4. Conclusion

This chapter discussed the general principles behind GANs so as to give you a foundation for the more advanced topics we'll now move on to, including improved GANs, disentangled representation GANs, and cross-domain GANs. We started this chapter by understanding how GANs are made up of two networks, called generator and discriminator. The role of the discriminator is to discriminate between real and fake signals. The aim of the generator is to fool the discriminator. The generator is normally combined with the discriminator to form an adversarial network. It is through training the adversarial network that the generator learns how to produce fake data that can trick the discriminator.

We also learned how GANs are easy to build but notoriously difficult to train. Two example implementations in tf.keras were presented. DCGAN demonstrated that it is possible to train GANs to generate fake images using deep CNNs. The fake images were MNIST digits. However...