Computers

55 pages, 2668 KiB

Open AccessSystematic Review

A Systematic Literature Review of Machine Unlearning Techniques in Neural Networks

by Ivanna Daniela Cevallos, Marco E. Benalcázar, Ángel Leonardo Valdivieso Caraguay, Jonathan A. Zea and Lorena Isabel Barona-López

Computers 2025, 14(4), 150; https://doi.org/10.3390/computers14040150 - 18 Apr 2025

Abstract

This review examines the field of machine unlearning in neural networks, an area driven by data privacy regulations such as the General Data Protection Regulation and the California Consumer Privacy Act. By analyzing 37 primary studies of machine unlearning applied to neural networks [...] Read more.

This review examines the field of machine unlearning in neural networks, an area driven by data privacy regulations such as the General Data Protection Regulation and the California Consumer Privacy Act. By analyzing 37 primary studies of machine unlearning applied to neural networks in both regression and classification tasks, this review thoroughly evaluates the foundational principles, key performance metrics, and methodologies used to assess these techniques. Special attention is given to recent advancements up to December 2023, including emerging approaches and frameworks. By categorizing and detailing these unlearning techniques, this work offers deeper insights into their evolution, effectiveness, efficiency, and broader applicability, thus providing a solid foundation for future research, development, and practical implementations in the realm of data privacy, model management, and compliance with evolving legal standards. Additionally, this review addresses the challenges of selectively removing data contributions at both the client and instance levels, highlighting the balance between computational costs and privacy guarantees. Full article

(This article belongs to the Special Issue Uncertainty-Aware Artificial Intelligence)

► Show Figures

Figure 1

31 pages, 5128 KiB

Open AccessArticle

Enhancing Smart Home Efficiency with Heuristic-Based Energy Optimization

by Yasir Abbas Khan, Faris Kateb, Ateeq Ur Rehman, Atif Sardar Khan, Fazal Qudus Khan, Sadeeq Jan and Ali Naser Alkhathlan

Computers 2025, 14(4), 149; https://doi.org/10.3390/computers14040149 - 16 Apr 2025

Viewed by 98

Abstract

In smart homes, heavy reliance on appliance automation has increased, along with the energy demand in developing urban areas, making efficient energy management an important factor. To address the scheduling of appliances under Demand-Side Management, this article explores the use of heuristic-based optimization [...] Read more.

In smart homes, heavy reliance on appliance automation has increased, along with the energy demand in developing urban areas, making efficient energy management an important factor. To address the scheduling of appliances under Demand-Side Management, this article explores the use of heuristic-based optimization techniques (HOTs) in smart homes (SHs) equipped with renewable and sustainable energy resources (RSERs) and energy storage systems (ESSs). The optimal model for minimization of the peak-to-average ratio (PAR), considering user comfort constraints, is validated by using different techniques, such as the Genetic Algorithm (GA), Binary Particle Swarm Optimization (BPSO), Wind-Driven Optimization (WDO), Bacterial Foraging Optimization (BFO) and the Genetic Modified Particle Swarm Optimization (GmPSO) algorithm, to minimize electricity costs, the PAR, carbon emissions and delay discomfort. This research investigates the energy optimization results of three real-world scenarios. The three scenarios demonstrate the benefits of gradually assembling RSERs and ESSs and integrating them into SHs employing HOTs. The simulation results show substantial outcomes, as in the scenario of Condition 1, GmPSO decreased carbon emissions from 300 kg to 69.23 kg, reducing emissions by 76.9%; bill prices were also cut from an unplanned value of 400.00 cents to 150 cents, a 62.5% reduction. The PAR was decreased from an unscheduled value of 4.5 to 2.2 with the GmPSO algorithm, which reduced the value by 51.1%. The scenario of Condition 2 showed that GmPSO reduced the PAR from 0.5 (unscheduled) to 0.2, a 60% reduction; the costs were reduced from 500.00 cents to 200.00 cents, a 60% reduction; and carbon emissions were reduced from 250.00 kg to 150 kg, a 60% reduction by GmPSO. In the scenario of Condition 3, where batteries and RSERs were integrated, the GmPSO algorithm reduced the carbon emission value to 158.3 kg from an unscheduled value of 208.3 kg, a reduction of 24%. The energy cost was decreased from an unplanned value of 500 cents to 300 cents with GmPSO, decreasing the overall cost by 40%. The GmPSO algorithm achieved a 57.1% reduction in the PAR value from an unscheduled value of 2.8 to 1.2. Full article

(This article belongs to the Special Issue Artificial Intelligence-Driven Innovations in Resilient Energy Systems)

► Show Figures

Figure 1

28 pages, 1521 KiB

Open AccessReview

Advancing Predictive Healthcare: A Systematic Review of Transformer Models in Electronic Health Records

by Azza Mohamed, Reem AlAleeli and Khaled Shaalan

Computers 2025, 14(4), 148; https://doi.org/10.3390/computers14040148 - 14 Apr 2025

Viewed by 119

Abstract

This systematic study seeks to evaluate the use and impact of transformer models in the healthcare domain, with a particular emphasis on their usefulness in tackling key medical difficulties and performing critical natural language processing (NLP) functions. The research questions focus on how [...] Read more.

This systematic study seeks to evaluate the use and impact of transformer models in the healthcare domain, with a particular emphasis on their usefulness in tackling key medical difficulties and performing critical natural language processing (NLP) functions. The research questions focus on how these models can improve clinical decision-making through information extraction and predictive analytics. Our findings show that transformer models, especially in applications like named entity recognition (NER) and clinical data analysis, greatly increase the accuracy and efficiency of processing unstructured data. Notably, case studies demonstrated a 30% boost in entity recognition accuracy in clinical notes and a 90% detection rate for malignancies in medical imaging. These contributions emphasize the revolutionary potential of transformer models in healthcare, and therefore their importance in enhancing resource management and patient outcomes. Furthermore, this paper emphasizes significant obstacles, such as the reliance on restricted datasets and the need for data format standardization, and provides a road map for future research to improve the applicability and performance of these models in real-world clinical settings. Full article

(This article belongs to the Special Issue Applications of Machine Learning and Artificial Intelligence for Healthcare)

► Show Figures

Figure 1

13 pages, 5610 KiB

Open AccessArticle

An Approach to Thermal Management and Performance Throttling for Federated Computation on a Low-Cost 3D ESP32-S3 Package Stack

by Yi Liu, Parth Sandeepbhai Shah, Tian Xia and Dryver Huston

Computers 2025, 14(4), 147; https://doi.org/10.3390/computers14040147 - 11 Apr 2025

Viewed by 48

Abstract

The rise of 3D heterogeneous packaging holds promise for increased performance in applications such as AI by bringing compute and memory modules into close proximity. This increased performance comes with increased thermal management challenges. This research explores the use of thermal sensing and [...] Read more.

