Keras Implementation of Semi Supervised GAN
Semi-supervised learning is the challenging problem of training a classifier in a dataset that contains a small number of labeled examples and a much larger number of unlabeled examples.
Semi Supervised GAN or simply SGAN is an extension of the GAN Architecture that involves simlultaneous training of Supervised Discriminator,Unsupervised Discriminator and Generator.
This is SGAN trained on MNIST Dataset using only 10 labeled examples per class i.e just 100 examples and was able to achieve a impressive accuracy of 90.67% whereas a normal supervised classifier trained on same data was able to reach around 65-70% accuracy