Teaching a GAN What Not to Learn

Siddarth Asokan, Chandra Sekhar Seelamantula

Generative adversarial networks (GANs) were originally envisioned as unsupervised generative models that learn to follow a target distribution. Variants such as conditional GANs, a…