The rise of 3D heterogeneous packaging holds promise for increased performance in applications such as AI by bringing compute and memory modules into close proximity. This increased performance comes with increased thermal management challenges. This research explores the use of thermal sensing and load throttling combined with federated computation to manage localized internal heating in a multi-3D chip package. The overall concept is that individual chiplets may heat at different rates due to operational and geometric factors. Shifting computational loads from hot to cooler chiplets can prevent local overheating while maintaining overall computational output. This concept is verified with experiments in a low-cost test vehicle. The test vehicle mimics a 3D chiplet stack with a tightly stacked assembly of SoC devices. These devices can sense and report internal temperature and dynamically adjust frequency. The configuration is for ESP32-S3 microcontrollers to work on a federated computational task, while reporting internal temperature to a host controller. The tight packing of processors causes temperatures to rise, with those internal to the stack rising more quickly than external ones. With real-time temperature monitoring, when the temperatures exceed a threshold, the AI system reduces the processor frequency, i.e., throttles the processor, to save power and dynamically shifts part of the workload to other ESP32-S3s with lower temperatures. This approach maximizes overall efficiency while maintaining thermal safety without compromising computational power. Experimental results with up to six processors confirm the validity of the concept. Full article

► Show Figures

Figure 1

23 pages, 1535 KiB

Open AccessArticle

Review of Maturity Models for Data Mining and Proposal of a Data Preparation Maturity Model Prototype for Data Mining

by Florian Hochkamp, Anne Antonia Scheidler and Markus Rabe

Computers 2025, 14(4), 146; https://doi.org/10.3390/computers14040146 - 11 Apr 2025

Viewed by 76

Abstract

Companies face uncertainties evaluating their own capabilities when implementing data analytics or data mining. Data mining is a valuable process used to analyze data and support decisions based on the knowledge generated, creating a pipeline from the phases of data collection to data [...] Read more.

Companies face uncertainties evaluating their own capabilities when implementing data analytics or data mining. Data mining is a valuable process used to analyze data and support decisions based on the knowledge generated, creating a pipeline from the phases of data collection to data preparation and data mining. In order to assess the target state of a company in terms of its data mining capabilities, maturity models are viable tools. However, no maturity models exist that focus on the data mining process and its particular phases. This article discusses existing maturity models in the broader field of data mining, such as data management, machine learning, or artificial intelligence. Focusing on the most cost-relevant phase of data preparation, the critical influences for the development of a maturity model for the phase of data preparation are identified and categorized. Additionally, a data preparation maturity model prototype is proposed. This provides the first step towards the design of a maturity model for data mining. Full article

(This article belongs to the Special Issue IT in Production and Logistics)

► Show Figures

Figure 1

25 pages, 436 KiB

Open AccessReview

Comparison of Bioelectric Signals and Their Applications in Artificial Intelligence: A Review

by Juarez-Castro Flavio Alfonso, Toledo-Rios Juan Salvador, Aceves-Fernández Marco Antonio and Tovar-Arriaga Saul

Computers 2025, 14(4), 145; https://doi.org/10.3390/computers14040145 - 11 Apr 2025

Viewed by 165

Abstract

This review examines the role of various bioelectrical signals in conjunction with artificial intelligence (AI) and analyzes how these signals are utilized in AI applications. The applications of electroencephalography (EEG), electroretinography (ERG), electromyography (EMG), electrooculography (EOG), and electrocardiography (ECG) in diagnostic and therapeutic [...] Read more.

This review examines the role of various bioelectrical signals in conjunction with artificial intelligence (AI) and analyzes how these signals are utilized in AI applications. The applications of electroencephalography (EEG), electroretinography (ERG), electromyography (EMG), electrooculography (EOG), and electrocardiography (ECG) in diagnostic and therapeutic systems are focused on. Signal processing techniques are discussed, and relevant studies that have utilized these signals in various clinical and research settings are highlighted. Advances in signal processing and classification methodologies powered by AI have significantly improved accuracy and efficiency in medical analysis. The integration of AI algorithms with bioelectrical signal processing for real-time monitoring and diagnosis, particularly in personalized medicine, is emphasized. AI-driven approaches are shown to have the potential to enhance diagnostic precision and improve patient outcomes. However, further research is needed to optimize these models for diverse clinical environments and fully exploit the interaction between bioelectrical signals and AI technologies. Full article

► Show Figures

Figure 1

35 pages, 1192 KiB

Open AccessArticle

Towards a Better Understanding of Mobile Banking App Adoption and Use: Integrating Security, Risk, and Trust into UTAUT2

by Richard Apau, Elzbieta Titis and Harjinder Singh Lallie

Computers 2025, 14(4), 144; https://doi.org/10.3390/computers14040144 - 10 Apr 2025

Viewed by 90

Abstract

This paper expands the extended unified theory of acceptance and use of technology (UTAUT2) to include four additional constructs (security, risk, institutional trust, and technology trust), providing a more comprehensive understanding of mobile banking applications (m-banking apps) adoption. It also highlights the significant [...] Read more.

This paper expands the extended unified theory of acceptance and use of technology (UTAUT2) to include four additional constructs (security, risk, institutional trust, and technology trust), providing a more comprehensive understanding of mobile banking applications (m-banking apps) adoption. It also highlights the significant role of demographic factors in moderating the impact of these constructs, offering practical insights for promoting the use of mobile devices to access and manage banking services. Data were collected using an online survey from 315 mobile banking users and analysed using covariance-based structural equation modelling (CB-SEM). Most constructs of the baseline UTAUT2 were validated in the m-banking context, with the additional constructs confirmed to affect user intention to adopt m-banking apps, except perceived risk. The model explained 79% of the variance in behavioural intention (BI), and 54.7% in use behaviour (UB), achieving higher fit than the baseline UTAUT2. Age, gender, experience, income, and education moderated the impact of perceived security and institutional trust on BI; age, education, and experience moderated technology trust on BI; and age, gender, and experience moderated perceived security on UB. The guarantee of enhanced security, advanced privacy mechanisms, and trust should be considered paramount in future strategies aimed at promoting m-banking app adoption and use. Overall, the paper advances scientific knowledge by providing a more nuanced and comprehensive framework for understanding m-banking app adoption, validating new constructs, and offering practical recommendations for promoting m-banking usage. Full article

(This article belongs to the Special Issue Multimedia Data and Network Security)

37 pages, 353 KiB

Open AccessReview

A State-of-the-Art Review of Artificial Intelligence (AI) Applications in Healthcare: Advances in Diabetes, Cancer, Epidemiology, and Mortality Prediction

by Mariano Vargas-Santiago, Diana Assaely León-Velasco, Christian Efraín Maldonado-Sifuentes and Liliana Chanona-Hernandez

Computers 2025, 14(4), 143; https://doi.org/10.3390/computers14040143 - 10 Apr 2025

Viewed by 199

Abstract

Artificial Intelligence (AI) methodologies have profoundly influenced healthcare research, particularly in chronic disease management and public health. This paper provides a comprehensive state-of-the-art review of AI’s applications across diabetes, cancer, epidemiology, and mortality prediction. The analysis highlights advancements in machine learning (ML), deep [...] Read more.

