An Optimized Federated Resource-Efficient IoT Security Framework with Uncertainty-Aware Distillation Loss

• Ogobuchi Daniel Okey UFABC https://orcid.org/0000-0003-0686-2763
• Demóstenes Zegarra Rodríguez UFLA https://orcid.org/0000-0001-5401-7551
• João Henrique Kleinschmidt UFABC https://orcid.org/0000-0001-6706-455X

DOI: https://doi.org/10.5753/sbseg.2026.27040

Resumo

This paper presents FedREUAD, a federated intrusion detection framework that jointly addresses client drift, communication overhead, and asymmetric knowledge transfer in distillation-based aggregation. Local model divergence under non-IID data is resolved through the FedProx’s proximal term that constrains local updates, limiting accuracy degradation on Dirichlet-simulated heterogeneity scenarios on CICIoT2023 and CICIoMT2024 datasets. During knowledge distillation, FedREUAD resolves the inherent asymmetry of Kullback-Leibler divergence (KL), which asymmetrically weights teacher-student discrepancies and propagates overconfident soft labels, by adopting a symmetric Jensen-Shannon divergence loss weighted by per-sample predictive uncertainty, penalizing the transfer of poorly calibrated soft labels from local teachers, thus reducing expected calibration error by 31% and 28% on the two datasets, respectively. Resource efficiency is achieved through distillation, structured magnitude pruning, and INT8 quantization, compressing local models by 90% and reducing per-round communication from 14.2 MB to 4.5 MB. The proposed method attains macro F1-scores of 98.22% on CICIoT2023 and 99.05% on CICIoMT2024, outperforming federated baselines under identical heterogeneity conditions, demonstrating that symmetric, uncertainty-calibrated distillation enables robust knowledge transfer without the calibration degradation inherent to asymmetric divergence formulations.

Palavras-chave: Federated Learning, Intrusion Detection System, Internet of Things, Knowledge Distillation, Jensen–Shannon Divergence, Resource Efficiency, Non-IID, CICIoT2023, CICIoMT2024

Referências

Bhatt, S., Gupta, A., and Rai, P. (2024). Federated learning with uncertainty via distilled predictive distributions. In Proceedings of the 15th Asian Conference on Machine Learning, volume 222 of Proceedings of Machine Learning Research, pages 153–168. PMLR.

Dadkhah, S., Neto, E. C. P., Ferreira, R., Molokwu, R. C., Sadeghi, S., and Ghorbani, A. A. (2024). Ciciomt2024: A benchmark dataset for multi-protocol security assessment in iomt. Internet of Things, 28:101351.

Friha, O., Ferrag, M. A., Shu, L., Maglaras, L., and Choo, K.-K. R. (2022). FELIDS: Federated learning-based intrusion detection system for agricultural Internet of Things. Journal of Parallel and Distributed Computing, 165:17–31.

Fuglede, B. and Topsoe, F. (2004). Jensen-shannon divergence and hilbert space embedding. In International symposium onInformation theory, 2004. ISIT 2004. Proceedings., page 31. IEEE.

Han, T., Nguyen, N. T., Peng, H., Xu, C., and Wang, S. (2024). A comprehensive review of model compression techniques in machine learning. Applied Intelligence, 54:11804–11844.

Hinton, G., Vinyals, O., and Dean, J. (2015). Distilling the knowledge in a neural network. arXiv preprint arXiv:1503.02531.

Itahara, S., Nishio, T., Koda, Y., Morikura, M., and Yamamoto, K. (2021). Distillation-based semi-supervised federated learning for communication-efficient collaborative training with non-iid private data. IEEE Transactions on Mobile Computing, 22(1):191–205.

Karimireddy, S. P., Kale, S., Mohri, M., Reddi, S., Stich, S., and Suresh, A. T. (2020). SCAFFOLD: Stochastic controlled averaging for federated learning. Proceedings of the 37th International Conference on Machine Learning (ICML), pages 5132–5143.

Lavaur, L., Pahl, M.-O., Busnel, Y., and Autrel, F. (2022). The evolution of federated learning-based intrusion detection and mitigation: A survey. IEEE Transactions on Network and Service Management, 19(3):2309–2332.

Li, Q., He, B., and Song, D. (2021). Model-contrastive federated learning. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition (CVPR), pages 10713–10722.

Li, T., Sahu, A. K., Zaheer, M., Sanjabi, M., Talwalkar, A., and Smith, V. (2020). Federated optimization in heterogeneous networks. Proceedings of Machine learning and systems, 2:429–450.

Liu, T., Xia, J., Ling, Z., Fu, X., Yu, S., and Chen, M. (2023). Efficient federated learning for aiot applications using knowledge distillation. IEEE Internet of Things Journal, 10(8):7229–7243.

Liu, Y., Ma, Z., Liu, X., Ma, S., Nepal, S., and Deng, R. (2022). Distributed intrusion detection system for IoT based on federated learning and edge computing. Computers & Security, 115:102622.

McMahan, B., Moore, E., Ramage, D., Hampson, S., and y Arcas, B. A. (2017). Communication-efficient learning of deep networks from decentralized data. Proceedings of the 20th International Conference on Artificial Intelligence and Statistics (AISTATS), 54:1273–1282.

Mothukuri, V., Khare, P., Parizi, R. M., Pouriyeh, S., Dehghantanha, A., and Srivastava, G. (2022). Federated-learning-based anomaly detection for IoT security attacks. IEEE Internet of Things Journal, 9(4):2545–2554.

Neto, E. C. P., Dadkhah, S., Ferreira, R., Zohourian, A., Lu, R., and Ghorbani, A. A. (2023). Ciciot2023: A real-time dataset and benchmark for large-scale attacks in iot environment. Sensors, 23(13):5941.

Okey, O. D., Rodŕıguez, D. Z., Guimarães, F. G., and Kleinschmidt, J. H. (2025). Cafiks: Communication-aware federated ids with knowledge sharing for secure iot connectivity. IEEE Access, 13:140694–140708.

Prakash, P., Ding, J., Chen, R., Qin, X., Shu, M., Cui, Q., Guo, Y., and Pan, M. (2022). Iot device friendly and communication-efficient federated learning via joint model pruning and quantization. IEEE Internet of Things Journal, 9(15):13638–13650.

Rahman, S. A., Tout, H., Talhi, C., and Mourad, A. (2020). Internet of things intrusion detection: centralized, on-device, or federated learning? IEEE Network, 34(6):310–317.

Shen, M., Zhu, L., Du, X., and Guizani, M. (2024). Federated learning ensemble knowledge distillation for intrusion detection in heterogeneous IoT networks. IEEE Transactions on Information Forensics and Security, 19:3918–3931.

Statista Research Department (2025). Number of iot devices 2015-2025. [link]. [Accessed, April 25, 2025].

Talpini, J., Civiero, N., Sartori, F., and Savi, M. (2025). A federated approach to enhance calibration of distributed ml-based intrusion detection systems. In In Proceedings of the 17th International Conference on Agents and Artificial Intelligence (ICAART 2025), volume 2, pages 840–848.

Talpini, J., Sartori, F., and Savi, M. (2024). Enhancing trustworthiness in ml-based network intrusion detection with uncertainty quantification. Journal of Reliable Intelligent Environments, 10(4):501–520.