Data Augmentation for Deep Graph Learning: A Survey

Graph neural networks, as powerful deep learning tools to model graph-structured data, have demonstrated remarkable performance on numerous graph learning tasks. To counter the data noise and data scarcity issues in deep graph learning (DGL), increasing graph data augmentation research has been conducted lately. However, conventional data augmentation methods can hardly handle graph-structured data which is defined on non-Euclidean space with multi-modality. In this survey, we formally formulate the problem of graph data augmentation and further review the representative techniques in this field. Specifically, we first propose a taxonomy for graph data augmentation and then provide a structured review by categorizing the related work based on the augmented information modalities. Focusing on the two challenging problems in DGL (i.e., optimal graph learning and low-resource graph learning), we also discuss and review the existing learning paradigms which are based on graph data augmentation. Finally, we point out a few directions and challenges on promising future works.