Semi-supervised Learning on Graphs with Generative Adversarial Nets

TL;DR

This paper introduces GraphSGAN, a novel semi-supervised learning method on graphs using GANs, which generates fake samples in low-density areas to improve classification accuracy and scalability.

Contribution

The paper proposes GraphSGAN, a new adversarial framework for semi-supervised learning on graphs that enhances traditional regularization with a theoretical guarantee.

Findings

GraphSGAN outperforms state-of-the-art methods on multiple datasets.

The approach is scalable with mini-batch training.

Theoretical analysis supports the method's effectiveness.

Abstract

We investigate how generative adversarial nets (GANs) can help semi-supervised learning on graphs. We first provide insights on working principles of adversarial learning over graphs and then present GraphSGAN, a novel approach to semi-supervised learning on graphs. In GraphSGAN, generator and classifier networks play a novel competitive game. At equilibrium, generator generates fake samples in low-density areas between subgraphs. In order to discriminate fake samples from the real, classifier implicitly takes the density property of subgraph into consideration. An efficient adversarial learning algorithm has been developed to improve traditional normalized graph Laplacian regularization with a theoretical guarantee. Experimental results on several different genres of datasets show that the proposed GraphSGAN significantly outperforms several state-of-the-art methods. GraphSGAN can be also trained using mini-batch, thus enjoys the scalability advantage.