Adversarial confidence and smoothness regularizations for scalable unsupervised discriminative learning

Open in new window