Abstract

Abstract
Deep-learning techniques can improve the efficiency of medical diagnosis while challenging human experts’ accuracy. However, the rationale behind these classifier’s decisions is largely opaque, which is dangerous in sensitive applications such as healthcare. Case-based explanations explain the decision process behind these mechanisms by exemplifying similar cases using previous studies from other patients. Yet, these may contain personally identifiable information, which makes them impossible to share without violating patients’ privacy rights. Previous works have used GANs to generate anonymous case-based explanations, which had limited visual quality. We solve this issue by employing a latent diffusion model in a three-step procedure: generating a catalogue of synthetic images, removing the images that closely resemble existing patients, and using this anonymous catalogue during an explanation retrieval process. We evaluate the proposed method on the MIMIC-CXR-JPG dataset and achieve explanations that simultaneously have high visual quality, are anonymous, and retain their explanatory value.

Abstract
Explainable Artificial Intelligence (XAI) aims to enhance the interpretability of Artificial Intelligence (AI) systems for humans. The goal is to ensure that algorithmic decisions and underlying data are understandable to non-technical stakeholders. Advanced Machine Learning (ML) models, such as deep neural networks, enable AI systems to process vast data and extract intricate patterns, akin to the human brain, but this complicates XAI development. Complex ML models require substantial data for training, exacerbating the challenge. Consequently, this paper proposes a novel approach to improve XAI for complex ML models, particularly those with large data needs. Using K-Means clustering, the paper proposes to identify relevant data instances to create similarity clusters. This filtering process focuses XAI on essential information, even with complex models, reducing the data set to find patterns and explanations, so that, using the same approach, only the best explanations are filtered efficiently. The paper proposes to implement and test this model with a case study on ML for PV generation forecasting in buildings. Results show that the proposed approach is able to generate explanations that are very similar to those generated when using the entire available data, in only a portion of the execution time, leveraging from the identification of a small number of representative data points. This approach, therefore, enhances the efficiency of XAI by achieving promising results with a smaller dataset. It also facilitates the development of more understandable and fastly provided solutions, which is essential for real-world XAI users such as electric mobility users that need PV forecasting explanations as support for their vehicles charging management.

Abstract
Quadruped robots have gained significant attention in the robotics world due to their capability to traverse unstructured terrains, making them advantageous in search and rescue and surveillance operations. However, their utility is substantially restricted in situations where object manipulation is necessary. A potential solution is to integrate a robotic arm, although this can be challenging since the arm's addition may unbalance the whole system, affecting the quadruped locomotion. To address this issue, the robotic arm must be adapted to the quadruped robot, which is not viable with commercially available products. This paper details the design and development of a robotic arm that has been specifically built to integrate with a quadruped robot to use in a variety of agricultural and industrial applications. The design of the arm, including its physical model and kinematic configuration, is presented. To assess the effectiveness of the prototype, a simulation was conducted with a motion-planning algorithm based on the arm's inverse kinematics. The simulation results confirm the system's stability and the functionality of the robotic arm's movement.

Abstract
The satellite monitoring of vegetation moisture content (VMC) and soil moisture content (SMC) in Southern European Atlantic mountains remains poorly understood but is a fundamental tool to better manage landscape moisture dynamics under climate change. In the Atlantic humid mountains of Portugal, we investigated an empirical model incorporating satellite (Sentinel-1 radar, S1; Sentinel-2 optical, S2) and ancillary predictors (topography and vegetation cover type) to monitor VMC (%) and SMC (%). Predictors derived from the S1 (VV, HH and VV/HH) and S2 (NDVI and NDMI) are compared to field measurements of VMC (n = 48) and SMC (n = 48) obtained during the early, mid and end of summer. Linear regression modelling was applied to uncover the feasibility of a landscape model for VMC and SMC, the role of vegetation type models (i.e. native forest, grasslands and shrubland) to enhance predictive capacity and the seasonal variation in the relationships between satellite predictors and VMC and SMC. Results revealed a significant but weak relationship between VMC and predictors at landscape level (R2 = 0.30, RMSEcv = 69.9 %) with S2_NDMI and vegetation cover type being the only significant predictors. The relationship improves in vegetation type models for grasslands (R2 = 0.35, RMSEcv = 95.0 % with S2_NDVI) and shrublands conditions (R2 = 0.52, RMSEcv = 45.3 %). A model incorporating S2_NDVI and S1_VV explained 52 % of the variation in VMC in shrublands. The relationship between SMC and satellite predictors at the landscape level was also weak, with only the S2_NDMI and vegetation cover type exhibiting a significant relationship (R2 = 0.28, RMSEcv = 18.9 %). Vegetation type models found significant associations with SMC only in shrublands (R2 = 0.31, RMSEcv = 9.03 %) based on the S2_NDMI and S1_VV/VH ratio. The seasonal analysis revealed however that predictors associated to VMC and SMC may vary over the summer. The relationships with VMC were stronger in the early summer (R2 = 0.31, RMSEcv = 90.1 %; based on S2_NDMI) and mid (R2 = 0.37, RMSEcv = 70.8 %; based on S2_NDVI), butnon-significant in the end of summer. Similar pattern was found for SMC, where the link with predictors decreases from the early summer (R2 = 0.33, RMSEcv = 16.0 %; based on S1_VH) and mid summer (R2 = 0.30, RMSEcv = 17.8 %; based on S2_NDMI) to the end of summer (non-significant). Overall, the hypothesis of a universal landscape model for VMC and SMC was not fully supported. Vegetation type models showed promise, particularly for VMC in shrubland conditions. Sentinel optical and radar data were the most significant predictors in all models, despite the inclusion of ancillary predictors. S2_NDVI, S2_NDMI, S1_VV and S1_VV/VH ratio were the most relevant predictors for VMC and, to a lesser extent, SMC. Future research should quantify misregistration effects using plot vs. moving window values for the satellite predictors, consider meteorological control factors, and enhance sampling to overcome a main limitation of our study, small sample size.

Abstract
Pupil size can be used as an important biomarker for occupational risks. In recent years, there has been an increase in the development of open-source tools dedicated to obtaining and measuring pupil diameter. However, it remains undetermined determined whether these tools are suitable for use in occupational settings. This study explores the significance of pupil size variation as a biomarker for occupational risks and evaluates existing opensource methods for potential use in both research and occupational settings, with the goal of to prevent occupational accidents and improve the health and performance of workers. To this end, a two-phase systematic literature review was conducted in the Web of Science TM, ScienceDirect (R), and Scopus (R) databases. For the relevance of monitoring pupil size variation in occupational settings, 15 articles were included. The articles were divided into three groups: mental workload, occupational stress, and mental fatigue. In most cases, pupil dilation increased with workload enhancement and with higher levels of stress. Regarding fatigue, it was noted that an increase in this condition corresponded with a decrease in pupil size. With respect to the open-source methodologies, 16 articles were identified, which were categorized into two groups: algorithms and software. Convolutional neural networks (CNN) 1 have exhibited superior performance among the various algorithmic approaches studied. Building on this insight, and considering the evaluations of software options, MEYE emerges as the premier open-source system for deployment in occupational settings due to its compatibility with a standard computer webcam. This feature positions MEYE as a particularly practical tool for workers in stable environments, like those of developers and administrators.