Paper 2025/192 Original: eprint.iacr.org/2025/192 ↗
Practical Electromagnetic Fault Injection on Intel Neural Compute Stick 2
Abstract
Machine learning (ML) has been widely deployed in various applications, with many applications being in critical infrastructures. One recent paradigm is edge ML, an implementation of ML on embedded devices for Internet-of-Things (IoT) applications. In this work, we have conducted a practical experiment on Intel Neural Compute Stick (NCS) 2, an edge ML device, with regard to fault injection (FI) attacks. More precisely, we have employed electromagnetic fault injection (EMFI) on NCS 2 to evaluate the practicality of the attack on a real target device. We have investigated multiple fault parameters with a low-cost pulse generator, aiming to achieve misclassification at the output of the inference. Our experimental results demonstrated the possibility of achieving practical and repeatable misclassifications.
Metadata
- Available format(s)
PDF- Category
- Implementation
- Publication info
- Preprint.
- History
- 2025-02-10: last updated
- 2025-02-10: published
- Short URL
- https://ia.cr/2025/192
- License
- https://creativecommons.org/publicdomain/zero/1.0/
BibTeX
@misc{cryptoeprint:2025/192,
author = {Shivam Bhasin and Dirmanto Jap and Marina Krček and Stjepan Picek and Prasanna Ravi},
title = {Practical Electromagnetic Fault Injection on Intel Neural Compute Stick 2},
howpublished = {Cryptology {ePrint} Archive, Paper 2025/192},
year = {2025},
url = {https://eprint.iacr.org/2025/192}
}