Abstract
Missing data is an issue that can negatively impact any task performed with the available data and it is often found in real -world domains such as healthcare. One of the most common strategies to address this issue is to perform imputation, where the missing values are replaced by estimates. Several approaches based on statistics and machine learning techniques have been proposed for this purpose, including deep learning architectures such as generative adversarial networks and autoencoders. In this work, we propose a novel siamese neural network suitable for missing data imputation, which we call Siamese Autoencoder-based Approach for Imputation (SAEI). Besides having a deep autoencoder architecture, SAEI also has a custom loss function and triplet mining strategy that are tailored for the missing data issue. The proposed SAEI approach is compared to seven state-of-the-art imputation methods in an experimental setup that comprises 14 heterogeneous datasets of the healthcare domain injected with Missing Not At Random values at a rate between 10% and 60%. The results show that SAEI significantly outperforms all the remaining imputation methods for all experimented settings, achieving an average improvement of 35%. This work is an extension of the article Siamese Autoencoder-Based Approach for Missing Data Imputation [1] presented at the International Conference on Computational Science 2023. It includes new experiments focused on runtime, generalization capabilities, and the impact of the imputation in classification tasks, where the results show that SAEI is the imputation method that induces the best classification results, improving the F1 scores for 50% of the used datasets.

Abstract
In fields that rely on high-stakes decisions, such as medicine, interpretability plays a key role in promoting trust and facilitating the adoption of deep learning models by the clinical communities. In the medical image analysis domain, gradient-based class activation maps are the most widely used explanation methods and the field lacks a more in depth investigation into inherently interpretable models that focus on integrating knowledge that ensures the model is learning the correct rules. A new approach, B-cos networks, for increasing the interpretability of deep neural networks by inducing weight-input alignment during training showed promising results on natural image classification. In this work, we study the suitability of these B-cos networks to the medical domain by testing them on different use cases (skin lesions, diabetic retinopathy, cervical cytology, and chest X-rays) and conducting a thorough evaluation of several explanation quality assessment metrics. We find that, just like in natural image classification, B-cos explanations yield more localised maps, but it is not clear that they are better than other methods' explanations when considering more explanation properties.

Abstract
Miniaturized three-electrode electrochemical sensors (MES) are widely used in the advancement of innovative technologies for remote sensing applications. MESs consist of conductive electrodes that are applied onto an inert solid substrate using various techniques, such as photolithography, electroplating, and screen printing. Typical MES systems comprise working (WE) and counter (CE) electrodes based on gold (Au), paired with a reference electrode (RE) based on silver (Ag). This configuration is commonly selected due to Au's high conductivity, low resistance, and compatibility with robust organothiol chemistries, especially for the WE. Moreover, Ag is often preferred for REs owing to its low toxicity, stability, and high conductivity. Nevertheless, in uncontrolled environments outside of cleanrooms, both Au and Ag surfaces are prone to atmospheric contamination, resulting in significant sensor variability and compromised analytical performance. Therefore, it is crucial to integrate a pre-processing stage into the sensor manufacturing process to guarantee the quality and cleanliness of MES electrode surfaces for sensor functionalization and precise electrochemical measurements. Considering the potential negative effects of methods tailored for a specific electrode material on another material, this study extensively investigates 18 different treatment methods for MESs incorporating Au CEs and WEs, along with Ag REs. Employing a multi-parametric analysis, this study aims to identify the most effective treatment for a variety of electrode materials, thereby improving analytical accuracy and reproducibility for subsequent MES (bio)sensor applications. Miniaturized three-electrode electrochemical sensors (MES) are essential for advancing remote sensing technologies. However, the inherent morpho-chemical heterogeneity of built-in electrodes challenges MES analytical performance. This study investigates treatments for the different electrode materials, providing new methods to enhance quality control, analytical accuracy, and reproducibility in MES biosensing applications. image

Abstract
Deep learning has rapidly increased in popularity, leading to the development of perception solutions for autonomous driving. The latter field leverages techniques developed for computer vision in other domains for accomplishing perception tasks such as object detection. However, the black-box nature of deep neural models and the complexity of the autonomous driving context motivates the study of explainability in these models that perform perception tasks. Moreover, this work explores explainable AI techniques for the object detection task in the context of autonomous driving. An extensive and detailed comparison is carried out between gradient-based and perturbation-based methods (e.g., D-RISE). Moreover, several experimental setups are used with different backbone architectures and different datasets to observe the influence of these aspects in the explanations. All the techniques explored consist of saliency methods, making their interpretation and evaluation primarily visual. Nevertheless, numerical assessment methods are also used. Overall, D-RISE and guided backpropagation obtain more localized explanations. However, D-RISE highlights more meaningful regions, providing more human-understandable explanations. To the best of our knowledge, this is the first approach to obtaining explanations focusing on the regression of the bounding box coordinates.

Abstract
Total evacuation time constitutes an important factor in the safety of any building. It is thus essential to devise an emergency evacuation plan, which will enable the safe evacuation of all the occupants in the shortest possible time. The main objective of this article was to examine and improve the evacuation process of a 4-star hotel located in the city of Porto, Portugal. To this end, one looked into 6 scenarios, by means of PathFinder simulation software, so as to determine the shortest total evacuation time and identify possible bottlenecks and congestion. The simulation model developed was tested to analyze the evacuation of 429 people from the hotel, based on the availability of the 3 accessible exit doors (central exit, side exit, spa exit) and elevators. Strategy 4 presented the shortest total evacuation time, with 536.0 s. Two other strategies which showed very similar times were 5 and 6, with 537.0 s and 537.5 s, respectively.

Abstract
Experimental assessment of different missing data imputation methods often compute error rates between the original values and the estimated ones. This experimental setup relies on complete datasets that are injected with missing values. The injection process is straightforward for the Missing Completely At Random and Missing At Random mechanisms; however, the Missing Not At Random mechanism poses a major challenge, since the available artificial generation strategies are limited. Furthermore, the studies focused on this latter mechanism tend to disregard a comprehensive baseline of state-of-the-art imputation methods. In this work, both challenges are addressed: four new Missing Not At Random generation strategies are introduced and a benchmark study is conducted to compare six imputation methods in an experimental setup that covers 10 datasets and five missingness levels (10% to 80%). The overall findings are that, for most missing rates and datasets, the best imputation method to deal with Missing Not At Random values is the Multiple Imputation by Chained Equations, whereas for higher missingness rates autoencoders show promising results.