Adversarial machine learning (AML) is a class of data manipulation techniques that cause alterations in the behavior of artificial intelligence (AI) systems while going unnoticed by humans. These alterations can cause serious vulnerabilities to mission-critical AI-enabled applications. This work introduces an AI architecture augmented with adversarial examples and defense algorithms to safeguard, secure, and make more reliable AI systems. This can be conducted by robustifying deep neural network (DNN) classifiers and explicitly focusing on the specific case of convolutional neural networks (CNNs) used in non-trivial manufacturing environments prone to noise, vibrations, and errors when capturing and transferring data. The proposed architecture enables the imitation of the interplay between the attacker and a defender based on the deployment and cross-evaluation of adversarial and defense strategies. The AI architecture enables (i) the creation and usage of adversarial examples in the training process, which robustify the accuracy of CNNs, (ii) the evaluation of defense algorithms to recover the classifiers’ accuracy, and (iii) the provision of a multiclass discriminator to distinguish and report on non-attacked and attacked data. The experimental results show promising results in a hybrid solution combining the defense algorithms and the multiclass discriminator in an effort to revitalize the attacked base models and robustify the DNN classifiers. The proposed architecture is ratified in the context of a real manufacturing environment utilizing datasets stemming from the actual production lines.

In the face of increasingly sophisticated and dynamic cyber threats, policy-based cybersecurity deployment has emerged as a critical strategy for securing modern, complex network infrastructures. This approach enables organizations to enforce adaptable, scalable and context-aware security policies that can dynamically respond to evolving threats and operational demands. However, the growing complexity of distributed systems and the interdependencies between diverse security components present significant challenges in ensuring consistent and efficient policy enforcement. Static and centralized security models are no longer sufficient to address these challenges, underscoring the need for intelligent, distributed solutions that can optimize policy deployment in real-time. This paper introduces a novel framework based on Message Passing algorithms that enables decentralized, adapted and optimized policy deployment across complex cybersecurity infrastructures. More specifically, we extend the traditional Min-Sum algorithm by incorporating dynamic trust weights that influence policy enforcement decisions, allowing more reliable nodes to have a greater impact while mitigating the risks posed by untrusted entities. The proposed framework is designed to handle large-scale network environments with minimal computational overhead, relying on localized message exchanges instead of centralized computations. We benchmark our framework against state-of-the-art approaches, demonstrating that the proposed framework achieves lower total enforcement cost while significantly improving convergence speed, particularly in heterogeneous trust environments. Finally, we provide an in-depth evaluation of how trust heterogeneity influences security policy enforcement. Extensive simulations were conducted to evaluate the framework’s effectiveness in optimizing policy deployment for varying network topologies, demonstrating significant improvements in threat detection accuracy, policy consistency and resource efficiency across diverse and dynamic network environments.

Data spaces consist of trusted frameworks that manage the entire data lifecycle, encompassing various data models, metadata descriptors, ontologies for semantic interpretation, and data services for accessing, processing, and analyzing data. Domain-specific data spaces are currently being designed and deployed in vertical sectors, following specifications and reference frameworks that enable interoperability and compatibility. One such vertical sector with high impact on economy is mobility. In this paper, we present our position towards mobility data spaces, highlighting the idiosyncrasy of mobility data, while presenting the MobiSpaces approach that encompasses services for mobility data management and advanced mobility analytics. By exploiting MobiSpaces, mobility data providers can break the barrier of participation in data spaces thereby sharing their dataset with less hurdle, whereas data consumers can find advanced data analysis tools readily applicable on mobility data sets to extract insights and discover mobility patterns.

This paper provides an analysis and comparison of the YOLOv5, YOLOv8 and YOLOv10 models for webpage CAPTCHAs detection using the datasets collected from the web and darknet as well as synthetized data of webpages. The study examines the nano (n), small (s), and medium (m) variants of YOLO architectures and use metrics such as Precision, Recall, F1 score, mAP@50 and inference speed to determine the real-life utility. Additionally, the possibility of tuning the trained model to detect new CAPTCHA patterns efficiently was examined as it is a crucial part of real-life applications. The image slicing method was proposed as a way to improve the metrics of detection on oversized input images which can be a common scenario in webpages analysis. Models in version nano achieved the best results in terms of speed, while more complexed architectures scored better in terms of other metrics.