How Explanations Leak the Decision Logic: Stealing Graph Neural Networks via Explanation Alignment

TL;DR

This paper reveals security risks in explainable GNNs by demonstrating how explanations can be exploited to steal models, proposing a novel framework that effectively replicates target models' behavior and reasoning patterns.

Contribution

It introduces { ool}, a framework combining explanation alignment and guided data augmentation for efficient model stealing of GNNs using explanation mechanisms.

Findings

Our approach outperforms traditional methods in model stealing accuracy.

Explanation-based attacks pose significant security threats to GNNs.

The work emphasizes the need for protective measures in deploying explainable GNNs.

Abstract

Graph Neural Networks (GNNs) have become essential tools for analyzing graph-structured data in domains such as drug discovery and financial analysis, leading to growing demands for model transparency. Recent advances in explainable GNNs have addressed this need by revealing important subgraphs that influence predictions, but these explanation mechanisms may inadvertently expose models to security risks. This paper investigates how such explanations potentially leak critical decision logic that can be exploited for model stealing. We propose {\method}, a novel stealing framework that integrates explanation alignment for capturing decision logic with guided data augmentation for efficient training under limited queries, enabling effective replication of both the predictive behavior and underlying reasoning patterns of target models. Experiments on molecular graph datasets demonstrate that our approach shows advantages over conventional methods in model stealing. This work highlights important security considerations for the deployment of explainable GNNs in sensitive domains and suggests the need for protective measures against explanation-based attacks. Our code is available at https://github.com/beanmah/EGSteal.