ImGAGN:Imbalanced Network Embedding via Generative Adversarial Graph Networks

Abstract

Imbalanced classification on graphs is ubiquitous yet challenging in many real-world applications, such as fraudulent node detection. Recently, graph neural networks (GNNs) have shown promising performance on many network analysis tasks. However, most existing GNNs have almost exclusively focused on the balanced networks, and would get unappealing performance on the imbalanced networks. To bridge this gap, in this paper, we present a generative adversarial graph network model, called ImGAGN to address the imbalanced classification problem on graphs. It introduces a novel generator for graph structure data, named GraphGenerator, which can simulate both the minority class nodes' attribute distribution and network topological structure distribution by generating a set of synthetic minority nodes such that the number of nodes in different classes can be balanced. Then a graph convolutional network (GCN) discriminator is trained to discriminate between real nodes and fake (i.e., generated) nodes, and also between minority nodes and majority nodes on the synthetic balanced network. To validate the effectiveness of the proposed method, extensive experiments are conducted on four real-world imbalanced network datasets. Experimental results demonstrate that the proposed method Im-GAGN outperforms state-of-the-art algorithms for semi-supervised imbalanced node classification task. CCS CONCEPTS• Computing methodologies → Neural networks.

Imbalanced Graph Classification via Graph-of-Graph Neural Networks

“…Imbalance on graph-structured data could lie either in the node or graph domain where nodes (graphs) in different classes have different amount of training data. Nearly all existing related GNN works focus on imbalanced node classification by either pre-training or adversarial training to reconstruct the graph topology [13,14,15,12,16], while to the best of our knowledge, imbalanced graph classification remains largely unexplored. On one hand, unlike node classification where we can derive extra supervision for minority nodes from their neighborhoods via propagation, graphs are individual instances that are isolated from each other and thus we cannot aggregate information directly from other graphs by propagation.…”

“…GraphSMOTE [15] attempts to generate edges by pre-training an edge generator for isolated synthetic nodes from SMOTE [12]. Most recently, imGAGN [16] simulates both distributions of node attributes in minority classes and graph structures via generative adversarial graph network model. However, all of these recent works are proposed for node imbalance classification and to the best of our knowledge, graph imbalance classification with GNNs remains largely unexplored.…”