Typically, the generative network is taught to map from a latent space to a particular data distribution of interest
