Application of GAN in Semi-Supervised Learning

The third wave of artificial intelligence are inseparable with deep learning. The success of deep learning is mainly attributed to three major factors – big data, big models and big calculations, but they also brought some constraints to the further development and popularization of deep learning. The first challenge is labeling data is expensive, we should design a new learning method that can learn from unlabeled data. Ian Goodfellow who is inventor of GAN, mentioned that the method by which GAN can be used for semi supervised learning is called SSGAN. The experimental results show that the discriminant model trained by this processing method has better effects than other methods in rationally using unlabeled data.

Paper

Full text

PDF

Application of GAN in Semi-Supervised Learning

Semantic Scholar · Computer Science · 2019

Abstract

The third wave of artificial intelligence are inseparable with deep learning. The success of deep learning is mainly attributed to three major factors – big data, big models and big calculations, but they also brought some constraints to the further development and popularization of deep learning. The first challenge is labeling data is expensive, we should design a new learning method that can learn from unlabeled data. Ian Goodfellow who is inventor of GAN, mentioned that the method by which GAN can be used for semi supervised learning is called SSGAN. The experimental results show that the discriminant model trained by this processing method has better effects than other methods in rationally using unlabeled data.

Similar papers

© 2026 NYSGPT2525 LLC