Hardware Malware Detection using Federated Learning Technique

DOI

10.2991/978-94-6239-789-7_21How to use a DOI?

Keywords

Federated Learning; Hardware malware; Hardware Trojans; Malicious firmware; Internet of Things

Abstract

Hardware malware poses a significant threat to modern computing systems, exploiting vulnerabilities in hardware components such as microprocessors, firmware, and integrated circuits. Unlike traditional software-based threats, hardware malware operates at a lower level, making it highly stealthy, persistent, and challenging to detect. This type of malware refers to malicious modifications or logic inserted into hardware components, including microprocessors, firmware, and integrated circuits, during the design, manufacturing, or operational phases. These modifications can take the form of hardware Trojans, malicious firmware, or micro-architectural exploits that compromise system integrity. For example, hardware Trojans can remain dormant until triggered by specific conditions, allowing attackers to manipulate, exfiltrate, or destroy critical data without being detected. This paper provides a comprehensive overview of the evolving landscape of hardware malware, categorizing attack types such as hardware Trojans, malicious firmware, and side-channel exploits. It also explores state-of-the-art detection techniques, ranging from static and dynamic analysis to emerging machine learning approaches, while discussing mitigation strategies such as secure hardware design, supply chain auditing, and runtime defenses like Trusted Execution Environments. Apart from the above discussion, the article suggests an intrusion detection framework based on Federated Learning (FL) for detecting microarchitecture attacks such as Spectre, Meltdown, Fallout, and ZombieLoad. This approach allows several clients to train a global model without exchanging their actual data, hence preserving the privacy of the data. The use of a feature optimization method based on dimensionality reduction methods is included to improve the efficiency of the learning algorithm and decrease computational overheads. Experiments reveal that the suggested model performs very well with high accuracy in classifying various attacks.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Sarthak Gupta
AU  - Deepankar Bharadwaj
AU  - Amit Kumar
AU  - Abhay Pratap Singh Bhadauria
AU  - Amrita Kumari
AU  - Ashish Saini
PY  - 2026
DA  - 2026/10/08
TI  - Hardware Malware Detection using Federated Learning Technique
BT  - Proceedings of theInternational Conference on Innovation and Practices in Artificial Intelligence (ICIPAI 2026)
PB  - Atlantis Press
SP  - 296
EP  - 302
SN  - 1951-6851
UR  - https://doi.org/10.2991/978-94-6239-789-7_21
DO  - 10.2991/978-94-6239-789-7_21
ID  - Gupta2026
ER  -