Can Modifying Data Address Graph Domain Adaptation?

TL;DR

This paper introduces GraphAlign, a data-centric approach for unsupervised graph domain adaptation that modifies the source graph to improve transferability, outperforming model-centric methods.

Contribution

The paper proposes GraphAlign, a novel data-centric UGDA method that generates a small, transferable graph, guided by theoretical principles, to enhance knowledge transfer in graph neural networks.

Findings

GraphAlign outperforms baselines by 2.16% on average.

Training on a graph as small as 0.25-1% of original yields strong results.

Data-centric approach shows significant potential over model-centric methods.

Abstract

Graph neural networks (GNNs) have demonstrated remarkable success in numerous graph analytical tasks. Yet, their effectiveness is often compromised in real-world scenarios due to distribution shifts, limiting their capacity for knowledge transfer across changing environments or domains. Recently, Unsupervised Graph Domain Adaptation (UGDA) has been introduced to resolve this issue. UGDA aims to facilitate knowledge transfer from a labeled source graph to an unlabeled target graph. Current UGDA efforts primarily focus on model-centric methods, such as employing domain invariant learning strategies and designing model architectures. However, our critical examination reveals the limitations inherent to these model-centric methods, while a data-centric method allowed to modify the source graph provably demonstrates considerable potential. This insight motivates us to explore UGDA from a data-centric perspective. By revisiting the theoretical generalization bound for UGDA, we identify two data-centric principles for UGDA: alignment principle and rescaling principle. Guided by these principles, we propose GraphAlign, a novel UGDA method that generates a small yet transferable graph. By exclusively training a GNN on this new graph with classic Empirical Risk Minimization (ERM), GraphAlign attains exceptional performance on the target graph. Extensive experiments under various transfer scenarios demonstrate the GraphAlign outperforms the best baselines by an average of 2.16%, training on the generated graph as small as 0.25~1% of the original training graph.