Neural Network Projects with Python

Generative adversarial networks

Generative adversarial networks (GANs) are a class of generative neural networks. To understand generative models, it's important to contrast them against discriminative models. So far in this book, we have focused only on discriminative models. Discriminative models are concerned with learning the mapping of features to a label. For example, when we created a CNN to classify images of cats and dogs, the CNN is a discriminative model that learns the mapping of features (images) to a label (a cat or a dog).

On the other hand, generative models are concerned with generating appropriate features, given the label. For example, given labeled images of cats and dogs, a generative model would learn to create the ...