A Self-Training Method for Semi-Supervised GANs

TL;DR

This paper introduces a novel self-training approach for semi-supervised GANs, leveraging their data generation ability to improve training with limited labeled data, demonstrating enhanced performance with simple and complex schemes.

Contribution

It presents the first integration of self-training methods into GANs for semi-supervised learning, exploiting their data generation capacity to improve training outcomes.

Findings

Self-training improves GAN performance in semi-supervised tasks.

Complex self-training schemes can match simple ones with less data augmentation.

Even basic self-training yields notable improvements.

Abstract

Since the creation of Generative Adversarial Networks (GANs), much work has been done to improve their training stability, their generated image quality, their range of application but nearly none of them explored their self-training potential. Self-training has been used before the advent of deep learning in order to allow training on limited labelled training data and has shown impressive results in semi-supervised learning. In this work, we combine these two ideas and make GANs self-trainable for semi-supervised learning tasks by exploiting their infinite data generation potential. Results show that using even the simplest form of self-training yields an improvement. We also show results for a more complex self-training scheme that performs at least as well as the basic self-training scheme but with significantly less data augmentation.