Federated Learning-Based Intrusion Detection System for IoT Environments with Locally Adapted Model
Souradip Roy, Juan Li, Yan Bai
Abstract
As the Internet of Things (IoT) becomes more prevalent, the need for intrusion detection systems (IDS) to protect against cyberattacks increases. However, the limited computing capabilities of IoT devices often require sending data to a centralized cloud for analysis, which can cause energy consumption, privacy issues, and data leakage. To address these problems, we propose a Federated Learning-based IDS that distributes learning to local devices without sending data to a centralized cloud. We also create lightweight local learners to accommodate IoT device limitations and locally adapted models to handle non-independent intrusion data distribution. We evaluate our method using NBaIoT and CICIDS-2017 datasets, and our results demonstrate comparable performance to centralized learning on metrics including accuracy, precision, and recall, while addressing privacy and data leakage concerns.