The Internet of Things (IoT) has become increasingly popular, and opened new possibilities for applications in various domains. However, the IoT also poses security challenges due to the limited resources of the devices and its dynamic network topology. Routing attacks on 6LoWPAN-based IoT devices can be particularly challenging to detect because of its unique characteristics of the network. In recent years, several techniques have been proposed for detecting routing attacks, including anomaly detection. These techniques leverage different features of network traffic to identify and classify routing attacks. This paper focuses on routing attacks that target the Routing Protocol for Low-Power and Lossy Networks (RPL), which are widely used in 6LoWPAN-based IoT systems. The attacks discussed in this paper can be categorized as either inherited from Wireless Sensor Networks or exploiting vulnerabilities unique to RPL (known as RPL-specific attacks). The paper describes various RPL attacks, including Flood Attacks, Data-DoS/DDoS Attacks, Wormhole Attacks, RPL Rank Attacks, Blackhole Attacks, Version Attacks, and Sinkhole Attacks. In this paper, a novel Hybrid Intrusion Detection System (HIDS) that combines a decision tree classifier and a one-class Support Vector Machine classifier is proposed to detect routing attacks. The HIDS draws on the strengths of both a Signature Intrusion Detection System (SIDS) and an Anomaly-based Intrusion Detection System (AIDS) to identify routing attacks with a high degree of accuracy and a low false alarm rate. The routing dataset, which features genuine IoT network traffic and various kinds of routing attacks, was used to test the proposed HIDS. According to the findings, the hybrid IDS proposed in this study outperforms SIDS and AIDS approaches, with higher detection rates and lower false positive rates.

A data stream can be considered as a sequence of examples that arrive continuously and are potentially unbounded, such as web page visits, sensor readings and call records. One of the serious and challenging problems that appears in a data stream is concept drift. This problem occurs when the relation between the input data and the target variable changes over time. Most existing works make an optimistic assumption that all incoming data are labelled and the class labels are available immediately. However, such an assumption is not always valid. Therefore, a lack of class labels aggravates the problem of concept drift detection. With this motivation, we propose a drift detector that reacts naturally to sudden drifts in the absence of class labels. In a novel way, the proposed detector reacts to concept drift in the absence of class labels, where the true label of an example is not necessary. Instead of monitoring the error estimates, the proposed detector monitors the diversity of a pair of classifiers, where the true label of an example is not necessary to determine whether components disagree. Using several datasets, an experimental evaluation and comparison is conducted against several existing detectors. The experiment results show that the proposed detector can detect drifts with less delay, runtime and memory usage.

Federated Learning (FL) is revolutionizing Machine Learning (ML) by enabling devices in different locations to collaborate and learn from user-generated data without centralizing it. In dynamic and non-stationary environments like Internet of Things (IoT), Concept Drift (CD) is the phenomenon of data changing/evolving over time. Traditional FL frameworks struggle to maintain performance when local data distributions evolve, as they lack mechanisms for detecting and adapting to concept drift. However, the use of FL in such environments, where data changing/evolving continuously and Continual Learning (CL) is required to adapt to concept drift, remains a relatively unexplored area. This study specifically addresses this gap by examining strategies for continuous adaptation within federated systems when faced with non-stationary data. Following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) guidelines, this study systematically reviews existing literature on FL adaptation to concept drift. To the best of our knowledge, this is the first systematic review that consolidates and reinterprets existing studies under the emerging framework of Federated Drift-Aware Learning (FDAL), bridging Federated and Continual Learning research toward adaptive and drift-resilient federated systems. We conducted an extensive systematic survey, including an analysis of state-of-the-art methods and the latest developments in this field. Our study highlights their strengths, weaknesses, and datasets used, identifies key challenges faced by FL systems in these scenarios, and explores potential future directions. Additionally, we categorize the limitations and future directions into major thematic areas that highlight core gaps and research opportunities. The results of this study will support researchers in overcoming the adaptation challenges that FL systems face when dealing with changing environments due to concept drift and serve as a critical resource for advancing adaptive federated intelligence.

As machine learning models are increasingly applied to real-world scenarios, it is essential to consider the possibility of changes in the data distribution over time. Concept drift detection and adaptation refers to the process of identifying and tracking these changes and updating the model accordingly. Researchers have devoted significant efforts to develop various techniques and tools for concept drift detection and adaptation, as this paper provides a generic roadmap and review of the field. In this paper, we begin by reviewing the background of data stream classification and its assumptions and requirements. Then, we explore the historical development of concept drift detection and adaptation and highlight the key points of approaches that have emerged over time. Next, we summarize the major findings, challenges, and limitations of past research, and provide insights into potential future directions of the field. The paper can benefit researchers and practitioners who seek to navigate the challenges and opportunities in concept drift detection and adaptation.