Abstract
Speech recognition aims to convert human speech into text and has applications in security, healthcare, commerce, automobiles, and technology, just to name a few. Inserting residual neural networks before recurrent neural network cells improves accuracy and cuts training time by a good margin. Furthermore, layer normalization instead of batch normalization is more effective in model training and performance enhancement. Also, the size of the datasets presents tremendous influences in achieving the best performance. Leveraging these tricks, this article proposes an automatic speech recognition model with a stacked five layers of customized Residual Convolution Neural Network and seven layers of Bi-Directional Gated Recurrent Units, including a logarithmic so f tmax for the model output. Each of them incorporates a learnable per-element affine parameter-based layer normalization technique. The training and testing of the new model were conducted on the LibriSpeech corpus and LJ Speech dataset. The experimental results demonstrate a character error rate (CER) of 4.7 and 3.61% on the two datasets, respectively, with only 33 million parameters without the requirement of any external language model.

Abstract
[No abstract available]

Abstract
Nowadays, cities face severe problems related to traffic management and mobility in general. Therefore, technologies have been developed that can handle these situations and somehow mitigate the caused impact, such as CCTV cameras. However, the techniques for analyzing the images collected by these cameras are increasingly complex and have numerous applications, being dispersed in the literature. Therefore, this article fills an important research gap by presenting a systematic review of the literature on the possible applications of data collected from CCTV cameras and the image analysis and processing techniques that have been developed and proposed in recent years. This systematic review followed the PRISMA statement guidelines and checklist, and three databases were searched, namely Scopus, Web of Science, and Inspec. From the analysis performed, the following applications were identified: Image/video analysis and traffic estimation, pedestrian detection, traffic data analysis, and forecasting, and traffic management. Regarding the image analysis and processing techniques YOLO (only look once), GMM (Gaussian mixture method), morphological methods, fuzzy logic, and other proprietary methods stand out. After a thorough analysis of traffic data, most works still implemented relatively trivial traffic management systems to generate a series of actions to be eventually applied to traffic controllers. Additionally, it was realized that these techniques could be implemented in industrial products from a future perspective. © 2023, Aracne Editrice. All rights reserved.

Abstract
Trade promotions are complex marketing agreements between a retailer and a manufacturer aiming to drive up sales. The retailer proposes numerous sales promotions that the manufacturer partially supports through discounts and deductions. In the Portuguese consumer packaged goods (CPG) sector, the proportion of price-promoted sales to regular-priced sales has increased significantly, making proper promotional planning crucial in ensuring manufacturer margins. In this context, a decision support system was developed to aid in the promotional planning process of two key product categories of a Portuguese CPG manufacturer. This system allows the manufacturer’s commercial team to plan and simulate promotional scenarios to better evaluate a proposed trade promotion and negotiate its terms. The simulation is powered by multiple gradient boosting machine models that estimate sales for a given promotion based solely on the scarce data available to the manufacturer. © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

Abstract
In a world facing unprecedented challenges, such as climate changes and growing social problems, the pharmaceutical industry must ensure that its supply chains are environmentally sustainable and resilient, guaranteeing access to key medications even when faced with unanticipated disruptions or crises. The core goal of this work is to develop an innovative simulation-based approach to support more informed and effective decision making, while establishing reasonable trade-offs between supply chain robustness and resiliency, operational efficiency, and environmental and social concerns. Such a decision-support system will contribute to the development of more resilient and sustainable pharmaceutical supply chains, which are, in general, critical for maintaining access to essential medicines, especially during times of crises or relevant disruptions. The system will help companies to better manage and design their supply chains, providing a valuable tool to achieve higher levels of resilience and sustainability. The study we conducted has two primary contributions that are noteworthy. Firstly, we present a new advanced approach that integrates multiple simulation techniques, allowing for the modeling of highly complex environments. Secondly, we introduce a new conceptual framework that helps to comprehend the interplay between resiliency and sustainability in decision-making processes. These two contributions provide valuable insights into understanding complex systems and can aid in designing more resilient and sustainable systems.

Abstract
Machine learning algorithms offer the capability to analyze large volumes of real-time data, providing transport authorities with valuable insights into traffic conditions, congestion hotspots, and incident detection from diverse data sources. However, these algorithms face challenges related to data quality and reliability. We conducted a comparative analysis of machine-learning models that can be used to identify and filter transportation content from social media or other sources that can provide small and concise text. The filtrated result can then feed models and/or tools used to improve and automate traffic control, operational management, and tactical management decision-making. We consider factors such as run time, generalization capacity, and performance metrics as criteria to assess their suitability for different decision levels. The analysis is supported by a dataset consisting of Twitter content. The predictions from three groups of algorithms are evaluated: traditional machine learning algorithms (Support Vector Machines, Logistic Regression, and Random Forest), a fine-tuned Google BERT model, and Google BERT models without training (BERT-base and BERT-large). The tests are performed using New York, London, and Melbourne data. The findings of this research aim to assist decision-makers in making informed choices when selecting the most appropriate method to filtrate information subsequently used for models that contribute to different traffic management tasks.