FL‐IDPP: A Federated Learning Based Intrusion Detection Approach With Privacy Preservation

ABSTRACT The growing adoption of Internet of Things (IoT) devices has led to a rising concern about the security of these networks. This paper proposes a proactive intrusion recognition method, FL‐IDPP, ensuring privacy preservation for IoT networks using federated learning (FL). The proposed approach employs bidirectional recurrent neural network (RNN) models to detect anomalies and identify potential intrusions. The proposed approach ensures data privacy and efficiency in the network by storing data locally on the IoT devices and only sharing the learned model weights with the central server for FL. A high accuracy of the global machine learning (ML) model is attained by incorporating a voting ensemble process for combining updates from multiple sources. The experimental results strongly advocate for the effectiveness of the proposed approach in recognizing potential intrusions in IoT networks with enhanced accuracy and data privacy.