Artificial Intelligence (AI) methodologies have profoundly influenced healthcare research, particularly in chronic disease management and public health. This paper provides a comprehensive state-of-the-art review of AI’s applications across diabetes, cancer, epidemiology, and mortality prediction. The analysis highlights advancements in machine learning (ML), deep learning (DL), and natural language processing (NLP) that enable robust predictive models and decision support systems, leading to significant clinical and public health outcomes. The study examines predictive modeling, pattern recognition, and decision support applications, addressing their respective challenges and potential in real-world healthcare settings. Emphasis is placed on the emerging role of explainable AI (XAI), multimodal data fusion, and privacy-preserving techniques such as federated learning, which aim to enhance interpretability, robustness, and ethical compliance. This paper underscores the vital role of interdisciplinary collaboration and adaptive AI systems in creating resilient, scalable, and patient-centric healthcare solutions. Full article

(This article belongs to the Special Issue Applications of Machine Learning and Artificial Intelligence for Healthcare)

15 pages, 3742 KiB

Open AccessArticle

An Innovative Approach to Topic Clustering for Social Media and Web Data Using AI

by Ioannis Kapantaidakis, Emmanouil Perakakis, George Mastorakis and Ioannis Kopanakis

Computers 2025, 14(4), 142; https://doi.org/10.3390/computers14040142 - 10 Apr 2025

Viewed by 111

Abstract

The vast amount of social media and web data offers valuable insights for purposes such as brand reputation management, topic research, competitive analysis, product development, and public opinion surveys. However, analysing these data to identify patterns and extract valuable insights is challenging due [...] Read more.

The vast amount of social media and web data offers valuable insights for purposes such as brand reputation management, topic research, competitive analysis, product development, and public opinion surveys. However, analysing these data to identify patterns and extract valuable insights is challenging due to the vast number of posts, which can number in the thousands within a single day. One practical approach is topic clustering, which creates clusters of mentions that refer to a specific topic. Following this process will create several manageable clusters, each containing hundreds or thousands of posts. These clusters offer a more meaningful overview of the discussed topics, eliminating the need to categorise each post manually. Several topic detection algorithms can achieve clustering of posts, such as LDA, NMF, BERTopic, etc. The existing algorithms, however, have several important drawbacks, including language constraints and slow or resource-intensive data processing. Moreover, the labels for the clusters typically consist of a few keywords that may not make sense unless one explores the mentions within the cluster. Recently, with the introduction of AI large language models, such as GPT-4, new techniques can be realised for topic clustering to address the aforementioned issues. Our novel approach (AI Mention Clustering) employs LLMs at its core to produce an algorithm for efficient and accurate topic clustering of web and social data. Our solution was tested on social and web data and compared to the popular existing algorithm of BERTopic, demonstrating superior resource efficiency and absolute accuracy of clustered documents. Furthermore, it produces summaries of the clusters that are easily understood by humans instead of just representative keywords. This approach enhances the productivity of social and web data researchers by providing more meaningful and interpretable results. Full article

(This article belongs to the Special Issue Harnessing Artificial Intelligence for Social and Semantic Understanding)

► Show Figures

Figure 1

36 pages, 1629 KiB

Open AccessReview

A Systematic Review of Blockchain-Based Initiatives in Comparison to Best Practices Used in Higher Education Institutions

by Diana Laura Silaghi and Daniela Elena Popescu

Computers 2025, 14(4), 141; https://doi.org/10.3390/computers14040141 - 8 Apr 2025

Viewed by 207

Abstract

Blockchain technology, originally introduced through Bitcoin cryptocurrency in 2008, has rapidly expanded beyond its financial roots, offering innovative solutions for secure data management across various sectors, including education. Higher education institutions, faced with challenges in managing academic records, verifying degrees, assessing skills, and [...] Read more.

Blockchain technology, originally introduced through Bitcoin cryptocurrency in 2008, has rapidly expanded beyond its financial roots, offering innovative solutions for secure data management across various sectors, including education. Higher education institutions, faced with challenges in managing academic records, verifying degrees, assessing skills, and safeguarding personal data, have increasingly looked to blockchain for answers. Blockchain’s transparent, immutable, and decentralized nature provides potential solutions to these longstanding problems. This systematic review assesses blockchain-based proposals for academic certificates management, aiming to highlight globally recognized best practices, explore the latest applications, and identify key challenges hindering the widespread adoption of blockchain technology in education. A thorough discussion based on the findings introduces potential solutions to mitigate these challenges and provides insights into possible future research directions that could help overcome these obstacles. Full article

(This article belongs to the Special Issue Harnessing the Blockchain Technology in Unveiling Futuristic Applications)

► Show Figures

Figure 1

30 pages, 14418 KiB

Open AccessArticle

LAVID: A Lightweight and Autonomous Smart Camera System for Urban Violence Detection and Geolocation

by Mohammed Azzakhnini, Houda Saidi, Ahmed Azough, Hamid Tairi and Hassan Qjidaa

Computers 2025, 14(4), 140; https://doi.org/10.3390/computers14040140 - 7 Apr 2025

Viewed by 117

Abstract

With the rise of digital video technologies and the proliferation of processing methods and storage systems, video-surveillance systems have received increasing attention over the last decade. However, the spread of cameras installed in public and private spaces makes it more difficult for human [...] Read more.

With the rise of digital video technologies and the proliferation of processing methods and storage systems, video-surveillance systems have received increasing attention over the last decade. However, the spread of cameras installed in public and private spaces makes it more difficult for human operators to perform real-time analysis of the large amounts of data produced by surveillance systems. Due to the advancement of artificial intelligence methods, many automatic video analysis tasks like violence detection have been studied from a research perspective, and are even beginning to be commercialized in industrial solutions. Nevertheless, most of these solutions adopt centralized architectures with costly servers utilized to process streaming videos sent from different cameras. Centralized architectures do not present the ideal solution due to the high cost, processing time issues, and network bandwidth overhead. In this paper, we propose a lightweight autonomous system for the detection and geolocation of violent acts. Our proposed system, named LAVID, is based on a depthwise separable convolution model (DSCNN) combined with a bidirectional long-short-term memory network (BiLSTM) and implemented on a lightweight smart camera. We provide in this study a lightweight video-surveillance system consisting of low-cost autonomous smart cameras that are capable of detecting and identifying harmful behavior and geolocate violent acts that occur over a covered area in real-time. Our proposed system, implemented using Raspberry Pi boards, represents a cost-effective solution with interoperability features making it an ideal IoT solution to be integrated with other smart city infrastructure. Furthermore, our approach, implemented using optimized deep learning models and evaluated on several public datasets, has shown good results in term of accuracy compared to state of the art methods while optimizing reducing power and computational requirements. Full article

(This article belongs to the Section Internet of Things (IoT) and Industrial IoT)

► Show Figures

Figure 1

27 pages, 744 KiB

Open AccessArticle

Microhooks: A Novel Framework to Streamline the Development of Microservices

by Omar Iraqi, Mohamed El Kadiri El Hassani and Anass Zouine

Computers 2025, 14(4), 139; https://doi.org/10.3390/computers14040139 - 7 Apr 2025

Viewed by 104

Abstract

The microservices architectural style has gained widespread adoption in recent years thanks to its ability to deliver high scalability and maintainability. However, the development process for microservices-based applications can be complex and challenging. Indeed, it often requires developers to manage a large number [...] Read more.

