Federated Learning for Privacy-Preserving Intrusion Detection: A Systematic Review, Taxonomy, Challenges and Future Directions
Dattatray Raghunath Kale, Swati Shirke-Deshmukh, Amulkumar Jadhav, Shrihari Khatawkar, Sunny Mohite, Prof. Dr. Sarang Patil, Madhav J. Salunkhe, Rahul Ganpatrao Sonkamble · Journal of Information Systems and Telecommunication (JIST) · 2026
This paper presents a systematic review of intrusion detection systems (IDS) that leverage federated learning (FL) to enhance privacy in distributed cybersecurity environments.A total of 78 peer-reviewed studies published between 2019 and 2024 were selected using PRISMA guidelines.We categorize FL-based IDS solutions based on architecture (centralized, decentralized, hierarchical), aggregation methods (e.g., FedAvg, DAFL), and privacy-preserving techniques (e.g., differential privacy, homomorphic encryption).The survey also examines solutions to key challenges such as communication overhead, data heterogeneity, and poisoning attacks.Furthermore, this study outlines unresolved issues and proposes future research directions, including adaptive federated optimization and cross-domain deployments.This review serves as a valuable resource for researchers and practitioners aiming to develop privacy-aware, scalable, and intelligent IDS using federated learning.