Securing Federated Learning through Blockchain and Explainable AI for Robust Intrusion Detection in IoT Networks

Federated learning (FL) is a distributed machine learning technique that allows multiple devices or nodes in a network to collaboratively train a machine learning model while keeping their data local to the device. This is particularly useful in the context of the Internet of Things (IoT) systems, where devices may have limited computational resources and may not be able to transmit their data to a central server due to privacy or bandwidth constraints. FL offers privacy protection but remains vulnerable to security and privacy attacks (e.g., Data Poisoning Attacks). To address this issue, in addition to the usual components of an FL system, an explainable FL framework for intrusion detection systems should also include mechanisms for explaining the model's predictions. This can be achieved through the use of techniques such as feature importance, which allows the model to identify the most important input features for a particular prediction. In this context, we propose a novel framework, called FedIoT, that leverages Explainable Artificial Intelligence (XAI) techniques and Blockchain to secure FL-based IDS in the IoT networks. FedIoT uses advanced XAI techniques to identify local model manipulations and mitigate FL-based attacks. Moreover, we propose a blockchain-based approach that uses an efficient reputation scheme that ensures the trustworthiness and reliability of the FL training process. We conduct experiments to validate FedIoT, an FL-based intrusion detection system in IoT networks. Using the UNSW-NB15 dataset, we confirm that FedIoT can effectively detect malicious activities and facilitate efficient FL collaboration among multiple users.