 (right) belong to a real human; they are all fake. Same thing for the handwritten digits in the left image in the same figure. This shows GAN’s ability to learn the features from the training images and imagine its own new images using these patterns learned.