Federated-Learning Intrusion Detection System Based Blockchain Technology

Authors:
  • EGEbrahim A. A. Ghaleb
    Ebrahim A. A. Ghaleb
  • NANur Arifin Akbar
    Nur Arifin Akbar
  • LALayla Asiri
    Layla Asiri
  • MAMeaad Alrehaili
    Meaad Alrehaili
  • AAAskar Altalidi
    Askar Altalidi
Open Access
Publisher website
Google Scholar
Abstract
This study presents the implementation of a blockchain-based federated-learning (FL) intrusion detection system. This approach utilizes machine learning (ML) instead of traditional signature-based methods, enabling the system to detect new attack types. The FL technique ensures the privacy of sensitive data while still utilizing the large amounts of data distributed across client devices. To achieve this, we employed the federated averaging method and incorporated a custom preprocessing stage for data standardization. The use of blockchain technology in combination with FL created a fully decentralized and open learning system capable of overcoming new security challenges.

This publication has 17 references indexed in Scilit: