Adversarial Attacks on Locally Private Graph Neural Networks

2026-03-21

AI Summary

This study investigates the security and vulnerability of graph neural networks (GNNs) under local differential privacy (LDP) in the presence of adversarial attacks. By systematically evaluating the effectiveness of existing adversarial attack methods under LDP constraints, the work reveals how privacy-preserving mechanisms alter adversarial behavior and influence model robustness. It is the first to elucidate the interplay between LDP and the adversarial robustness of GNNs, demonstrating that privacy-induced perturbations can both mitigate attack efficacy and introduce new vulnerabilities. Building on these insights, the paper identifies key challenges in generating adversarial examples for LDP-protected GNNs and offers novel directions for designing graph learning frameworks that jointly optimize privacy guarantees and adversarial robustness.

Technology Category

• Machine Learning: Adversarial Learning & Robustness
• Computer Vision: Adversarial Attacks & Robustness
• Game Theory and Economic Paradigms: Adversarial Learning

Application Category

• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs
• Security and Privacy: Security and privacy of machine learning and AI applications
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

Graph neural network (GNN) is a powerful tool for analyzing graph-structured data. However, their vulnerability to adversarial attacks raises serious concerns, especially when dealing with sensitive information. Local Differential Privacy (LDP) offers a privacy-preserving framework for training GNNs, but its impact on adversarial robustness remains underexplored. This paper investigates adversarial attacks on LDP-protected GNNs. We explore how the privacy guarantees of LDP can be leveraged or hindered by adversarial perturbations. The effectiveness of existing attack methods on LDP-protected GNNs are analyzed and potential challenges in crafting adversarial examples under LDP constraints are discussed. Additionally, we suggest directions for defending LDP-protected GNNs against adversarial attacks. This work investigates the interplay between privacy and security in graph learning, highlighting the need for robust and privacy-preserving GNN architectures.

Problem

• Adversarial Attacks
• Locally Private GNNs
• Local Differential Privacy
• Graph Neural Networks
• Privacy-Security Tradeoff

Innovation

• Adversarial Attacks
• Local Differential Privacy
• Graph Neural Networks
• Privacy-Security Trade-off
• Robustness