The microservices architectural style has gained widespread adoption in recent years thanks to its ability to deliver high scalability and maintainability. However, the development process for microservices-based applications can be complex and challenging. Indeed, it often requires developers to manage a large number of distributed components with the burden of handling low-level, recurring needs, such as inter-service communication, brokering, event management, and data replication. In this article, we present Microhooks: a novel framework designed to streamline the development of microservices by allowing developers to focus on their business logic while declaratively expressing the so-called low-level needs. Based on the inversion of control and the materialized view patterns, among others, our framework automatically generates and injects the corresponding artifacts, leveraging 100% build time code introspection and instrumentation, as well as context building, for optimized runtime performance. We provide the first implementation for the Java world, supporting the most popular containers and brokers, and adhering to the standard Java/Jakarta Persistence API. From the user perspective, Microhooks exposes an intuitive, container-agnostic, broker-neutral, and ORM framework-independent API. Microhooks evaluation against state-of-the-art practices has demonstrated its effectiveness in drastically reducing code size and complexity, without incurring any considerable cost on performance. Based on such promising results, we believe that Microhooks has the potential to become an essential component of the microservices development ecosystem. Full article

► Show Figures

Figure 1

14 pages, 3430 KiB

Open AccessArticle

Optimal Selection of Sampling Rates and Mother Wavelet for an Algorithm to Classify Power Quality Disturbances

by Jonatan A. Medina-Molina, Enrique Reyes-Archundia, José A. Gutiérrez-Gnecchi, Javier A. Rodríguez-Herrejón, Marco V. Chávez-Báez, Juan C. Olivares-Rojas and Néstor F. Guerrero-Rodríguez

Computers 2025, 14(4), 138; https://doi.org/10.3390/computers14040138 - 6 Apr 2025

Viewed by 109

Abstract

The introduction of renewable energy sources, distributed energy systems, and power electronics equipment has led to the emergence of the Smart Grid. However, these developments have also caused the worsening of power quality. Selecting the correct sampling frequency and feature extraction techniques are [...] Read more.

The introduction of renewable energy sources, distributed energy systems, and power electronics equipment has led to the emergence of the Smart Grid. However, these developments have also caused the worsening of power quality. Selecting the correct sampling frequency and feature extraction techniques are essential for appropriately analyzing power quality disturbances. This work compares the performance of an algorithm based on a Support Vector Machine and Discrete Wavelet Transform for the classification of power quality disturbances using eight sampling rates and five different mother wavelets. The algorithm was tested in noisy and noiseless scenarios to show the methodology. The results indicate that a success rate of 99.9% is obtained for the noiseless signals using a sampling rate of 9.6 kHz and 95.2% for signals with a signal-to-noise ratio of 30 dB with a sampling rate of 30 kHz. Full article

(This article belongs to the Special Issue Artificial Intelligence-Driven Innovations in Resilient Energy Systems)

► Show Figures

Figure 1

16 pages, 434 KiB

Open AccessArticle

Quantum Testing of Recommender Algorithms on GPU-Based Quantum Simulators

by Chenxi Liu, W. Bernard Lee and Anthony G. Constantinides

Computers 2025, 14(4), 137; https://doi.org/10.3390/computers14040137 - 6 Apr 2025

Viewed by 107

Abstract

This study explores the application of quantum computing in asset management, focusing on the use of the Quantum Approximate Optimization Algorithm (QAOA) to solve specific classes of financial asset recommendation problems. While quantum computing holds promise for combinatorial optimization tasks, its application to [...] Read more.

This study explores the application of quantum computing in asset management, focusing on the use of the Quantum Approximate Optimization Algorithm (QAOA) to solve specific classes of financial asset recommendation problems. While quantum computing holds promise for combinatorial optimization tasks, its application to portfolio management faces significant challenges in scalability for practical implementations. In this work, we model the problem using a graph representation where nodes represent investors, and edges reflect significant similarities in asset choices. We test the proposed method using quantum simulators, including cuQuantum, Cirq-GPU, and Cirq with IonQ, and compare the performance of quantum optimization against classical brute-force methods. Our results suggest that quantum algorithms may offer computational advantages for certain use cases, though classical heuristics also provide competitive performance for smaller datasets. This study contributes to the ongoing investigation into the potential of quantum computing for real-time financial decision-making, providing insights into both its applicability and limitations in asset management for larger and more complex investor datasets. Full article

► Show Figures

Figure 1

19 pages, 2173 KiB

Open AccessArticle

Digital Twins and the Stendhal Syndrome

by Franco Niccolucci and Achille Felicetti

Computers 2025, 14(4), 136; https://doi.org/10.3390/computers14040136 - 6 Apr 2025

Viewed by 97

Abstract

The “Stendhal Syndrome” mentioned in the title refers to the first (early 19th century) documented perception of the role of intangible aspects in characterising cultural heritage. This paper addresses the semantic organisation of data concerning the digital documentation of cultural heritage, considering its [...] Read more.

The “Stendhal Syndrome” mentioned in the title refers to the first (early 19th century) documented perception of the role of intangible aspects in characterising cultural heritage. This paper addresses the semantic organisation of data concerning the digital documentation of cultural heritage, considering its intangible dimension in the framework of Digital Twins. The intangible component was one of the aspects motivating the need of setting up the Heritage Digital Twin (HDT) ontology and its extensions, published in a series of papers since early 2023. In this paper, we analyse how places, persons, and things may give value to a heritage asset, being linked to and supporting its intrinsic cultural significance. This development stems from the consideration of heritage studies and research carried out by scholars and organisations such as UNESCO and ICOMOS, which underline the paramount role of the intangible component in defining heritage assets. The paper then expands the previous semantic structure of the Heritage Digital Twin ontology as concerns the intangible aspects of a heritage asset, extending the HDT concepts by defining new classes and properties related to its intangible component. These are discussed in various cases concerning places, monuments, objects, and persons, and fully developed in examples. Full article

► Show Figures

Figure 1

32 pages, 469 KiB

Open AccessArticle

Enhancing Cryptographic Solutions for Resource-Constrained RFID Assistive Devices: Implementing a Resource-Efficient Field Montgomery Multiplier

by Atef Ibrahim and Fayez Gebali

Computers 2025, 14(4), 135; https://doi.org/10.3390/computers14040135 - 6 Apr 2025

Viewed by 80

Abstract

Radio Frequency Identification (RFID) assistive systems, which integrate RFID devices with IoT technologies, are vital for enhancing the independence, mobility, and safety of individuals with disabilities. These systems enable applications such as RFID navigation for blind users and RFID-enabled canes that provide real-time [...] Read more.

