Semi-supervised Learning with GANs: Manifold Invariance with Improved Inference

TL;DR

This paper introduces a semi-supervised learning approach using GANs to estimate data manifold tangent spaces, enhancing classifier invariance and improving performance especially with limited labeled data.

Contribution

It proposes a novel method to estimate the data manifold tangent space with GANs and improves inverse mapping, leading to better semi-supervised learning results.

Findings

Empirical gains in semi-supervised learning with few labeled examples.

Enhanced inverse mapping improves semantic similarity of reconstructions.

Insights into the role of fake examples in semi-supervised learning.

Abstract

Semi-supervised learning methods using Generative Adversarial Networks (GANs) have shown promising empirical success recently. Most of these methods use a shared discriminator/classifier which discriminates real examples from fake while also predicting the class label. Motivated by the ability of the GANs generator to capture the data manifold well, we propose to estimate the tangent space to the data manifold using GANs and employ it to inject invariances into the classifier. In the process, we propose enhancements over existing methods for learning the inverse mapping (i.e., the encoder) which greatly improves in terms of semantic similarity of the reconstructed sample with the input sample. We observe considerable empirical gains in semi-supervised learning over baselines, particularly in the cases when the number of labeled examples is low. We also provide insights into how fake examples influence the semi-supervised learning procedure.