Abstract
This article focuses on the creation and evaluation of synthetic data to address the challenges of imbalanced datasets in machine learning (ML) applications, using fake news detection as a case study. We conducted a thorough literature review on generative adversarial networks (GANs) for tabular data, synthetic data generation methods, and synthetic data quality assessment. By augmenting a public news dataset with synthetic data generated by different GAN architectures, we demonstrate the potential of synthetic data to improve ML models' performance in fake news detection. Our results show a significant improvement in classification performance, especially in the underrepresented class. We also modify and extend a data usage approach to evaluate the quality of synthetic data and investigate the relationship between synthetic data quality and data augmentation performance in classification tasks. We found a positive correlation between synthetic data quality and performance in the underrepresented class, highlighting the importance of high-quality synthetic data for effective data augmentation.

Abstract
SENTINEL-Orb is an ongoing research and development initiative focused on creating a compact, fully autonomous flying robotic sphere. This sphere has been meticulously engineered to provide a comprehensive range of assistance to astronauts during both Extravehicular (EVA) and Intravehicular Activities (IVA). The system's design is intended to ensure operational effectiveness in microgravity and vacuum environments, thereby facilitating inspection, mapping, and navigation capabilities in areas that are either inaccessible or hazardous for human access. The development of the UX1-NEO system for flooded mine exploration was predicated on the advancement of underwater robotics technology pioneered at the Center for Robotics and Autonomous Systems (CRAS) of the Institute for Systems and Computer Engineering, Technology and Science (INESC TEC). This technology enabled the miniaturization, modular design, and integration of advanced perception and control algorithms for the SENTINEL-Orb. The current prototype incorporates a safe propulsion system designed to ensure the well-being of astronauts, a translation and attitude control mechanism, and a high-resolution camera that enhances the visual-inertial odometry module for object detection and anomaly identification. The propulsion and control subsystems were tested independently, while the six degrees of freedom (6-DoF) control performance has been validated through high-fidelity simulations, demonstrating precise maneuvering capabilities. The platform's modular architecture facilitates straightforward assembly and accommodates future enhancements. Preparations are underway for a parabolic flight campaign to evaluate the complete system's performance in a real microgravity environment. Preliminary findings suggest that the design meets rigorous safety standards and has significant potential for operational autonomy in space missions. A comparison of SENTINEL-Orb to the existing robotic solutions reveals that it addresses limitations related to atmospheric dependency and astronaut safety. This establishes SENTINEL-Orb as a promising tool for space maintenance, inspection, and exploration activities. © 2025 by INESC-TEC.

Abstract

Abstract

Abstract
In the context of Strategic Governance in Public Administration (SGPA) aimed at promoting Digital Transformation and Innovation, and improving the quality of services made available to citizens, this work proposes the design and specification of an open-source, citizen-centered application architecture to support SGPA, leveraging Generative Artificial Intelligence (AI) approaches. Based on a literature review and highlighting the lack of fully open-source virtual assistant (chatbot) solutions for SGPA, this work, in addition to the review, presents a prototype that adopts a hybrid AI approach for the development of chatbots tailored to SGPA purposes. It combines rule-based methods for quick responses with Large Language Models and Retrieval-Augmented Generation to improve accuracy in responding to complex queries. To encourage future research in this field, the code and pre-trained models are available in a public GitHub repository.

Abstract