Radio Frequency Identification (RFID) assistive systems, which integrate RFID devices with IoT technologies, are vital for enhancing the independence, mobility, and safety of individuals with disabilities. These systems enable applications such as RFID navigation for blind users and RFID-enabled canes that provide real-time location data. Central to these systems are resource-constrained RFID devices that rely on RFID tags to collect and transmit data, but their limited computational capabilities make them vulnerable to cyberattacks, jeopardizing user safety and privacy. Implementing the Elliptic Curve Cryptography (ECC) algorithm is essential to mitigate these risks; however, its high computational complexity exceeds the capabilities of these devices. The fundamental operation of ECC is finite field multiplication, which is crucial for securing data. Optimizing this operation allows ECC computations to be executed without overloading the devices’ limited resources. Traditional multiplication designs are often unsuitable for such devices due to their excessive area and energy requirements. Therefore, this work tackles these challenges by proposing an efficient and compact field multiplier design optimized for the Montgomery multiplication algorithm, a widely used method in cryptographic applications. The proposed design significantly reduces both space and energy consumption while maintaining computational performance, making it well-suited for resource-constrained environments. ASIC synthesis results demonstrate substantial improvements in key metrics, including area, power consumption, Power-Delay Product (PDP), and Area-Delay Product (ADP), highlighting the multiplier’s efficiency and practicality. This innovation enables the implementation of ECC on RFID assistive devices, enhancing their security and reliability, thereby allowing individuals with disabilities to engage with assistive technologies more safely and confidently. Full article

(This article belongs to the Special Issue Wearable Computing and Activity Recognition)

► Show Figures

Figure 1

17 pages, 5373 KiB

Open AccessArticle

Real-Time Overhead Power Line Component Detection on Edge Computing Platforms

by Nico Surantha

Computers 2025, 14(4), 134; https://doi.org/10.3390/computers14040134 - 5 Apr 2025

Viewed by 89

Abstract

Regular inspection of overhead power line (OPL) systems is required to detect damage early and ensure the efficient and uninterrupted transmission of high-voltage electric power. In the past, these checks were conducted utilizing line crawling, inspection robots, and a helicopter. Yet, these traditional [...] Read more.

Regular inspection of overhead power line (OPL) systems is required to detect damage early and ensure the efficient and uninterrupted transmission of high-voltage electric power. In the past, these checks were conducted utilizing line crawling, inspection robots, and a helicopter. Yet, these traditional solutions are slow, costly, and hazardous. Advancements in drones, edge computing platforms, deep learning, and high-resolution cameras may enable real-time OPL inspections using drones. Some research has been conducted on OPL inspection with autonomous drones. However, it is essential to explore how to achieve real-time OPL component detection effectively and efficiently. In this paper, we report our research on OPL component detection on edge computing devices. The original OPL dataset is generated in this study. In this paper, we evaluate the detection performance with several sizes of training datasets. We also implement simple data augmentation to extend the size of datasets. The performance of the YOLOv7 model is also evaluated on several edge computing platforms, such as Raspberry Pi 4B, Jetson Nano, and Jetson Orin Nano. The model quantization method is used to improve the real-time performance of the detection model. The simulation results show that the proposed YOLOv7 model can achieve mean average precision (mAP) over 90%. While the hardware evaluation shows the real-time detection performance can be achieved in several circumstances. Full article

(This article belongs to the Special Issue Distributed Computing Paradigms for the Internet of Things: Exploring Cloud, Edge, and Fog Solutions)

► Show Figures

Figure 1

33 pages, 1066 KiB

Open AccessReview

The Ontology-Based Mapping of Microservice Identification Approaches: A Systematic Study of Migration Strategies from Monolithic to Microservice Architectures

by Idris Oumoussa and Rajaa Saidi

Computers 2025, 14(4), 133; https://doi.org/10.3390/computers14040133 - 5 Apr 2025

Viewed by 70

Abstract

The Microservice Architecture Style (MSA) has emerged as a significant computing paradigm in software engineering, with companies increasingly restructuring their monolithic systems to enhance digital performance and competitiveness. However, the migration process, particularly the microservice identification phase, presents complex challenges that require careful [...] Read more.

The Microservice Architecture Style (MSA) has emerged as a significant computing paradigm in software engineering, with companies increasingly restructuring their monolithic systems to enhance digital performance and competitiveness. However, the migration process, particularly the microservice identification phase, presents complex challenges that require careful consideration. This study aimed to provide developers and researchers with a practical roadmap for microservice identification during legacy system migration while highlighting crucial migration steps and research requirements. Through a systematic mapping study following Kitchenham and Petersen’s guidelines, we analyzed various microservice identification approaches and developed a middleweight ontology that can be queried for key inputs, data modeling, identification algorithms, and performance evaluation metrics. Our research makes several significant contributions: a comprehensive analysis of existing identification methodologies, a multi-dimensional framework for categorizing and evaluating approaches, an examination of current research trajectories and literature gaps, an ontological framework specifically designed for microservice identification, and an outline of pressing challenges and future research directions. The study concluded that microservice identification remains a significant barrier in system migration efforts, highlighting the need for more research focused on developing effective identification techniques that consider various aspects, including roles and dependencies within a microservice architecture. This comprehensive analysis provides valuable insights for professionals and researchers working on microservice migration projects. Full article

(This article belongs to the Special Issue Best Practices, Challenges and Opportunities in Software Engineering)

► Show Figures

Figure 1

23 pages, 3151 KiB

Open AccessArticle

Scalability and Efficiency Analysis of Hyperledger Fabric and Private Ethereum in Smart Contract Execution

by Maaz Muhammad Khan, Fahd Sikandar Khan, Muhammad Nadeem, Taimur Hayat Khan, Shahab Haider and Dani Daas

Computers 2025, 14(4), 132; https://doi.org/10.3390/computers14040132 - 3 Apr 2025

Viewed by 229

Abstract

Blockchain technology has emerged as a transformative solution for secure, immutable, and decentralized data management across diverse domains, including economics, healthcare, and supply chain management. Given its soaring adoption, it is crucial to assess the suitability of various blockchain platforms for specific applications. [...] Read more.

Blockchain technology has emerged as a transformative solution for secure, immutable, and decentralized data management across diverse domains, including economics, healthcare, and supply chain management. Given its soaring adoption, it is crucial to assess the suitability of various blockchain platforms for specific applications. This study evaluates the performance of Hyperledger Fabric (HF) and private Ethereum (Geth) to analyze their scalability (node count), throughput (transactions per second (TPS)), and latency (measured in milliseconds). A benchmarking tool was developed in-house to assess the execution of key smart contract functions—QueryUser, CreateUser, TransferMoney, and IssueMoney—under varying transaction loads (10–1000 transactions) and network sizes (2–16 node count). The results indicate that HF performs significantly better than private Ethereum in terms of invoke functions, achieving up to 5× throughput and up to 26× lower latency. However, private Ethereum excels in query operations because of its account-based ledger model. While Hyperledger Fabric scales efficiently within moderate transaction volumes, it experiences concurrency limitations beyond 1000 transactions, whereas private Ethereum processes up to 10,000 transactions, albeit with performance fluctuations due to gas fees. The findings offer valuable insights into the strengths and tradeoffs of both platforms, informing optimal blockchain selection for enterprise applications that require high transaction efficiency. Full article

(This article belongs to the Special Issue Next Generation Blockchain, Information Security and Soft Computing for Future IoT Networks)

► Show Figures

Figure 1

36 pages, 2159 KiB

Open AccessReview

Employing Blockchain, NFTs, and Digital Certificates for Unparalleled Authenticity and Data Protection in Source Code: A Systematic Review

by Leonardo Juan Ramirez Lopez and Genesis Gabriela Morillo Ledezma

Computers 2025, 14(4), 131; https://doi.org/10.3390/computers14040131 - 2 Apr 2025

Viewed by 214

Abstract

In higher education, especially in programming-intensive fields like computer science, safeguarding students’ source code is crucial to prevent theft that could impact learning and future careers. Traditional storage solutions like Google Drive are vulnerable to hacking and alterations, highlighting the need for stronger [...] Read more.

In higher education, especially in programming-intensive fields like computer science, safeguarding students’ source code is crucial to prevent theft that could impact learning and future careers. Traditional storage solutions like Google Drive are vulnerable to hacking and alterations, highlighting the need for stronger protection. This work explores digital technologies that enhance source code security, with a focus on Blockchain and NFTs. Due to Blockchain’s decentralized and immutable nature, NFTs can be used to control code ownership, improving security, traceability, and preventing unauthorized access. This approach effectively addresses existing gaps in protecting academic intellectual property. However, as Bennett et al. highlight, while these technologies have significant potential, challenges remain in large-scale implementation and user acceptance. Despite these hurdles, integrating Blockchain and NFTs presents a promising opportunity to enhance academic integrity. Successful adoption in educational settings may require a more inclusive and innovative strategy. Full article

(This article belongs to the Section Blockchain Infrastructures and Enabled Applications)

► Show Figures

Figure 1

23 pages, 1956 KiB

Open AccessArticle

Artificial Intelligence in Neoplasticism: Aesthetic Evaluation and Creative Potential

by Su Jin Mun and Won Ho Choi

Computers 2025, 14(4), 130; https://doi.org/10.3390/computers14040130 - 2 Apr 2025

Viewed by 256

Abstract

This research investigates the aesthetic evaluation of AI-generated neoplasticist artworks, exploring how well artificial intelligence systems, specifically Midjourney, replicate the core principles of neoplasticism, such as geometric forms, balance, and color harmony. The background of this study stems from ongoing debates about the [...] Read more.

This research investigates the aesthetic evaluation of AI-generated neoplasticist artworks, exploring how well artificial intelligence systems, specifically Midjourney, replicate the core principles of neoplasticism, such as geometric forms, balance, and color harmony. The background of this study stems from ongoing debates about the legitimacy of AI-generated art and how these systems engage with established artistic movements. The purpose of the research is to assess whether AI can produce artworks that meet aesthetic standards comparable to human-created works. The research utilized Monroe C. Beardsley’s aesthetic emotion criteria and Noël Carroll’s aesthetic experience criteria as a framework for evaluating the artworks. A logistic regression analysis was conducted to identify key compositional elements in AI-generated neoplasticist works. The findings revealed that AI systems excelled in areas such as unity, color diversity, and overall artistic appeal but showed limitations in handling monochromatic elements. The implications of this research suggest that while AI can produce high-quality art, further refinement is needed for more subtle aspects of design. This study contributes to understanding the potential of AI as a tool in the creative process, offering insights for both artists and AI developers. Full article

(This article belongs to the Special Issue AI in Its Ecosystem)

► Show Figures

Figure 1

35 pages, 1880 KiB

Open AccessArticle

Strengthening Cybersecurity Resilience: An Investigation of Customers’ Adoption of Emerging Security Tools in Mobile Banking Apps

by Irfan Riasat, Mahmood Shah and M. Sinan Gonul

Computers 2025, 14(4), 129; https://doi.org/10.3390/computers14040129 - 1 Apr 2025

Viewed by 183

Abstract

The rise in internet-based services has raised risks of data exposure. The manipulation and exploitation of sensitive data significantly impact individuals’ resilience—the ability to protect and prepare against cyber incidents. Emerging technologies seek to enhance cybersecurity resilience by developing various security tools. This [...] Read more.

The rise in internet-based services has raised risks of data exposure. The manipulation and exploitation of sensitive data significantly impact individuals’ resilience—the ability to protect and prepare against cyber incidents. Emerging technologies seek to enhance cybersecurity resilience by developing various security tools. This study aims to explore the adoption of security tools using a qualitative research approach. Twenty-two semi-structured interviews were conducted with users of mobile banking apps from Pakistan. Data were analyzed using thematic analysis, which revealed that biometric authentication and SMS alerts are commonly used. Limited use of multifactor authentication has been observed, mainly due to a lack of awareness or implementation knowledge. Passwords are still regarded as a trusted and secure mechanism. The findings indicate that the adoption of security tools is based on perceptions of usefulness, perceived trust, and perceived ease of use, while knowledge and awareness play a moderating role. This study also proposes a framework by extending TAM to include multiple security tools and introducing knowledge and awareness as a moderator influencing users’ perceptions. The findings inform practical implications for financial institutions, application developers, and policymakers to ensure standardized policy to include security tools in online financial platforms, thereby enhancing overall cybersecurity resilience. Full article

► Show Figures

Figure 1

19 pages, 13596 KiB

Open AccessArticle

SMS3D: 3D Synthetic Mushroom Scenes Dataset for 3D Object Detection and Pose Estimation

by Abdollah Zakeri, Bikram Koirala, Jiming Kang, Venkatesh Balan, Weihang Zhu, Driss Benhaddou and Fatima A. Merchant

Computers 2025, 14(4), 128; https://doi.org/10.3390/computers14040128 - 1 Apr 2025

Viewed by 97

Abstract

The mushroom farming industry struggles to automate harvesting due to limited large-scale annotated datasets and the complex growth patterns of mushrooms, which complicate detection, segmentation, and pose estimation. To address this, we introduce a synthetic dataset with 40,000 unique scenes of white Agaricus [...] Read more.

The mushroom farming industry struggles to automate harvesting due to limited large-scale annotated datasets and the complex growth patterns of mushrooms, which complicate detection, segmentation, and pose estimation. To address this, we introduce a synthetic dataset with 40,000 unique scenes of white Agaricus bisporus and brown baby bella mushrooms, capturing realistic variations in quantity, position, orientation, and growth stages. Our two-stage pose estimation pipeline combines 2D object detection and instance segmentation with a 3D point cloud-based pose estimation network using a Point Transformer. By employing a continuous 6D rotation representation and a geodesic loss, our method ensures precise rotation predictions. Experiments show that processing point clouds with 1024 points and the 6D Gram–Schmidt rotation representation yields optimal results, achieving an average rotational error of

1.67 °

on synthetic data, surpassing current state-of-the-art methods in mushroom pose estimation. The model, further, generalizes well to real-world data, attaining a mean angle difference of

3.68 °

on a subset of the M18K dataset with ground-truth annotations. This approach aims to drive automation in harvesting, growth monitoring, and quality assessment in the mushroom industry. Full article

(This article belongs to the Special Issue Advanced Image Processing and Computer Vision (2nd Edition))

► Show Figures

Figure 1

23 pages, 2184 KiB

Open AccessArticle

Lossless Compression of Malaria-Infected Erythrocyte Images Using Vision Transformer and Deep Autoencoders

by Md Firoz Mahmud, Zerin Nusrat and W. David Pan

Computers 2025, 14(4), 127; https://doi.org/10.3390/computers14040127 - 1 Apr 2025

Viewed by 103

Abstract

Lossless compression of medical images allows for rapid image data exchange and faithful recovery of the compressed data for medical image assessment. There are many useful telemedicine applications, for example in diagnosing conditions such as malaria in resource-limited regions. This paper presents a [...] Read more.

Lossless compression of medical images allows for rapid image data exchange and faithful recovery of the compressed data for medical image assessment. There are many useful telemedicine applications, for example in diagnosing conditions such as malaria in resource-limited regions. This paper presents a novel machine learning-based approach where lossless compression of malaria-infected erythrocyte images is assisted by cutting-edge classifiers. To this end, we first use a Vision Transformer to classify images into two categories: those cells that are infected with malaria and those that are not. We then employ distinct deep autoencoders for each category, which not only reduces the dimensions of the image data but also preserves crucial diagnostic information. To ensure no loss in reconstructed image quality, we further compress the residuals produced by these autoencoders using the Huffman code. Simulation results show that the proposed method achieves lower overall bit rates and thus higher compression ratios than traditional compression schemes such as JPEG 2000, JPEG-LS, and CALIC. This strategy holds significant potential for effective telemedicine applications and can improve diagnostic capabilities in regions impacted by malaria. Full article

(This article belongs to the Special Issue Applications of Machine Learning and Artificial Intelligence for Healthcare)

► Show Figures

Figure 1

15 pages, 387 KiB

Open AccessArticle

Analyzing Digital Political Campaigning Through Machine Learning: An Exploratory Study for the Italian Campaign for European Union Parliament Election in 2024

by Paolo Sernani, Angela Cossiri, Giovanni Di Cosimo and Emanuele Frontoni

Computers 2025, 14(4), 126; https://doi.org/10.3390/computers14040126 - 30 Mar 2025

Viewed by 145

Abstract

The rapid digitalization of political campaigns has reshaped electioneering strategies, enabling political entities to leverage social media for targeted outreach. This study investigates the impact of digital political campaigning during the 2024 EU elections using machine learning techniques to analyze social media dynamics. [...] Read more.

The rapid digitalization of political campaigns has reshaped electioneering strategies, enabling political entities to leverage social media for targeted outreach. This study investigates the impact of digital political campaigning during the 2024 EU elections using machine learning techniques to analyze social media dynamics. We introduce a novel dataset—Political Popularity Campaign—which comprises social media posts, engagement metrics, and multimedia content from the electoral period. By applying predictive modeling, we estimate key indicators such as post popularity and assess their influence on campaign outcomes. Our findings highlight the significance of micro-targeting practices, the role of algorithmic biases, and the risks associated with disinformation in shaping public opinion. Moreover, this research contributes to the broader discussion on regulating digital campaigning by providing analytical models that can aid policymakers and public authorities in monitoring election compliance and transparency. The study underscores the necessity for robust frameworks to balance the advantages of digital political engagement with the challenges of ensuring fair democratic processes. Full article

(This article belongs to the Special Issue Machine Learning and Statistical Learning with Applications 2025)

► Show Figures

Figure 1

23 pages, 839 KiB

Open AccessArticle

Introducing a New Genetic Operator Based on Differential Evolution for the Effective Training of Neural Networks

by Ioannis G. Tsoulos, Vasileios Charilogis and Dimitrios Tsalikakis

Computers 2025, 14(4), 125; https://doi.org/10.3390/computers14040125 - 28 Mar 2025

Viewed by 213

Abstract

Artificial neural networks are widely established models used to solve a variety of real-world problems in the fields of physics, chemistry, etc. These machine learning models contain a series of parameters that must be appropriately tuned by various optimization techniques in order to [...] Read more.

Artificial neural networks are widely established models used to solve a variety of real-world problems in the fields of physics, chemistry, etc. These machine learning models contain a series of parameters that must be appropriately tuned by various optimization techniques in order to effectively address the problems that they face. Genetic algorithms have been used in many cases in the recent literature to train artificial neural networks, and various modifications have been made to enhance this procedure. In this article, the incorporation of a novel genetic operator into genetic algorithms is proposed to effectively train artificial neural networks. The new operator is based on the differential evolution technique, and it is periodically applied to randomly selected chromosomes from the genetic population. Furthermore, to determine a promising range of values for the parameters of the artificial neural network, an additional genetic algorithm is executed before the execution of the basic algorithm. The modified genetic algorithm is used to train neural networks on classification and regression datasets, and the results are reported and compared with those of other methods used to train neural networks. Full article

(This article belongs to the Special Issue Emerging Trends in Machine Learning and Artificial Intelligence)

► Show Figures

Figure 1

44 pages, 29360 KiB

Open AccessReview

Advances in Federated Learning: Applications and Challenges in Smart Building Environments and Beyond

by Mohamed Rafik Aymene Berkani, Ammar Chouchane, Yassine Himeur, Abdelmalik Ouamane, Sami Miniaoui, Shadi Atalla, Wathiq Mansoor and Hussain Al-Ahmad

Computers 2025, 14(4), 124; https://doi.org/10.3390/computers14040124 - 27 Mar 2025

Viewed by 483

Abstract

Federated Learning (FL) is a transformative decentralized approach in machine learning and deep learning, offering enhanced privacy, scalability, and data security. This review paper explores the foundational concepts, and architectural variations of FL, prominent aggregation algorithms like FedAvg, FedProx, and FedMA, and diverse [...] Read more.

Federated Learning (FL) is a transformative decentralized approach in machine learning and deep learning, offering enhanced privacy, scalability, and data security. This review paper explores the foundational concepts, and architectural variations of FL, prominent aggregation algorithms like FedAvg, FedProx, and FedMA, and diverse innovative applications in thermal comfort optimization, energy prediction, healthcare, and anomaly detection within smart buildings. By enabling collaborative model training without centralizing sensitive data, FL ensures privacy and robust performance across heterogeneous environments. We further discuss the integration of FL with advanced technologies, including digital twins and 5G/6G networks, and demonstrate its potential to revolutionize real-time monitoring, and optimize resources. Despite these advances, FL still faces challenges, such as communication overhead, security issues, and non-IID data handling. Future research directions highlight the development of adaptive learning methods, robust privacy measures, and hybrid architectures to fully leverage FL’s potential in driving innovative, secure, and efficient intelligence for the next generation of smart buildings. Full article

► Show Figures

Figure 1

20 pages, 2914 KiB

Open AccessArticle

Cross-Dataset Data Augmentation Using UMAP for Deep Learning-Based Wind Speed Prediction

by Eder Arley Leon-Gomez, Andrés Marino Álvarez-Meza and German Castellanos-Dominguez

Computers 2025, 14(4), 123; https://doi.org/10.3390/computers14040123 - 27 Mar 2025

Viewed by 141

Abstract

Wind energy has emerged as a cornerstone in global efforts to transition to renewable energy, driven by its low environmental impact and significant generation potential. However, the inherent intermittency of wind, influenced by complex and dynamic atmospheric patterns, poses significant challenges for accurate [...] Read more.

