Graph Neural Networks for Hardware Vulnerability Analysis - Can you Trust your GNN?

Lilas Alrahis; Ozgur Sinanoglu

doi:10.1109/VTS56346.2023.10140095

Graph Neural Networks for Hardware Vulnerability Analysis - Can you Trust your GNN?

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Abstract

The participation of third-party entities in the globalized semiconductor supply chain introduces potential security vulnerabilities, such as intellectual property piracy and hardware Trojan (HT) insertion. Graph neural networks (GNNs) have been employed to address various hardware security threats, owing to their superior performance on graph-structured data, such as circuits. However, GNNs are also susceptible to attacks.This work examines the use of GNNs for detecting hardware threats like HTs and their vulnerability to attacks. We present BadGNN, a backdoor attack on GNNs that can hide HTs and evade detection with a 100% success rate through minor circuit perturbations. Our findings highlight the need for further investigation into the security and robustness of GNNs before they can be safely used in security-critical applications.

Original language	English (US)
Title of host publication	Proceedings - 2023 IEEE 41st VLSI Test Symposium, VTS 2023
Publisher	IEEE Computer Society
ISBN (Electronic)	9798350346305
DOIs	https://doi.org/10.1109/VTS56346.2023.10140095
State	Published - 2023
Event	41st IEEE VLSI Test Symposium, VTS 2023 - San Diego, United States Duration: Apr 24 2023 → Apr 26 2023

Publication series

Name	Proceedings of the IEEE VLSI Test Symposium
Volume	2023-April

Conference

Conference	41st IEEE VLSI Test Symposium, VTS 2023
Country/Territory	United States
City	San Diego
Period	4/24/23 → 4/26/23

Keywords

Backdoor attacks
Graph neural networks
Hardware Trojans
Hardware security
Intellectual property

ASJC Scopus subject areas

Computer Science Applications
Electrical and Electronic Engineering

Access to Document

10.1109/VTS56346.2023.10140095

Cite this

Alrahis, L & Sinanoglu, O 2023, Graph Neural Networks for Hardware Vulnerability Analysis - Can you Trust your GNN? in Proceedings - 2023 IEEE 41st VLSI Test Symposium, VTS 2023. Proceedings of the IEEE VLSI Test Symposium, vol. 2023-April, IEEE Computer Society, 41st IEEE VLSI Test Symposium, VTS 2023, San Diego, United States, 4/24/23. https://doi.org/10.1109/VTS56346.2023.10140095

@inproceedings{0fc303f02df8416eb7c55483ede64b8d,

title = "Graph Neural Networks for Hardware Vulnerability Analysis - Can you Trust your GNN?",

abstract = "The participation of third-party entities in the globalized semiconductor supply chain introduces potential security vulnerabilities, such as intellectual property piracy and hardware Trojan (HT) insertion. Graph neural networks (GNNs) have been employed to address various hardware security threats, owing to their superior performance on graph-structured data, such as circuits. However, GNNs are also susceptible to attacks.This work examines the use of GNNs for detecting hardware threats like HTs and their vulnerability to attacks. We present BadGNN, a backdoor attack on GNNs that can hide HTs and evade detection with a 100% success rate through minor circuit perturbations. Our findings highlight the need for further investigation into the security and robustness of GNNs before they can be safely used in security-critical applications.",

keywords = "Backdoor attacks, Graph neural networks, Hardware Trojans, Hardware security, Intellectual property",

author = "Lilas Alrahis and Ozgur Sinanoglu",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 41st IEEE VLSI Test Symposium, VTS 2023 ; Conference date: 24-04-2023 Through 26-04-2023",

year = "2023",

doi = "10.1109/VTS56346.2023.10140095",

language = "English (US)",

series = "Proceedings of the IEEE VLSI Test Symposium",