Graph Out-of-Distribution Generalization With Controllable Data Augmentation

Graph Neural Network (GNN) has demonstrated extraordinary performance in classifying graph properties. However, due to the selection bias of training and testing data (e.g., training on small graphs and testing on large graphs, or training on dense graphs and testing on sparse graphs), distribution...

Full description

Saved in:
Bibliographic Details
Published inIEEE transactions on knowledge and data engineering Vol. 36; no. 11; pp. 6317 - 6329
Main Authors Lu, Bin, Zhao, Ze, Gan, Xiaoying, Liang, Shiyu, Fu, Luoyi, Wang, Xinbing, Zhou, Chenghu
Format Journal Article
LanguageEnglish
Published IEEE 01.11.2024
Subjects
Correlation
Data augmentation
domain generalization
Extraterrestrial measurements
graph neural network
Manifolds
Out-of-distribution generalization
Perturbation methods
Testing
Training
Online AccessGet full text

Cover

More Information
Summary:Graph Neural Network (GNN) has demonstrated extraordinary performance in classifying graph properties. However, due to the selection bias of training and testing data (e.g., training on small graphs and testing on large graphs, or training on dense graphs and testing on sparse graphs), distribution deviation is widespread. More importantly, we often observe hybrid structure distribution shift of both scale and density, despite of one-sided biased data partition. The spurious correlations over hybrid distribution deviation degrade the performance of previous GNN methods and show large instability among different datasets. To alleviate this problem, we propose OOD-GMixup to jointly manipulate the training distribution with controllable data augmentation in metric space. Specifically, we first extract the graph rationales to eliminate the spurious correlations due to irrelevant information. Second, we generate virtual samples with perturbation on graph rationale representation domain to obtain potential OOD training samples. Finally, we propose OOD calibration to measure the distribution deviation of virtual samples by leveraging Extreme Value Theory, and further actively control the training distribution by emphasizing the impact of virtual OOD samples. Extensive studies on several real-world datasets on graph classification demonstrate the superiority of our proposed method over state-of-the-art baselines.
ISSN:1041-4347
1558-2191
DOI:10.1109/TKDE.2024.3393109