Wind energy has emerged as a cornerstone in global efforts to transition to renewable energy, driven by its low environmental impact and significant generation potential. However, the inherent intermittency of wind, influenced by complex and dynamic atmospheric patterns, poses significant challenges for accurate wind speed prediction. Existing approaches, including statistical methods, machine learning, and deep learning, often struggle with limitations such as non-linearity, non-stationarity, computational demands, and the requirement for extensive, high-quality datasets. In response to these challenges, we propose a novel neighborhood preserving cross-dataset data augmentation framework for high-horizon wind speed prediction. The proposed method addresses data variability and dynamic behaviors through three key components: (i) the uniform manifold approximation and projection (UMAP) is employed as a non-linear dimensionality reduction technique to encode local relationships in wind speed time-series data while preserving neighborhood structures, (ii) a localized cross-dataset data augmentation (DA) approach is introduced using UMAP-reduced spaces to enhance data diversity and mitigate variability across datasets, and (iii) recurrent neural networks (RNNs) are trained on the augmented datasets to model temporal dependencies and non-linear patterns effectively. Our framework was evaluated using datasets from diverse geographical locations, including the Argonne Weather Observatory (USA), Chengdu Airport (China), and Beijing Capital International Airport (China). Comparative tests using regression-based measures on RNN, GRU, and LSTM architectures showed that the proposed method was better at improving the accuracy and generalizability of predictions, leading to an average reduction in prediction error. Consequently, our study highlights the potential of integrating advanced dimensionality reduction, data augmentation, and deep learning techniques to address critical challenges in renewable energy forecasting. Full article

(This article belongs to the Special Issue Machine Learning and Statistical Learning with Applications 2025)

► Show Figures

Figure 1

25 pages, 2924 KiB

Open AccessArticle

Learning Analytics to Guide Serious Game Development: A Case Study Using Articoding

by Antonio Calvo-Morata, Cristina Alonso-Fernández, Julio Santilario-Berthilier, Iván Martínez-Ortiz and Baltasar Fernández-Manjón

Computers 2025, 14(4), 122; https://doi.org/10.3390/computers14040122 - 27 Mar 2025

Viewed by 161

Abstract

Serious games are powerful interactive environments that provide more authentic experiences for learning or training different skills. However, developing effective serious games is complex, and a more systematic approach is needed to create better evidence-based games. Learning analytics—based on the analysis of collected [...] Read more.

Serious games are powerful interactive environments that provide more authentic experiences for learning or training different skills. However, developing effective serious games is complex, and a more systematic approach is needed to create better evidence-based games. Learning analytics—based on the analysis of collected in-game user interactions—can support game development and the players’ learning process, providing assessment information to teachers, students, and other stakeholders. However, empirical studies applying and demonstrating the use of learning analytics in the context of serious games in real environments remain scarce. In this paper, we study the application of learning analytics throughout the whole lifecycle of a serious game, in order to assess the game’s design and players’ learning using a serious game that introduces basic programming concepts through a visual programming language. The game was played by N = 134 high school students in two 50-min sessions. During the game sessions, all player interactions were collected, including the time spent solving levels, their programming solutions, and the number of replays. We analyzed these interaction traces to gain insights that can facilitate teachers’ use of serious games in their lessons and assessments, as well as guide developers in making possible improvements to the game. Among these insights, knowing which tasks students struggle with is critical for both teachers and game developers, and can also reveal game design issues. Among the results obtained through analysis of the interaction data, we found differences between boys and girls when playing. Girls play in a more reflexive way and, in terms of acceptance of the game, a higher percentage of girls had neutral opinions. We also found the most repeated errors, the level each player reached, and how long it took them to reach those levels. These data will help to make further improvements to the game’s design, resulting in a more effective educational tool in the future. The process and results of this study can guide other researchers when applying learning analytics to evaluate and improve the educational design of serious games, as well as supporting teachers—both during and after the game activity—in applying an evidence-based assessment of the players based on the collected learning analytics. Full article

(This article belongs to the Special Issue Smart Learning Environments)

► Show Figures

Figure 1

25 pages, 11555 KiB

Open AccessArticle

Scalable Data Transformation Models for Physics-Informed Neural Networks (PINNs) in Digital Twin-Enabled Prognostics and Health Management (PHM) Applications

by Atuahene Kwasi Barimah, Ogwo Precious Onu, Octavian Niculita, Andrew Cowell and Don McGlinchey

Computers 2025, 14(4), 121; https://doi.org/10.3390/computers14040121 - 26 Mar 2025

Viewed by 251

Abstract

Digital twin (DT) technology has become a key enabler for prognostics and health management (PHM) in complex industrial systems, yet scaling predictive models for multi-component degradation (MCD) scenarios remains challenging, particularly when transferring insights from predictive models of smaller systems developed with limited [...] Read more.

Digital twin (DT) technology has become a key enabler for prognostics and health management (PHM) in complex industrial systems, yet scaling predictive models for multi-component degradation (MCD) scenarios remains challenging, particularly when transferring insights from predictive models of smaller systems developed with limited data to larger systems. To address this, a physics-informed neural network (PINN) framework that integrates a standardized scaling methodology, enabling scalable DT analytics for MCD prognostics, was developed in this paper. Our approach employs a systematic DevOps workflow that features containerized PINN DT analytics deployed on a Kubernetes cluster for dynamic resource optimization, a real-time DT platform (PTC ThingWorx™), and a custom API for bidirectional data exchange that connects the cluster to the DT platform. A key contribution of this paper is the scalable DT model, which facilitates transfer learning of degradation patterns across heterogeneous hydraulic systems. Three (3) hydraulic system configurations were modeled, analyzing multi-component filter degradation under pump speeds of 700–900 RPM. Trained on limited data from a reference system, the scaled PINN model achieved 88.98% accuracy for initial degradation detection at 900 RPM—outperforming an unscaled baseline of 64.13%—with consistent improvements across various speeds and thresholds. This work advances PHM analytics by reducing costs and development time, providing a scalable framework for cross-system DT deployment. Full article

(This article belongs to the Special Issue Generative Artificial Intelligence and Machine Learning in Industrial Processes and Manufacturing)

► Show Figures

Figure 1

Journal Menu

Journal Browser

Computers, Volume 14, Issue 4 (April 2025) – 38 articles

Further Information

Guidelines

MDPI Initiatives

Follow MDPI