Collaborate to Adapt: Source-Free Graph Domain Adaptation via Bi-directional Adaptation

Zhen Zhang, Meihan Liu, an hui wang, Hongyang Chen, Zhao Li, Jiajun Bu, Bingsheng He

Keywords: Graph Neural Networks, Source-Free Unsupervised Graph Domain Adaptation

TL;DR: This paper proposes a novel source-free graph domain adaptation framework that performs model adaptation and graph adaptation collaboratively.

Abstract: Unsupervised graph domain adaptation has emerged as a practical solution to transfer knowledge from a label-rich source graph to a completely unlabelled target graph, when there is a scarcity of labels in target graph. However, most of existing methods require a labelled source graph to provide supervision signals, which might not be accessible in the real-world scenarios due to regulations and privacy concerns. In this paper, we explore the scenario of source-free unsupervised graph domain adaptation, which tries to address the domain adaptation problem without accessing the labelled source graph. Specifically, we present a novel paradigm called GraphCTA, which performs model adaptation and graph adaptation collaboratively through a series of procedures: (1) conduct model adaptation based on node's neighborhood predictions in target graph considering both local and global information; (2) perform graph adaptation by updating graph structure and node attributes via neighborhood constrastive learning; and (3) the updated graph serves as an input to facilitate the subsequent iteration of model adaptation, thereby establishing a collaborative loop between model adaptation and graph adaptation. Comprehensive experiments are conducted on various public datasets including transaction, social, and citation graphs. The experimental results demonstrate that our proposed model outperforms recent source free baselines by large margins. Our source code and datasets are available at https://anonymous.4open.science/r/GraphCTA-code.

Track: Graph Algorithms and Learning for the Web

Submission Guidelines Scope: Yes

Submission Guidelines Blind: Yes

Submission Guidelines Format: Yes

Submission Guidelines Limit: Yes

Submission Guidelines Authorship: Yes

Student Author: No

Submission Number: 1186