Multi-source Unsupervised Domain Adaptation on Graphs with Transferability Modeling

TL;DR

This paper introduces a novel framework for multi-source unsupervised domain adaptation on graphs, leveraging transferability modeling and selective data alignment to improve node classification across diverse graph structures.

Contribution

The work proposes a new framework with transferability-based source selection and bi-level alignment for effective MSUDA on graphs, addressing challenges of diverse structures and distribution discrepancies.

Findings

Outperforms existing methods on five graph datasets.

Effectively selects informative source data via transferability measures.

Achieves better node classification accuracy across domains.

Abstract

In this paper, we tackle a new problem of \textit{multi-source unsupervised domain adaptation (MSUDA) for graphs}, where models trained on annotated source domains need to be transferred to the unsupervised target graph for node classification. Due to the discrepancy in distribution across domains, the key challenge is how to select good source instances and how to adapt the model. Diverse graph structures further complicate this problem, rendering previous MSUDA approaches less effective. In this work, we present the framework Selective Multi-source Adaptation for Graph ({\method}), with a graph-modeling-based domain selector, a sub-graph node selector, and a bi-level alignment objective for the adaptation. Concretely, to facilitate the identification of informative source data, the similarity across graphs is disentangled and measured with the transferability of a graph-modeling task set, and we use it as evidence for source domain selection. A node selector is further incorporated to capture the variation in transferability of nodes within the same source domain. To learn invariant features for adaptation, we align the target domain to selected source data both at the embedding space by minimizing the optimal transport distance and at the classification level by distilling the label function. Modules are explicitly learned to select informative source data and conduct the alignment in virtual training splits with a meta-learning strategy. Experimental results on five graph datasets show the effectiveness of the proposed method.