Graph Transplant: Node Saliency-Guided Graph Mixup with Local Structure Preservation
This work addresses a domain-specific problem for graph neural network practitioners by providing an incremental but effective augmentation technique for graph classification tasks.
The paper tackles the challenge of applying Mixup-like data augmentation to irregular graph-structured data by introducing Graph Transplant, a method that mixes graphs at the graph-level using node saliency to select meaningful subgraphs and preserve local structure, resulting in consistent performance improvements over baselines across multiple benchmarks.
Graph-structured datasets usually have irregular graph sizes and connectivities, rendering the use of recent data augmentation techniques, such as Mixup, difficult. To tackle this challenge, we present the first Mixup-like graph augmentation method at the graph-level called Graph Transplant, which mixes irregular graphs in data space. To be well defined on various scales of the graph, our method identifies the sub-structure as a mix unit that can preserve the local information. Since the mixup-based methods without special consideration of the context are prone to generate noisy samples, our method explicitly employs the node saliency information to select meaningful subgraphs and adaptively determine the labels. We extensively validate our method with diverse GNN architectures on multiple graph classification benchmark datasets from a wide range of graph domains of different sizes. Experimental results show the consistent superiority of our method over other basic data augmentation baselines. We also demonstrate that Graph Transplant enhances the performance in terms of robustness and model calibration.