Semi-supervised Learning on Graphs with Generative Adversarial Nets

By Ming Ding, Jie Tang, Jie Zhang. 2018

123 citations recorded by the source corpus.

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.

Subfields: Artificial Intelligence, Statistical and Nonlinear Physics

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Artificial Intelligence

In Statistical and Nonlinear Physics

Citations in the corpus

Deep Graph Infomax (2018, methodology): “It should be noted that, while we are capable of outperforming equivalent supervised encoder architectures, our performance still does not surpass the current supervised transductive state of the art (which is held by methods such as GraphSGAN (Ding et al., 2018)).”
Learning Graph Embedding With Adversarial Training Methods (2019, methodology): “[51] proposed a GraphSGAN for semisupervised node classification with the GAN principle, and Hu et al.”
Graph Random Neural Networks for Semi-Supervised Learning on Graphs (2020, background): “The focus of this work is on the problem of semisupervised learning on graphs [46, 20, 10], which aims to predict the categories of unlabeled nodes of a given graph with only a small proportion of labeled nodes.”
Graph Wavelet Neural Network (2019, background): “me works devote to making graph convolutional networks more powerful. Monti et al. (2018) alternated convolutions on vertices and edges, generalizing GAT and leading to better performance. GraphsGAN (Ding et al., 2018) generalizes GANs to graph, and generates fake samples in low-density areas betwee”
Graph Representation Learning via Graphical Mutual Information Maximization (2020, background): “Among them, there is a rich literature on supervised representation learning over graphs [7, 9, 25, 40, 48].”
LARA: Attribute-to-feature Adversarial Learning for New-item Recommendation (2020, background): “2) In most situations, GANs have an advantage in generating dense feature vectors, just as the original GANs are applied to generating images [7, 8].”
Are we really making much progress?: Revisiting, benchmarking and refining heterogeneous graph neural networks (2021, background): “Many following works [1, 7, 11, 37] improve GCN and GAT furthermore, with focuses on homogeneous graphs.”
A Review on Generative Adversarial Networks: Algorithms, Theory, and Applications (2020, background): “A research field where GANs are very successful is the application of generative models to semi-supervised learning [228], [229], as proposed but not shown in the first GANs paper [3].”
Deep Learning on Graphs: A Survey (2018, unknown): “GraphSGAN [126] used a GAN to enhance semi-supervised learning on graphs.”
Graph Contrastive Learning with Augmentations (2020, unknown): “xplored, with some work along these lines but requiring prohibitive additional computation cost [40]. Traditional self-training methods [40, 13] utilize the trained model to annotate unlabelled data; [41] 2 proposes to train a generator-classiﬁer network in the adversarial learning setting to genera”