Graph Data Selection for Domain Adaptation: A Model-Free Approach

TL;DR

This paper introduces GRADATE, a model-free data selection framework for graph domain adaptation that improves performance and efficiency by selecting optimal training data without relying on GNN predictions.

Contribution

The paper presents a novel, scalable, model-free data selection method for graph domain adaptation using optimal transport, complementing existing GDA techniques.

Findings

GRADATE outperforms existing data selection methods.

It enhances GDA methods with fewer training data.

Demonstrates effectiveness across multiple real-world datasets.

Abstract

Graph domain adaptation (GDA) is a fundamental task in graph machine learning, with techniques like shift-robust graph neural networks (GNNs) and specialized training procedures to tackle the distribution shift problem. Although these model-centric approaches show promising results, they often struggle with severe shifts and constrained computational resources. To address these challenges, we propose a novel model-free framework, GRADATE (GRAph DATa sElector), that selects the best training data from the source domain for the classification task on the target domain. GRADATE picks training samples without relying on any GNN model's predictions or training recipes, leveraging optimal transport theory to capture and adapt to distribution changes. GRADATE is data-efficient, scalable and meanwhile complements existing model-centric GDA approaches. Through comprehensive empirical studies on several real-world graph-level datasets and multiple covariate shift types, we demonstrate that GRADATE outperforms existing selection methods and enhances off-the-shelf GDA methods with much fewer training data.