Overview: What’s Next in GANs

Now that we have been thoroughly exposed to deep learning and GANs, in this chapter, we will learn about the possible future avenues for GANs! We start with a summary of this course, the topics that we covered, and the knowledge that we have gained so far.

Next, we address important open questions related to GANs that are essential for interacting with GAN models. We briefly pose questions related to how important architectures are, whether GANs really learn the target distribution, whether they are dependent on the inductive bias of architectures, and how to identify GAN samples.

Following this, we consider the artistic use of GANs in the visual and sonic arts. In the visual arts, we provide examples of painting and video generation; while in the sonic arts, we provide examples of instrument synthesis and music generation.

Finally, we look at recent and yet-to-be-explored domains within GANs. We offer a brief discussion on verified artificial intelligence (AI) and explore GAN examples in biology, audio, and other domains.

Topics covered in this chapter

The following topics will be covered in this chapter:

• What we’ve learned so far

• Unanswered questions in GANs

• Artistic GANs

• Recent and yet-to-be-explored GAN topics