Abstract
As machine learning technologies are increasingly adopted, the demand for responsible AI practices to ensure transparency and accountability grows. To better understand the decision-making processes of machine learning models, GASTeN was developed to generate realistic yet ambiguous synthetic data near a classifier's decision boundary. However, the results were inconsistent, with few images in the low-confidence region and noise. Therefore, we propose a new GASTeN version with a modified architecture and a novel loss function. This new loss function incorporates a multi-objective measure with a Gaussian loss centered on the classifier probability, targeting the decision boundary. Our study found that while the original GASTeN architecture yields the highest Frechet Inception Distance (FID) scores, the updated version achieves lower Average Confusion Distance (ACD) values and consistent performance across low-confidence regions. Both architectures produce realistic and ambiguous images, but the updated one is more reliable, with no instances of GAN mode collapse. Additionally, the introduction of the Gaussian loss enhanced this architecture by allowing for adjustable tolerance in image generation around the decision boundary.

Abstract
The development of a robust annotation scheme and corresponding guidelines is crucial for pro- ducing annotated datasets that advance both lin- guistic and computational research. This paper presents a case study that outlines a method- ology for designing an annotation scheme and its guidelines, specifically aimed at represent- ing morphosyntactic and semantic information regarding temporal features, as well as medi- cal information in medical reports written in Portuguese. We detail a multi-step process that includes reviewing existing frameworks, con- ducting an annotation experiment to determine the optimal approach, and designing a model based on these findings. We validated the ap- proach through a pilot experiment where we assessed the reliability and applicability of the annotation scheme and guidelines. In this ex- periment, two annotators independently anno- tated a patient's medical report consisting of six documents using the proposed model, while a curator established the ground truth. The analy- sis of inter-annotator agreement and the annota- tion results enabled the identification of sources of human variation and provided insights for further refinement of the annotation scheme and guidelines.

Abstract
Pedestrian safety is a pressing subject in urban areas. The disorderly sharing of streets and roads between pedestrians and vehicles leads to potentially serious accidents for pedestrians. This student project aims to tackle the issue by placing an interactive gaming device at traffic lights. SMASHY by Stempe Safety offers pedestrians an amusing and active way to discourage jaywalking. The multipurpose solution features a smashing game with buttons on one side and a screen displaying useful information on the other side. While the traffic light remains red for pedestrians, the module buttons light up and the players can start smashing the buttons as fast as possible, until the light turns green and consequently, the game ends. Ultimately, the modules are connected to an app where, if desired by the player, scores can be tracked and difficulty can vary based on user performance. Multiple modules can be placed around the city and the app will track player scores by location. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

Abstract
Most human time is spent indoors, and due to the pandemic, monitoring indoor air quality (IAQ) has become more crucial. In this study, an IoT (Internet of Things) architecture is implemented to monitor IAQ parameters, including CO2 and particulate matter (PM). An ESP32-C6-based device is developed to measure sensor data and send them, using the MQTT protocol, to a remote InfluxDBv2 database instance, where the data are stored and visualized. The Python 3.11 scripting programming language is used to automate Flux queries to the database, allowing a more in-depth data interpretation. The implemented system allows to analyze two measured scenarios during sleep: one with the door slightly open and one with the door closed. Results indicate that sleeping with the door slightly open causes CO2 levels to ascend slowly and maintain lower concentrations compared to sleeping with the door closed, where CO2 levels ascend faster and the maximum recommended values are exceeded. This demonstrates the benefits of ventilation in maintaining IAQ. The developed system can be used for sensing in different environments, such as schools or offices, so an IAQ assessment can be made. Based on the generated data, predictive models can be designed to support decisions on intelligent natural ventilation systems, achieving an optimized, efficient, and ubiquitous solution to moderate the IAQ.

Abstract

Abstract
In recent years, code-reuse attacks have been used to exploit software vulnerabilities and gain control of numerous software programs and embedded devices. Several measures have been put in place to prevent this type of attack, such as Control-Flow Integrity (CFI) systems, and some of these systems have already been integrated into hardware. Nevertheless, Function-Oriented Programming (FOP) attacks, a form of code-reuse that chains functions to carry out malicious actions, continue to persist. In this work, we present the first analysis of the implications and feasibility of FOP attacks on microcontrollers, focusing on ARM Cortex-M processors that support PACBTI, that is, a hardware feature designed for CFI system implementation. During this process, we identified multiple dispatch gadgets in two common Real-time Operating System (RTOS). Since these gadgets reside within core OS functionalities, they are inherently included in a broad range of embedded operating systems. Furthermore, we also present CortexMFopper - a tool specially built to identify FOP gadgets in embedded devices and to raise awareness of this technique.