Federated Learning-Based Anomaly Detection in 5G Networks

Strategic Insights -

Researchers at the University of Portsmouth have proposed a novel framework, FedLLMGuard, to address the rising security challenges in 5G networks. This framework integrates Federated Learning (FL) with Large Language Models (LLMs) for real-time, privacy-preserving, and adaptive anomaly detection. FedLLMGuard has been evaluated against traditional Machine learning-based Intrusion Detection Systems (IDS) and demonstrated superior performance, detecting threats rapidly and mitigating attacks effectively while ensuring data privacy.

Key Takeaways:

  • FedLLMGuard is a novel framework that integrates Federated Learning (FL) with Large Language Models (LLMs) for real-time, privacy-preserving, and adaptive anomaly detection in 5G networks.
  • FedLLMGuard uses FL for decentralized learning and LLM for contextual traffic analysis and interoperability.
  • The framework introduces the CorruptNet adversarial attack, a novel poisoning strategy targeting FL-based anomaly detection, ensuring robustness evaluation under adversarial conditions.
  • FedLLMGuard outperformed three benchmark methods - Random Forest, LSTM, and PSO Autoencoder LSTM - on three benchmark datasets: TII-SSRC-23, CICDDoS2019, and NF-UNSW-NB15.
  • Under the CorruptNet attack, FedLLMGuard achieved an accuracy of 98.64%, a false positive rate of only 2.16%, and ultra-low detection latency (0.0113s).
  • The research concluded that FedLLMGuard is a scalable and resource-efficient security solution for 5G networks, capable of detecting threats rapidly and mitigating attacks effectively while ensuring data privacy.
  • The study has been peer-reviewed and published in the journal Computer Networks.

Statistics:

  • Accuracy: 98.64% under the CorruptNet attack
  • False Positive Rate: 2.16% under the CorruptNet attack
  • Detection Latency: 0.0113s under the CorruptNet attack
  • Number of Benchmark Methods: 3 (Random Forest, LSTM, and PSO Autoencoder LSTM)
  • Number of Benchmark Datasets: 3 (TII-SSRC-23, CICDDoS2019, and NF-UNSW-NB15)

Sources:

  • NewsRx. Investigators at University of Portsmouth Describe Findings in Technology (Fedllmguard: a Federated Large Language Model for Anomaly Detection In 5g Networks). Journal of Engineering. September 1, 2025; p 1244.
  • Fedllmguard: a Federated Large Language Model for Anomaly Detection In 5g Networks. Computer Networks, 2025;269.