Imbalanced Graph Classification with Multi-scale Oversampling Graph Neural Networks

TL;DR

This paper introduces MOSGNN, a multi-scale oversampling graph neural network that enhances minority class graph representations by leveraging intra- and inter-graph semantics across multiple scales, significantly improving imbalanced graph classification.

Contribution

The paper proposes a novel multi-scale oversampling framework that jointly optimizes subgraph, graph, and pairwise graph tasks to better learn discriminative features for minority classes.

Findings

MOSGNN outperforms five state-of-the-art models on 16 datasets.

The framework is flexible and improves performance with different imbalanced loss functions.

Extensive experiments validate the effectiveness of multi-scale oversampling in imbalanced graph classification.

Abstract

One main challenge in imbalanced graph classification is to learn expressive representations of the graphs in under-represented (minority) classes. Existing generic imbalanced learning methods, such as oversampling and imbalanced learning loss functions, can be adopted for enabling graph representation learning models to cope with this challenge. However, these methods often directly operate on the graph representations, ignoring rich discriminative information within the graphs and their interactions. To tackle this issue, we introduce a novel multi-scale oversampling graph neural network (MOSGNN) that learns expressive minority graph representations based on intra- and inter-graph semantics resulting from oversampled graphs at multiple scales - subgraph, graph, and pairwise graphs. It achieves this by jointly optimizing subgraph-level, graph-level, and pairwise-graph learning tasks to learn the discriminative information embedded within and between the minority graphs. Extensive experiments on 16 imbalanced graph datasets show that MOSGNN i) significantly outperforms five state-of-the-art models, and ii) offers a generic framework, in which different advanced imbalanced learning loss functions can be easily plugged in and obtain significantly improved classification performance.