Graph Structure and Feature Extrapolation for Out-of-Distribution Generalization
This addresses a domain-specific problem for graph machine learning applications, offering a tailored solution for handling distribution shifts.
The paper tackled the problem of out-of-distribution generalization in graph learning by proposing a data augmentation method that extrapolates graph structure and features, resulting in substantial and constant improvements across various tasks.
Out-of-distribution (OOD) generalization deals with the prevalent learning scenario where test distribution shifts from training distribution. With rising application demands and inherent complexity, graph OOD problems call for specialized solutions. While data-centric methods exhibit performance enhancements on many generic machine learning tasks, there is a notable absence of data augmentation methods tailored for graph OOD generalization. In this work, we propose to achieve graph OOD generalization with the novel design of non-Euclidean-space linear extrapolation. The proposed augmentation strategy extrapolates both structure and feature spaces to generate OOD graph data. Our design tailors OOD samples for specific shifts without corrupting underlying causal mechanisms. Theoretical analysis and empirical results evidence the effectiveness of our method in solving target shifts, showing substantial and constant improvements across various graph OOD tasks.