This study proposes a glucose regulation approach that integrates PID control with Insulin Feedback (IFB) and Takagi-Sugeno fuzzy logic, applied to the Bergman minimal model in patients with type 1 diabetes, within the context of artificial pancreas systems. The hybrid control strategy combines the simplicity of the Bergman model with the adaptive features of fuzzy logic and the robust performance of PID-IFB control. Simulation results show significant improvements in glycemic control, with minimal overshoot and better prediction of hyperglycemia sequences, particularly in response to meal-related disturbances, compared to traditional PID control. The system effectively maintains glucose levels within the normoglycemic range (60-120 mg/dL), while avoiding both hyperglycemic and hypoglycemic episodes.
Accurate short-term electricity consumption forecasting is essential for optimizing energy management in Smart Campus Grids, particularly within a Smart City infrastructure. However, electricity consumption patterns exhibit variability due to fluctuating user behavior and external weather variables. To address these complexities, ensemble learning techniques have been explored. Stacking, in particular, leverages multiple base models through a meta-learner, offering an improved prediction accuracy for dynamic patterns.In this paper, we propose a stacking ensemble learning model to enhance short-term electricity consumption forecasting in a campus-scale Automated Metering infrastructure (AMI): a first step towards a fully integrated Smart Campus Grid. Our approach combines many base models including Extreme Gradient Boosting, Random Forest, Lasso Regression, and Support Vector Regression, with a Linear Regression model serving as the meta-learner. Our work uses on-site electricity consumption data collected from an AMI we already implemented consisting of 18 smart meters deployed at the National Institute of Posts and Telecommunications (INPT), a Moroccan engineering school with on-campus housing. Our ensemble model outperforms individual base models by an average of 8.7 %. This highlights the potential of advanced machine learning for accurate energy management in Smart Cities.
Viral evolution is a complex and natural process that determine how viruses transmit, evolve, and cause disease. Understanding how viruses evolve and mutate is important from a public health perspective to guide vaccine development, early diagnosis and outbreak response. Traditional approaches of studying viral evolution (i.e., phylogenetic analysis, epidemiologic modeling) can struggle to keep pace with the rapid and often unpredictable nature of viral evolution. Recently, machine learning (ML) has emerged as a promising means of analyzing the vast amount of data created through viral genomics that can help distinguish evolutionary patterns and predict mutations. This paper reviews some of the machine learning techniques being utilized in the study of viral evolution through applied machine learning, with a focus on machine learning model components and procedures (i.e., no virological basis). We review model types used in the analysis of viral prediction, including deep learning architectures, such as, convolutional neural networks (CNNs), long short term memory (LSTM) networks, transformers and traditional modeling approaches in machine learning such as support vector machines (SVMs) and hybrid approaches. In addition to models study, we also review data sources (e.g GISAID, GenBank, NCBI) and how data was implemented to feed the model. Next we will present evaluation metrics used to validate and evaluate models. Finally, we discuss future issues and directions for machine learning and viral evolution prediction methods including data size, model generalizability, computational issues and model interpretability. Overall, these advancements with respect to ML and viral evolution research have made important inroads to the modeling of viral evolution prediction. However, issues such as stability, model types, interpretability, and real world applications remain.
Extracting semantic relations between words is crucial for the development and enrichment of lexical resources, especially for under-resourced languages like Moroccan Darija. This paper presents an automated methodology for identifying synonyms, antonyms, hypernyms, and hyponyms by leveraging bilingual Darija–English resources, Princeton WordNet (PWN), the Suggested Upper Merged Ontology (SUMO), and the NLTK toolkit. Experimental evaluation was conducted on a dataset of 361 Darija nouns, selected as a preliminary testbed to validate the methodology before scaling it to the full lexicon. The results show that 83.10% were successfully aligned with PWN synsets, resulting in the extraction of 14,201 semantic relations, of which 5,475 (38.55%) were validated through back-translation. These findings confirm the potential of transferring semantic knowledge from English into Darija, despite cultural and lexical mismatches. The proposed pipeline substantially enriches Darija’s lexical coverage and offers a scalable and replicable approach for developing semantic resources in other low-resource dialects.
