Secure Aggregation Protocols for Federated Learning in IoT Intrusion Detection
DOI:
https://doi.org/10.15680/IJCTECE.2021.0404002Keywords:
Secure Aggregation, Federated Learning, IoT Intrusion Detection Systems (IDS), Data Privacy, Cybersecurity, IoT Security, Decentralized Machine LearningAbstract
The Internet of Things (IoT) offers vast opportunities for automation and smart devices, but its widespread adoption has significantly increased the risk of cyber threats. Intrusion detection in IoT systems is crucial to ensuring the integrity and security of IoT networks. Federated learning, an emerging approach that allows decentralized machine learning on IoT devices, can enhance intrusion detection systems without compromising privacy. However, the challenge of securely aggregating model updates across distributed devices remains a critical issue. This paper investigates Secure Aggregation Protocols in federated learning for IoT-based Intrusion Detection Systems (IDS). We propose a protocol that ensures data privacy and integrity while minimizing communication overhead. The results demonstrate that secure aggregation techniques can effectively improve the accuracy and privacy of IoT intrusion detection systems, ensuring the safety of sensitive data while reducing the risk of attacks.
References
1. Agarwal, S., & Ruan, W. Lightweight encryption for federated learning in IoT environments. IEEE Transactions on Mobile Computing, 19(2), 422-435.
2. Bonawitz, K., McMahan, H. B., & Ramage, D. Practical secure aggregation for privacy-preserving machine learning. In Proceedings of the 2017 ACM SIGSAC Conference on Computer and Communications Security.
3. Li, T., & Zhang, Y. Efficient secure aggregation for federated learning. IEEE Transactions on Cloud Computing, 8(4), 1-13.
4. McMahan, H. B., Moore, E., & Ramage, D. Communication-efficient learning of deep networks from decentralized data. In Proceedings of the 20th International Conference on Artificial Intelligence and Statistics
5. Shokri, R., & Shmatikov, V. Privacy-preserving deep learning. In Proceedings of the 22nd ACM SIGSAC Conference on Computer and Communications Security (CCS 2015).
