S2cGAN: Semi-Supervised Training of Conditional GANs with Fewer Labels

TL;DR

This paper introduces S2cGAN, a semi-supervised approach for training conditional GANs that reduces the need for extensive labeled data by leveraging both sparse labels and large amounts of unlabeled data.

Contribution

It presents a novel semi-supervised training framework for cGANs that effectively combines sparse labels with unlabeled data to learn conditional distributions.

Findings

Effective on multiple datasets and tasks

Reduces label requirements for training cGANs

Maintains high-quality sample generation

Abstract

Generative adversarial networks (GANs) have been remarkably successful in learning complex high dimensional real word distributions and generating realistic samples. However, they provide limited control over the generation process. Conditional GANs (cGANs) provide a mechanism to control the generation process by conditioning the output on a user defined input. Although training GANs requires only unsupervised data, training cGANs requires labelled data which can be very expensive to obtain. We propose a framework for semi-supervised training of cGANs which utilizes sparse labels to learn the conditional mapping, and at the same time leverages a large amount of unsupervised data to learn the unconditional distribution. We demonstrate effectiveness of our method on multiple datasets and different conditional tasks.