Federated Learning (FL) has become a game-changing strategy. in machine learning, allowing for decentralized model training while safeguarding data privacy. In the healthcare industry, where private patient information is dispersed around numerous organizations and gadgets, this approach is essential. FL facilitates the development of reliable predictive models without compromising compliance, as it retains data locally. In this research, we calculate the performance of FL on two distinct datasets: a Categorical, Integer dataset, the "Diabetes 130-US Hospitals," and an image dataset, "Diabetic Retinopathy 224x224". Using FedAP personalized federated learning strategy designed for the medical field. The results demonstrate FL FedAP’s effectiveness in handling diverse data types. These metrics underscore FL’s capacity to adapt and perform across heterogeneous healthcare data. By enhancing predictive accuracy and fostering patient-centric innovations, FL sets a new standard for privacy-preserving machine learning, paving the way for a future where healthcare solutions are both effective and secure.
Stream ciphers play energetic mechanisms in securing cryptographic protocols by providing fast, efficient, and robust symmetric encryption systems to protect sensitive data in weak environments like wireless network. In this paper, we propose a novel stream cipher algorithm founded on the mathematical foundations of quadratic fields, primitive polynomials, and Linear Feedback Shift Registers (LFSRs) to achieve internal encryptions secure against structural and probabilistic attacks. The quadratic fields provides us a rich, random and robust algebraic framework characterized by unique properties that enhance the generation of secure key stream, while primitive polynomials ensure maximal periodicity and strong randomness properties in the key stream generation. LFSRs are exploited for their computational efficiency and great simplicity of implementation, further improving the algorithm's ability to generate high-entropy pseudo-random binary sequences. The proposed cipher demonstrates strong resistance against cryptographic attacks, including linear and differential attacks, outstanding to its innovative design founded on robust synchronous stream cipher generator. Through its experimental evaluation, we validate the efficiency, security, and practical applicability of the proposed algorithm in modern communication and resource-constrained environments like wireless network, Electric Vehicles (EVs), Internet of Things networks, or embedded systems. This work underscores the potential of advanced algebraic techniques in the design of next-generation cryptographic systems.
Modernizing legacy systems in public administration is more than a technical task, it is an institutional transformation. This paper presents a novel architectural design approach tailored for national e-government platforms, based on microservices (MSA) principles and domain-driven design (DDD). We explore the case of Morocco’s civil registry System (E-FES), a national civil registration platform, in order to demonstrate how monolithic public service platforms can be restructured into scalable, modular, and legally compliant systems. Our framework decomposes domain logic into independently deployable services, each aligned with specific civil registration processes such as births, marriages, divorces, and deaths. This case-based implementation offers a replicable approach for governments seeking to modernize core civil services through scalable digital infrastructures.
The winter season in mountainous regions such as Ifrane, Morocco, is known for its extreme weather conditions. To maintain thermal comfort in residential households, people commonly use wood logs, also known as traditional biomass, for heating. However, this method contributes significantly to deforestation in the region while at the same time emits greenhouse gases into the atmosphere. This project proposes an alternative solution for heating by using pellet stoves. The wood pellets are manufactured from raw materials native to the Ifrane region. The study includes software simulations in which key performance indicators of the stove are assessed. The proposed solution is environmentally friendly and economically beneficial for low-income families.
This systematic literature review aims to analyze the state-of-the-art and scientometric trends in the application of the Technology Readiness Level (TRL) framework within supply chain management. Following PRISMA guidelines, a systematic search identified 44 relevant publications (published between 2013 and 2025), which were subjected to bibliometric and thematic analyses. The bibliometric analysis reveals increasing research activity in this area, highlighting key journals, authors, and prevalent keywords related to technology adoption and innovation. The thematic analysis uncovers core topics including technology adoption strategies and readiness assessment. Findings indicate that TRL frameworks play a pivotal role in optimizing supply chain performance by aligning technological maturity with operational deployment, thereby enhancing efficiency and flexibility. However, small and medium-sized enterprises (SMEs) face specific challenges in applying TRL, including limited resources, insufficient technical expertise, and organizational constraints, which can hinder their ability to fully leverage emerging technologies. This review synthesizes current knowledge on TRL in supply chain ecosystems and highlights gaps for future research, providing guidance for scholars and practitioners on effectively leveraging TRL to improve supply chain performance. Keywords : Technology Readiness Level, TRL, Supply Chain, Logistics, Manufacturing, Industrial Innovation
This survey examines recent advancements in 3D U-Net architectures for brain tumor segmentation, highlighting their vital role in accurate medical image analysis. It reviews significant modifications to the original 3D U-Net by Çiçek et al. [2], including residual and dense connections, separable convolutions, and hybrid CNN-Transformer models. Advanced attention mechanisms such as double attention and vision transformers, are also discussed.The survey critically evaluates studies based on performance metrics like the Dice Similarity Coefficient, sensitivity, precision, Intersection over Union (IoU), and Hausdorff Distance (HD95), using standard MRI benchmark datasets and the BraTS challenges.Focusing on 3D U-Net-based models, the survey emphasizes their ability to manage complex spatial relationships in volumetric imaging. It also identifies ongoing research challenges, such as class imbalance and computational efficiency, offering recommendations for enhancing segmentation accuracy and real-world clinical applicability.
