Recent technological innovations along with the vast amount of available data worldwide have led to the rise of cyberattacks against network systems. Intrusion Detection Systems (IDS) play a crucial role as a defense mechanism in networks against adversarial attackers. Machine Learning methods provide various cybersecurity tools. However, these methods require plenty of data to be trained efficiently, which may be hard to collect or to use due to privacy reasons. One of the most notable Machine Learning tools is the Generative Adversarial Network (GAN), and it has great potential for tabular data synthesis. In this work, we start by briefly presenting the most popular GAN architectures, VanillaGAN, WGAN, and WGAN-GP. Focusing on tabular data generation, CTGAN, CopulaGAN, and TableGAN models are used for the creation of synthetic IDS data. Specifically, the models are trained and evaluated on an NSL-KDD dataset, considering the limitations and requirements that this procedure needs. Finally, based on certain quantitative and qualitative methods, we argue and evaluate the most prominent GANs for tabular network data synthesis.

Network intrusion detection is a key pillar towards the sustainability and normal operation of information systems. Complex threat patterns and malicious actors are able to cause severe damages to cyber-systems. In this work, we propose novel Deep Learning formulations for detecting threats and alerts on network logs that were acquired by pfSense, an open-source software that acts as firewall on FreeBSD operating system. pfSense integrates several powerful security services such as firewall, URL filtering, and virtual private networking among others. The main goal of this study is to analyse the logs that were acquired by a local installation of pfSense software, in order to provide a powerful and efficient solution that controls traffic flow based on patterns that are automatically learnt via the proposed, challenging DL architectures. For this purpose, we exploit the Convolutional Neural Networks (CNNs), and the Long Short Term Memory Networks (LSTMs) in order to construct robust multi-class classifiers, able to assign each new network log instance that reaches our system into its corresponding category. The performance of our scheme is evaluated by conducting several quantitative experiments, and by comparing to state-of-the-art formulations.

The latest technological developments are challenging for finding new solutions to mitigate the massive integration of renewable-based electricity generation in the electrical networks and to support new and dynamic energy and ancillary services markets. Smart meters have become ubiquitous equipment in the low voltage grid, enabled by the decision made in many countries to support massive deployments. The smart meter is the only equipment mandatory to be mounted when supplying a grid connected user, as it primarily has the function to measure delivered and/or produced energy on its common coupling point with the network, as technical and legal support for billing. Active distribution networks need new functionalities, to cope with the bidirectional energy flow behaviour of the grid, and many smart grid requirements need to be implemented in the near future. However there is no real coupling between smart metering systems and smart grids, as there is not yet a synergy using the opportunity of the high deployment level in smart metering. The paper presents a new approach for managing the smart metering and smart grid orchestration by presenting a new general design based on an unbundled smart meter (USM) concept, labelled as next generation open real-time smart meters (NORM), for integrating the smart meter, phasor measurement unit (PMU) and cyber-security through an enhanced smart metering gateway (SMG). NORM is intended to be deployed everywhere at the prosumer’s interface to the grid, as it is usually now done with the standard meter. Furthermore, rich data acquired from NORM is used to demonstrate the potential of assessing grid data inconsistencies at a higher level, as function to be deployed in distribution security monitoring centers, to address the higher level cyber-security threats, such as false data injections and to allow secure grid operations and complex market activities at the same time. The measures are considering only non-sensitive data from a privacy perspective, and is therefore able to be applied everywhere in the grid, down to the end-customer level, where a citizen’s personal data protection is an important aspect.

Federated learning (FL) is an emerging machine learning technique where machine learning models are trained in a decentralized manner. The main advantage of this approach is the data privacy it provides because the data are not processed in a centralized device. Moreover, the local client models are aggregated on a server, resulting in a global model that has accumulated knowledge from all the different clients. This approach, however, is vulnerable to attacks because clients can be malicious or malicious actors may interfere within the network. In the first case, these types of attacks may refer to data or model poisoning attacks where the data or model parameters, respectively, may be altered. In this paper, we investigate the data poisoning attacks and, more specifically, the label-flipping case within a federated learning system. For an image classification task, we introduce two variants of data poisoning attacks, namely model degradation and targeted label attacks. These attacks are based on synthetic images generated by a generative adversarial network (GAN). This network is trained jointly by the malicious clients using a concatenated malicious dataset. Due to dataset sample limitations, the architecture and learning procedure of the GAN are adjusted accordingly. Through the experiments, we demonstrate that these types of attacks are effective in achieving their task and managing to fool common federated defenses (stealth). We also propose a mechanism to mitigate these attacks based on clean label training on the server side. In more detail, we see that the model degradation attack causes an accuracy degradation of up to 25%, while common defenses can only alleviate this for a percentage of ∼5%. Similarly, the targeted label attack results in a misclassification of 56% compared to 2.5% when no attack takes place. Moreover, our proposed defense mechanism is able to mitigate these attacks.

Autonomous fault detection plays a major role in the Critical Energy Infrastructure (CEI) domain, since sensor faults cause irreparable damage and lead to incorrect results on the condition monitoring of Cyber-Physical (CP) systems. This paper focuses on the challenging application of wind turbine (WT) monitoring. Specifically, we propose the two challenging architectures based on learning deep features, namely—Long Short Term Memory-Stacked Autoencoders (LSTM-SAE), and Convolutional Neural Network (CNN-SAE), for semi-supervised fault detection in wind CPs. The internal learnt features will facilitate the classification task by assigning each upcoming measurement into its corresponding faulty/normal operation status. To illustrate the quality of our schemes, their performance is evaluated against real-world’s wind turbine data. From the experimental section we are able to validate that both LSTM-SAE and CNN-SAE schemes provide high classification scores, indicating the high detection rate of the fault level of the wind turbines. Additionally, slight modification on our architectures are able to be applied on different fault/anomaly detection categories on variant Cyber-Physical systems.

This paper addresses the challenge of olive tree segmentation using drone imagery, which is crucial for precision agriculture applications. We tackle the data scarcity issue by augmenting existing detection datasets. Additionally, lightweight model variations of state-of-the-art models like YOLOv8n, RepViT-SAM, and EdgeSAM are combined into two proposed pipelines to meet computational constraints while maintaining segmentation accuracy. Our multifaceted approach successfully achieves an equilibrium among model size, inference time, and accuracy, thereby facilitating efficient olive tree segmentation in precision agriculture scenarios with constrained datasets. Following comprehensive evaluations, YOLOv8n appears to surpass the other models in terms of inference time and accuracy, albeit necessitating a more intricate fine-tuning procedure. Conversely, SAM-based pipelines provide a significantly more streamlined fine-tuning process, compatible with existing detection datasets for olive trees. However, this convenience incurs the disadvantages of a more elaborate inference architecture that relies on dual models, consequently yielding lower performance metrics and prolonged inference durations.