Semi-supervised Conditional GANs

TL;DR

This paper presents a semi-supervised conditional GAN model that effectively learns data and attribute distributions using a pair of discriminators, outperforming existing models in generating conditioned data.

Contribution

The paper introduces a novel semi-supervised GAN architecture with stacked discriminators for improved conditional data generation.

Findings

Significantly better performance than existing semi-supervised conditional GANs

Effective learning of marginal and conditional distributions from labeled and unlabeled data

Demonstrated improvements across multiple datasets

Abstract

We introduce a new model for building conditional generative models in a semi-supervised setting to conditionally generate data given attributes by adapting the GAN framework. The proposed semi-supervised GAN (SS-GAN) model uses a pair of stacked discriminators to learn the marginal distribution of the data, and the conditional distribution of the attributes given the data respectively. In the semi-supervised setting, the marginal distribution (which is often harder to learn) is learned from the labeled + unlabeled data, and the conditional distribution is learned purely from the labeled data. Our experimental results demonstrate that this model performs significantly better compared to existing semi-supervised conditional GAN models.