GraphCrop: Subgraph Cropping for Graph Classification

TL;DR

GraphCrop introduces a node-centric subgraph cropping data augmentation technique for GNNs, improving graph classification by promoting global structural understanding and enhancing model generalization.

Contribution

The paper proposes GraphCrop, a simple, parameter-free subgraph cropping method that improves GNN performance by regularizing training and simulating real-world structural noise.

Findings

Significant accuracy improvements on multiple datasets.

Enhanced GNN generalization and robustness.

Easy integration into existing GNN frameworks.

Abstract

We present a new method to regularize graph neural networks (GNNs) for better generalization in graph classification. Observing that the omission of sub-structures does not necessarily change the class label of the whole graph, we develop the \textbf{GraphCrop} (Subgraph Cropping) data augmentation method to simulate the real-world noise of sub-structure omission. In principle, GraphCrop utilizes a node-centric strategy to crop a contiguous subgraph from the original graph while maintaining its connectivity. By preserving the valid structure contexts for graph classification, we encourage GNNs to understand the content of graph structures in a global sense, rather than rely on a few key nodes or edges, which may not always be present. GraphCrop is parameter learning free and easy to implement within existing GNN-based graph classifiers. Qualitatively, GraphCrop expands the existing training set by generating novel and informative augmented graphs, which retain the original graph labels in most cases. Quantitatively, GraphCrop yields significant and consistent gains on multiple standard datasets, and thus enhances the popular GNNs to outperform the baseline methods.