Production planning is a key challenge for optimizing resources and responding to demand fluctuations. Demand Driven Material Requirements Planning (DDMRP) is an innovative approach to production planning. Integrating artificial intelligence and optimization algorithms into DDMRP can improve planning process.This study proposes a methodology combining the DDMRP method and genetic algorithms to optimize inventory management and buffer placement. It also incorporates Demand Driven Material Requirements Planning (DDMRP) to enhance responsiveness to demand variations. The main goal is to reduce storage costs, minimize delays, and improve the efficiency of production systems. The results of an industrial case study confirm the effectiveness of this approach in enhancing the competitiveness of production systems.
This analysis focuses on password-free Electronic IDentity (eID) solutions for eGovernment services under the federated identity management framework Electronic IDentification, Authentication and trust Services (eIDAS). The scope of eID systems is centred on their alignment of the associated technical, legal, and procedural challenges. Through an analysis of five password-free eID solutions—Fast IDentity Online 2 (FIDO2) tokens, Secure Identity Across Borders Linked (STORK), distributed ledgers, mobile authenticators, and eID cards—the study evaluates their compliance with eIDAS standards and identifies key gaps in their design and implementation. While certain solutions, such as FIDO2 tokens and mobile authenticators, demonstrate full compliance, others, including STORK and distributed ledger-based systems, face challenges in achieving interoperability, privacy, and regulatory alignment. This research contributes to the discourse on digital identity management by offering insights into current limitations and recommending pathways for advancing the design, standardization, and deployment of eID systems. These results support the larger objectives of the European digital single market by highlighting the significance of regulations, innovations, and user-oriented design in creating password-free eID systems.
Format differences present a significant challenge to the interoperability of Text Analysis tools. It is essential to consider format conversions within a robust theoretical framework that can effectively manage these conversions while ensuring that they adhere to specific properties. This paper presents an approach based on "functors" to address format conversion for electronic textual documents. This method ensures that the properties of text and tools are preserved during the process. Functors are key concepts in Category Theory as they enable us to reformulate problems from a category where they are complicated to solve to another category where solutions are more easily attainable.The main concept of this paper is to model a specific scenario. Within the category of documents that conform to a particular format f, there arises a need to parse a document D using a Text Analysis (TA) tool t that cannot interpret the format f. The challenge can be solved with the help of format conversion from f to f′, where f′ fits with t. However, we propose and discuss a method that uses functors to "transform" D and t so that the transformed t can read D with f′.
With the emergence of Industry 4.0 and 5.0, industrial ecosystems are undergoing profound transformations. This study provides a systematic review of the literature to analyze the technological, human, and environmental dimensions of these paradigms. Using Kitchenham's methodological model and the VOSviewer bibliometric tool, 56 articles were examined. The results highlight the convergence of advanced technologies (digital twins, blockchain, augmented reality) with the challenges of sustainability and human-centered innovation. While Industry 4.0 emphasizes automation and intelligent systems, Industry 5.0 introduces a new dynamic based on human-machine collaboration and ethical values. This summary identifies key technologies and major challenges and proposes concrete avenues for future research.
