Abstract
The advancement of the use of Artificial Intelligence (AI) in the healthcare sector makes it possible to use computational intelligence applications to assist healthcare professionals in the diagnosis process, facilitating and optimizing early detection and allowing for a more accurate diagnosis (He et al., 2019). The application of machine learning methods, and, more recently, deep learning, has shown promising results (Barbosa et al., 2022; da Silva et al., 2021; De Oliveira et al., 2020; Espinola et al., 2021a, b; Gomes et al., 2021, 2023; Santana et al., 2018; Torcate et al., 2022). These approaches allow powerful tools to support diagnostic imaging and signs to be built, through the extraction of image features and the creation of a classification system, for example (Yu et al., 2018). There are several diseases known and classified by man, with different causes and prevalence. Therefore, contributing to the early detection of diseases defined as neglected was the initial motivation for this work.

Abstract
Most forecasting methods use recent past observations (lags) to model the future values of univariate time series. Selecting an adequate number of lags is important for training accurate forecasting models. Several approaches and heuristics have been devised to solve this task. However, there is no consensus about what the best approach is. Besides, lag selection procedures have been developed based on local models and classical forecasting techniques such as ARIMA. We bridge this gap in the literature by carrying out an extensive empirical analysis of different lag selection methods. We focus on deep learning methods trained in a global approach, i.e., on datasets comprising multiple univariate time series. Specifically, we use NHITS, a recently proposed architecture that has shown competitive forecasting performance. The experiments were carried out using three benchmark databases that contain a total of 2411 univariate time series. The results indicate that the lag size is a relevant parameter for accurate forecasts. In particular, excessively small or excessively large lag sizes have a considerable negative impact on forecasting performance. Cross-validation approaches show the best performance for lag selection, but this performance is comparable with simple heuristics.

Abstract
We present LAZYFS, a new fault injection tool that simplifies the debugging and reproduction of complex data durability bugs experienced by databases, key-value stores, and other data-centric systems in crashes. Our tool simulates persistence properties of POSIX file systems (e.g., operations ordering and atomicity) and enables users to inject lost and torn write faults with a precise and controlled approach. Further, it provides profiling information about the system's operations flow and persisted data, enabling users to better understand the root cause of errors. We use LAZYFS to study seven important systems: PostgreSQL, etcd, Zookeeper, Redis, LevelDB, PebblesDB, and Lightning Network. Our fault injection campaign shows that LAZYFS automates and facilitates the reproduction of five known bug reports containing manual and complex reproducibility steps. Further, it aids in understanding and reproducing seven ambiguous bugs reported by users. Finally, LAZYFS is used to find eight new bugs, which lead to data loss, corruption, and unavailability.

Abstract
Cyberattacks exploit deceptions involving the Domain Name Service (DNS) to direct users to fake websites, such as typosquatting attacks, which exploit natural typographical errors, and homograph attacks, where different Unicode characters resemble the legitimate ones. The deception attacks may also exploit the confusion between DNS domain names, specifically Top-Level Domains (TLDs), and file extensions. Recently, two new TLDs were added, zip and mov, sharing names with certain file types. This overlapping can be explored by malicious actors in a range of threat scenarios to compromise user security. This paper provides an overview of threats originating from the confusion between specific TLDs and file extensions, such as the recent zip and mov. The threats are grouped into 6 threat scenarios that are described and discussed. This research can be part of a more comprehensive strategy that includes addressing the risks associated with these threats and designing future strategies to address the threats associated with exploiting this ambiguity.

Abstract
In recent years, the energy landscape has undergone significant transformations, characterized by the integration of renewable energy sources, smart grids, and the proliferation of IoT-enabled devices. As a result, the efficient management of energy resources has become paramount, requiring advanced methodologies in load forecasting and clustering. This article presents an enhanced methodology for short-term load forecasting that focuses on load consumption profile recognition within a smart building environment. The methodology is designed to analyze and identify recurring load consumption profiles and measures of sensors, thereby enhancing load consumption profile recognition capabilities within the smart building context. The interaction between single and grouped datasets is explored to enhance the accuracy and interpretability of predictions, contributing to optimized energy consumption and providing valuable information for demand response programs. The default forecasting methods used in the methodology are artificial neural networks and k-nearest neighbors. For comparing results and evaluating the proposed approach, XGBoost is also employed. The dataset is selected from a specific database, and the clustering method, partitioning type, is applied with k-means. The results, validated with error evaluation models and statistics, reveal the advantages of the proposed approach, especially with three clusters, where the results achieved by the Artificial Neural Network are the best. The clustering process, particularly the partitioning type, demonstrates a strong capability in improving load forecasting in smart buildings and helps understand load consumption patterns and achieve energy savings. Copyright (C) 2024 The Authors. This is an open access article under the CC BY-NC-ND license (https://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract
The literature on the isothermal model gas flow is extensive, but the effect of temperature variation on the hydraulic characteristics has been rarely addressed. Additionally, the impact of hydrogen blending on the thermal condition of NG pipelines is also an emergent topic that requires new approaches to the gas flow problem formulation and resolution. In this paper, a model for the gas flow problem was developed to optimise the operation of natural gas distribution networks with hydrogen injection while maintaining pressure, gas flows, and gas quality indexes within admissible limits. The goal is to maximise the injection of hydrogen and investigate the influences of thermal variations in the gas blending. Also, this model enables the calculation of the maximum permitted volume of hydrogen in the network, quantifying the total savings in natural gas usage and carbon dioxide emissions in different temperature conditions.