Real-world data streams pose a unique challenge to the implementation of machine learning (ML) models and data analysis. A notable problem that has been introduced by the growth of Internet of Things (IoT) deployments across the smart city ecosystem is that the statistical properties of data streams can change over time, resulting in poor prediction performance and ineffective decisions. While concept drift detection methods aim to patch this problem, emerging communication and sensing technologies are generating a massive amount of data, requiring distributed environments to perform computation tasks across smart city administrative domains. In this article, we implement and test a number of state-of-the-art active concept drift detection algorithms for time series analysis within a distributed environment. We use real-world data streams and provide critical analysis of results retrieved. The challenges of implementing concept drift adaptation algorithms, along with their applications in smart cities, are also discussed.

The application of machine learning models, particularly in cybersecurity, has surged significantly in the past few years. However, the effectiveness of these models is predominantly tethered to the quality and breadth of the training data they ingest. The scarcity of realistic datasets within the cybersecurity field constitutes a considerable challenge to the development of industry-grade tools intended for real-world application scenarios. Specifically, current datasets are either significantly outdated or fall short on both qualitative and quantitative fronts, primarily because many organizations exhibit reluctance in data sharing, stemming from privacy concerns or the potential threat to trade secrets. To address this challenge, the paper introduces PAC-GPT, a novel framework to generate reliable synthetic data for machine learning methods based on Open AI’s Generative Pre-trained Transformer 3 (GPT-3). The core components of this framework are two modules, namely a Flow Generator, which is responsible for capturing and regenerating patterns in a series of network packets, and Packet Generator, which can generate individual network packets given the network flow. We also propose a packet generator based on LLM chaining and then proceed to assess, compare, and evaluate its performance using metrics such as loss, accuracy and success rate, concluding that transformers are a suitable approach for synthetic packet generation with minimal fine-tuning performed. Lastly, a streamlined command line interface (CLI) tool has been devised to facilitate the seamless access of this innovative data generation strategy by professionals from various disciplines.

Cities serve as vital hubs of economic activity and knowledge generation and dissemination. As such, cities bear a significant responsibility to uphold environmental protection measures while promoting the welfare and living comfort of their residents. There are diverse views on the development of smart cities, from integrating Information and Communication Technologies into urban environments for better operational decisions to supporting sustainability, wealth, and comfort of people. However, for all these cases, data are the key ingredient and enabler for the vision and realization of smart cities. This article explores the challenges associated with smart city data. We start with gaining an understanding of the concept of a smart city, how to measure that the city is a smart one, and what architectures and platforms exist to develop one. Afterwards, we research the challenges associated with the data of the cities, including availability, heterogeneity, management, analysis, privacy, and security. Finally, we discuss ethical issues. This article aims to serve as a “one-stop shop” covering data-related issues of smart cities with references for diving deeper into particular topics of interest.

The rapid integration of IoT, cloud, and edge computing has resulted in highly interconnected networks, emphasizing the need for advanced Intrusion Detection Systems (IDS) to maintain security. Successful AI-based IDS relies on high-quality data for model training. Even though a vast array of datasets from controlled settings are accessible, many fall short as they are outdated and lack the representative data of network traffic dynamics typically seen in public networks. This paper aims to advance understanding in designing testbed architectures for defense mechanisms within public networks. At its core, this research introduces a unique testbed utilizing the connectivity of panOULU Municipal public network in the city of Oulu, Finland. This experimental setup examines AI-driven security across the public network. It utilizes edge-to-cloud infrastructures, incorporating Software-Defined Networking (SDN) and Network Function Virtualization (NFV) via the VMware vSphere platform. During the training phase, a script distinguishes incoming packets as either benign or malicious based on well-defined local parameters and simulated attack scenarios. This labeled data is then utilized for training machine learning models within the Federated Learning framework, FED-ML. Subsequently, these models are evaluated on previously unseen data. The entire procedure, from traffic gathering to model training, operates without human involvement. The evaluation dataset and testbed configuration we have made publicly available through this research can deepen our understanding of the challenges in safeguarding public networks, especially those that blend various technologies in diverse environments.