Explaining and Adapting Graph Conditional Shift

TL;DR

This paper analyzes the vulnerability of Graph Neural Networks to distribution shifts, quantifies the causes of this shift, and proposes a method to mitigate it, improving robustness in graph classification tasks.

Contribution

It provides a theoretical analysis of conditional shift in GNNs and introduces an unsupervised domain adaptation method to reduce this shift, enhancing model robustness.

Findings

Conditional shift is worsened by graph heterophily and model architecture.

Proposed method reduces conditional shift and improves ROC AUC by up to 10%.

Method demonstrates robustness across node and graph classification under distribution shifts.

Abstract

Graph Neural Networks (GNNs) have shown remarkable performance on graph-structured data. However, recent empirical studies suggest that GNNs are very susceptible to distribution shift. There is still significant ambiguity about why graph-based models seem more vulnerable to these shifts. In this work we provide a thorough theoretical analysis on it by quantifying the magnitude of conditional shift between the input features and the output label. Our findings show that both graph heterophily and model architecture exacerbate conditional shifts, leading to performance degradation. To address this, we propose an approach that involves estimating and minimizing the conditional shift for unsupervised domain adaptation on graphs. In our controlled synthetic experiments, our algorithm demonstrates robustness towards distribution shift, resulting in up to 10% absolute ROC AUC improvement versus the second-best algorithm. Furthermore, comprehensive experiments on both node classification and graph classification show its robust performance under various distribution shifts.