Abstract
Traumatic Brain Injury (TBI) is a form of brain injury caused by external forces, resulting in temporary or permanent impairment of brain function. Despite advancements in healthcare, TBI mortality rates can reach 30%–40% in severe cases. This study aims to assist clinical decision-making and enhance patient care for TBI-related complications by employing Artificial Intelligence (AI) methods and data-driven approaches to predict decompensation. This study uses learning models based on sequential data from Electronic Health Records (EHR). Decompensation prediction was performed based on 24-h in-mortality prediction at each hour of the patient's stay in the Intensive Care Unit (ICU). A cohort of 2261 TBI patients was selected from the MIMIC-III dataset based on age and ICD-9 disease codes. Logistic Regressor (LR), Long-short term memory (LSTM), and Transformers architectures were used. Two sets of features were also explored combined with missing data strategies by imputing the normal value, data imbalance techniques with class weights, and oversampling. The best performance results were obtained using LSTMs with the original features with no unbalancing techniques and with the added features and class weight technique, with AUROC scores of 0.918 and 0.929, respectively. For this study, using EHR time series data with LSTM proved viable in predicting patient decompensation, providing a helpful indicator of the need for clinical interventions. © 2025 Elsevier Ltd

Abstract
Two essential tasks in production planning and control are the generation and the release of orders to the shop floor. In this study order, generation is based on the Demand Driven Materials Requirement Planning system, while order release is based on the CONstant Work-in-Process system. Although the two systems alone have been extensively studied, their combination has received much less attention. In this paper, we address the problem of sequencing replenishment orders generated by the Demand Driven Materials Requirement Planning system to be released by the CONstant Work-in-Process system. Four pool-sequencing rules have been considered. Two of these are used by Demand Driven Materials Requirement Planning for establishing priorities for order planning and order execution. The other two are the First-Come-First-Served rule and a virtual due date rule. Results of a simulation study show that the rules proposed in the Demand Driven Materials Requirement Planning literature for planning and for execution are not the best options for pool-sequencing, particularly for restricted levels of workload allowed on the shop floor. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

Abstract
Bees play a critical role as pollinators in ecosystem services, contributing significantly to the sexual reproduction and diversity of plants. The Caatinga biome in Brazil, home to around 200 bee species, provides an ideal habitat for these species due to its unique climate conditions. However, this biome faces threats from anthropogenic processes, making it urgent to characterise the local bee populations efficiently. Traditional taxonomic surveys for bee identification are complex due to the lack of suitable keys and expertise required. As a result, molecular barcoding has emerged as a valuable tool, using genome regions to compare and identify bee species. However, little is known about Caatinga bees to develop these molecular tools further. This study addresses this gap, providing an updated list of 262 Caatinga bee species across 86 genera and identifying similar to 40 primer sets to aid in barcoding these species. The findings highlight the ongoing work needed to fully characterise the Caatinga biome's bee distribution and species or subspecies to support more effective monitoring and conservation efforts.

Abstract
The textile industry faces economic and environmental challenges due to low recycling rates and contamination from fasteners like buttons, rivets, and zippers. This paper proposes an Red, Green, Blue (RGB) vision system using You Only Look Once version 11 (YOLOv11) with a sliding window technique for automated fastener detection. The system addresses small object detection, occlusion, and fabric variability, incorporating Grounding DINO for garment localization and U2-Net for segmentation. Experiments show the sliding window method outperforms full-image detection for buttons and rivets (precision 0.874, recall 0.923), while zipper detection is less effective due to dataset limitations. This work advances scalable AI-driven solutions for textile recycling, supporting circular economy goals. Future work will target hidden fasteners, dataset expansion and fastener removal.

Abstract
Supporting sustainability through modelling and analysis has become an active area of research in Software Engineering. Therefore, it is important and timely to survey the current state of the art in sustainability in Cyber-Physical Systems (CPS), one of the most rapidly evolving classes of complex software systems. This work presents the findings of a Systematic Mapping Study (SMS) that aims to identify key primary studies reporting on CPS modelling approaches that address sustainability over the last 10 years. Our literature search retrieved 2209 papers, of which 104 primary studies were deemed relevant fora detailed characterisation. These studies were analysed based on nine research questions designed to extract information on sustainability attributes, methods, models/meta-models, metrics, processes, and tools used to improve the sustainability of CPS. These questions also aimed to gather data on domain-specific modelling approaches and relevant application domains. The final results report findings for each of our questions, highlight interesting correlations among them, and identify literature gaps worth investigating in the near future.

Abstract
We show how (well-established) type systems based on non-idempotent intersection types can be extended to characterize termination properties of functional programming languages with pattern matching features. To model such programming languages, we use a (weak and closed) lambda-calculus integrating a pattern matching mechanism on algebraic data types (ADTs). Remarkably, we also show that this language not only encodes Plotkin's CBV and CBN lambda-calculus as well as other subsuming frameworks, such as the bang-calculus, but can also be used to interpret the semantics of effectful languages with exceptions. After a thorough study of the untyped language, we introduce a type system based on intersection types, and we show through purely logical methods that the set of terminating terms of the language corresponds exactly to that of well-typed terms. Moreover, by considering non-idempotent intersection types, this characterization turns out to be quantitative, i.e. the size of the type derivation of a term t gives an upper bound for the number of evaluation steps from t to its normal form.