Homophily Enhanced Graph Domain Adaptation

2025-05-26

AI Summary

Graph Domain Adaptation (GDA) suffers from performance degradation when source and target graphs exhibit heterophily—i.e., structural dissimilarity in node homophily—yet this critical issue has been largely overlooked by existing methods. This work is the first to theoretically and empirically demonstrate that heterophily mismatch severely impairs knowledge transfer in GDA. To address it, we propose the first explicit paradigm for modeling and aligning graph homophily across domains: a hybrid graph convolutional filter jointly optimizes graph signal smoothness and homophily-aware alignment, thereby enforcing structural consistency between source and target graphs. Evaluated on multiple benchmark datasets, our method achieves an average 5.2% improvement in cross-graph classification accuracy. These results validate homophily alignment as a key factor for generalization in GDA, offering both an interpretable, optimization-friendly theoretical framework and a practical technical pathway for robust graph domain adaptation.

Technology Category

• Machine Learning: Transfer, Domain Adaptation, Multi-Task Learning
• Data Mining & Knowledge Management: Graph Mining, Social Network Analysis & Community
• Application Domains: Humanities & Computational Social Science

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Semantics and Knowledge: Scalable techniques for the creation, curation, publication, maintenance, and consumption of large, Web-based, structured, reusable, knowledge graphs and ontologies
• Search and Retrieval-Augmented AI: Vertical and domain-specific search

Abstract

Graph Domain Adaptation (GDA) transfers knowledge from labeled source graphs to unlabeled target graphs, addressing the challenge of label scarcity. In this paper, we highlight the significance of graph homophily, a pivotal factor for graph domain alignment, which, however, has long been overlooked in existing approaches. Specifically, our analysis first reveals that homophily discrepancies exist in benchmarks. Moreover, we also show that homophily discrepancies degrade GDA performance from both empirical and theoretical aspects, which further underscores the importance of homophily alignment in GDA. Inspired by this finding, we propose a novel homophily alignment algorithm that employs mixed filters to smooth graph signals, thereby effectively capturing and mitigating homophily discrepancies between graphs. Experimental results on a variety of benchmarks verify the effectiveness of our method.

Problem

• Addressing label scarcity in graph domain adaptation
• Highlighting overlooked homophily discrepancies in benchmarks
• Proposing homophily alignment to mitigate performance degradation

Innovation

• Employs mixed filters for graph signal smoothing
• Addresses homophily discrepancies in graph alignment
• Enhances Graph Domain Adaptation via homophily alignment