Title: Collaborative AI-based malware detection through reliable clustered federated learning

Authors: Elaf Almushiti; Tarek Moulahi

Addresses: Department of Information Technology, College of Computer, Qassim University, Buraydah, Saudi Arabia ' Department of Information Technology, College of Computer, Qassim University, Buraydah, Saudi Arabia

Abstract: Artificial intelligence (AI) has become integral to enhancing decision-making and personalising experiences across various domains. However, data privacy remains a major challenge when training machine learning models. This study proposes a novel framework combining an innovative aggregation technique and lightweight differential privacy (DP) to secure communications in federated learning (FL). The framework supports AI-based malware detection using eight clients organised into three clustering configurations. Performance is evaluated using five classifiers: random forest (RFC), decision tree (DTree), extreme gradient boosting (XGB), support vector classifier (SVC), and multi-layer perceptron (MLP). The XGB classifier achieved the highest accuracy, up to 99.6%. Results show that the framework maintains high accuracy while preserving data privacy, offering a promising solution for secure AI deployment in sensitive sectors such as finance, healthcare, and cybersecurity.

Keywords: federated learning; clustering; security; differential privacy; collaboration; malware detection.

DOI: 10.1504/IJCCBS.2024.146784

International Journal of Critical Computer-Based Systems, 2024 Vol.11 No.4, pp.307 - 326

Received: 22 Jun 2024
Accepted: 21 Feb 2025
Published online: 17 Jun 2025 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article