The rapid expansion of the Internet of Things (IoT) poses crucial challenges for the deployment of IoT applications in IT infrastructures, in particular the latency issues inherent in cloud-based platforms, despite their cost and productivity advantages. This highlights the need to rethink compliance testing frameworks for cloud environments, with a focus on improving coordination and observability during distributed data processing. To address these challenges, this study proposes a federated learning (FL) framework, in which fog nodes independently train local models and share only parameter updates - not raw data - with a central server. These updates are aggregated into a global model, which is then redistributed to IoT users for refinement using local datasets. This iterative process improves the accuracy of the global model while preserving data confidentiality at device level. Focusing on IoT edge computing, where computational tasks are decentralized to optimize resource efficiency, we present a new FL-driven test architecture designed to streamline coordination and fault detection in distributed cloud systems. The methodology is rigorously evaluated to demonstrate its effectiveness in balancing workload distribution and improving system reliability.
Faced with the accelerated growth of the automotive industry, companies are focusing their efforts on the optimization of the value chain, quality assurance, technological innovation, improvement of operational performance and achievement of operational excellence. Gradually, Lean Management has proven its ability to minimize non-value-added operations, reduce costs and optimize operations and activities with the ultimate goal of improving performance. Industry 4.0 has also confirmed its ability to create interconnected and intelligent factories, improve the flexibility, effectiveness and efficiency of processes while exploiting real-time data and optimizing the decision-making process. Compliance of quality management systems with the international automotive standard IATF 16949:2016 is a requirement of the automotive market and a determining factor in the supplier selection process. Based on an empirical study conducted among automotive companies located in Morocco and conducting bivariate correlation analysis with SPSS software, this article analyzes the synergistic relationships between Lean Management tools, IATF requirements, Industry 4.0 pillars and operational excellence. The results showed that the different paradigms interact in a synergistic way, thus demonstrating the importance of an integrated approach for better results.
Software testing plays a significant role in the development cycle process because it is necessary to ensure high-quality software and meet the client’s requirements. However, it is considered one of the most time and cost consuming activities in the development process. On the other hand, the integration of AI in software testing promises vital advances in terms of efficiency, coverage, and time saving, but the evolving software testing tools sector remains relatively conservative due to several challenges related to trustworthiness, scalability, customization, and ethical considerations. In this paper, we propose a maturity model for industry adoption of AI in software testing, designed to help organizations assess and advance their integration efforts. Grounded in a multi-stakeholder ecosystem perspective incorporating academia, PhD researchers, industry players, and governmental agencies our model identifies key stages of adoption and the conditions necessary for progress. By reframing the gap between innovation and practice through this structured lens, we offer actionable insights to align research outputs with industrial readiness and accelerate effective AI adoption in testing environments.
Continuous monitoring of fiber optic network equipment is essential to ensure optimal service quality, especially in remote rural areas. This study proposes a cost-effective real-time supervision solution for FTTH networks based on the SNMP protocol. An automated and secure diagnostic system was developed to rapidly detect faults, optimize performance, and ensure proactive maintenance, thereby contributing to a reliable user experience. The system queries network equipment via SNMP, collects data from various FTTH devices, notably OLTs and ONUs, and displays them through a user-friendly interface. The results demonstrate that this application reduces maintenance costs for companies, enables rapid fault detection, and improves service quality for users.
Academic libraries have long been essential institutions for preserving and providing access to written heritage. However, in the digital age, their role has expanded beyond traditional tasks of safeguarding and disseminating information to include more active involvement in research. This paper examines how academic libraries have evolved – parallel to the transition from physical collections to digital repositories – to include the development research infrastructures, curation of metadata, and promotion of interdisciplinary collaboration. Using the University of Bergen Library as a case study, the paper highlights the increasing engagement of academic libraries with scholarly communities and the public through active partnerships, particularly in the field of digital humanities. By contributing to collection-led research on written heritage and serving as intermediaries between research, technology, and cultural heritage institutions, academic libraries play a crucial role in advancing knowledge. The paper concludes by discussing future directions for academic libraries, emphasizing the importance of ongoing dialogue among stakeholders to sustain the relevance and impact of digital heritage initiatives.
The evaluation and assessment of government data quality have become crucial in ensuring the reliability of public services and effective decision-making. With the increasing adoption of artificial intelligence (AI) in governance, maintaining high-quality data is essential. Several models for assessing data quality have been proposed to establish a stable foundation for public service delivery. This paper focuses on the proactive publication of government data by local administrations. We propose a two-phase framework: first, applying a data quality model tailored to local government data, and second, evaluating whether open government practices contribute to improving data quality. Our findings offer insights into whether transparency initiatives enhance the quality of proactively published government data.
