Search Results (158)

 COVID-19 infection is often accompanied by psychological symptoms, which may persist long after the end of the infection (long COVID). The symptoms include fatigue, cognitive impairment, and anxiety. The reason for these long-term effects is currently unclear. Therapeutic approaches have included cognitive rehabilitation therapy, physical activity, and serotonin reuptake inhibitors (SSRIs) if depression co-exists. The neuropsychological evaluation of subjects with suspected cognitive issues is essential for the correct diagnosis. Most of the COVID-19 studies used the Montreal Cognitive Assessment (MoCA) or the Mini Mental State Examination (MMSE). However, MoCA scores can be confusing if not interpreted correctly. For this reason, we have developed an original technique to map cognitive domains and motor performance on various brain areas in COVID-19 patients aiming at improving the follow-up of long-COVID-19 symptoms. To this end, we retrospectively reanalyzed data from a cohort of 40 patients hospitalized for COVID-19 without requiring intubation or hemodialysis. Cognitive function was tested during hospitalization and six months after. Global cognitive function and cognitive domains were retrieved using MoCA tests. Laboratory data were retrieved regarding kidney function, electrolytes, acid–base, blood pressure, TC score, and P/F ratio. The dimensionality of cognitive functions was represented over cortical brain structures using a transformation matrix derived from fMRI data from the literature and the Cerebroviz mapping tool. Memory function was linearly dependent on the P/F ratio. We also used the UMAP method to reduce the dimensionality of the data and represent them in low-dimensional space. Six months after hospitalization, no cases of severe cognitive deficit persisted, and the number of moderate cognitive deficits reduced from 14% to 4%. Most cognitive domains (visuospatial abilities, executive functions, attention, working memory, spatial–temporal orientation) improved over time, except for long-term memory and language skills, which remained reduced or slightly decreased. The Cerebroviz algorithm helps to visualize which brain regions might be involved in the process. Many patients with COVID-19 continue to suffer from a subclinical cognitive deficit, particularly in the memory and language domains. Cerebroviz’s representation of the results provides a new tool for visually representing the data.  Full article

Accurate and robust power quality disturbance (PQD) classification is critical for modern electrical grids, particularly in noisy environments. This study presents a comprehensive comparative evaluation of machine learning (ML) and deep learning (DL) models for automatic PQD identification. The models evaluated include Support Vector Machines (SVM), Decision Trees (DT), Random Forest (RF), k-Nearest Neighbors (kNN), Gradient Boosting (GB), and Dense Neural Networks (DNN). For experimentation, a hybrid dataset, comprising both synthetic and real signals, was used to assess model performance. The robustness of the models was evaluated by systematically introducing Gaussian noise across a wide range of Signal-to-Noise Ratios (SNRs). A central objective was to directly benchmark the practical implementation and performance of these models across two widely used platforms: MATLAB R2024a and Python 3.11. Results show that ML models achieve high accuracies, exceeding $95\%$ at an SNR of 10 dB. DL models exhibited remarkable stability, maintaining $97\%$ accuracy for SNRs above 10 dB. However, their performance degraded significantly at lower SNRs, revealing specific confusion patterns. The analysis underscores the importance of multi-domain feature extraction and adaptive preprocessing for achieving resilient PQD classification. This research provides valuable insights and a practical guide for implementing and optimizing robust PQD classification systems in real-world, noisy scenarios. Full article

This study investigates urban–remote disparities in the science performance of Taiwanese eighth-grade students, particularly in matter-related domains, using an explanatory–sequential mixed-methods design. For the quantitative phase, we applied differential item functioning (DIF) analysis with Mantel–Haenszel statistics and logistic regression to the TIMSS 2019 science assessment, while in the qualitative phase, we employed think-aloud interviews and the repertory grid technique (RGT) with 12 students (6 urban, 6 remote) to explore cognitive structures. The quantitative phase identified 26 items (12.3% of 211) disadvantaging remote students, with DIF most pronounced in constructed-response formats and matter-related domains: “Composition of Matter”, “Physical States and Changes in Matter”, and “Properties of Matter”. The follow-up qualitative analyses revealed fragmented, associative cognitive structures in remote learners, marked by reliance on observable (macroscopic) properties rather than microscopic explanations, terminological confusion, microscopic gaps, and misconceptions, contrasting with urban students’ hierarchical integration. Triangulation suggests that the observed disparities are linked to experiential constraints, potentially accounted for by hindered micro–macro connections. Our findings suggest that resource inequities may play a role in sustaining certain biases, indicating that targeted measures could help to make science education more inclusive. Based on these results, we tentatively outline possible educational interventions to improve equity in science education. Full article

Background/Objectives: Antimicrobial stewardship (AMS) interventions are critical to reducing inappropriate antibiotic prescribing for acute respiratory infections (ARIs) in primary care and mitigating antimicrobial resistance (AMR). While interventions are routinely employed in hospitals, implementation in general practice is nascent. This qualitative study, part of the OPTIMAS-GP project, explored determinants influencing the implementation of evidence-based AMS strategies in Australian general practice. Methods: Using Experience-Based Co-Design, three rounds of online focus groups were conducted with ten healthcare professionals (GPs, pharmacists, microbiologist, practice staff) and ten adult patients who had experienced ARI management in primary care. Participants discussed the feasibility and acceptability of AMS interventions: shared decision-making (SDM) tools, delayed prescribing (DP) and point-of-care testing (PoCT) for C-reactive protein (CRP). Results: Thematic analysis of focus group transcriptions identified four interrelated themes: ‘Patient acceptance and engagement’, ‘Practising within a system’, ‘Prescribing stewardship’, and ‘Diagnostic stewardship’. Patient engagement was dependent upon expectations, trust, and personalised care, while systemic factors such as continuity of care, practice culture, and resource availability influenced implementation. DP was viewed as a pragmatic but potentially confusing strategy, requiring clear patient guidance and interprofessional collaboration. SDM tools were conceptually supported but challenged by time constraints and poor health literacy. PoCT-CRP was cautiously welcomed for selective use, with concerns expressed about workflow integration and overreliance on testing. Findings were mapped to the Capability, Opportunity, Motivation-Behaviour (COM-B) and Theoretical Domains Framework (TDF) to identify behavioural determinants and inform future implementation strategies. Recommendations include co-designing patient-centred AMS tools with clear instructions and red flags, enhancing GP-pharmacist collaboration, and addressing barriers to PoCT integration. Conclusions: These insights highlight the complexity of implementing AMS interventions in general practice and underscore the need for tailored, system-supported approaches to optimise antibiotic use and reduce AMR. Full article

Next-point-of-interest (POI) recommendation aims to model user preferences based on historical information to predict future mobility behavior, which has significant application value in fields such as urban planning, traffic management, and optimizing business decisions. However, existing methods often overlook the differences in location, time, and category information features, fail to fully utilize information from various modalities, and lack effective solutions for addressing users’ incidental behavior. Additionally, existing methods are somewhat lacking in capturing users’ personalized preferences. To address these issues, we propose a new method called Multi-Branch Spectral Network with Contrastive Learning (MBSCL-Net) for next-POI recommendation. We use a multihead attention mechanism to separately capture the distinct features of location, time, and category information, and then fuse the captured features to effectively integrate cross-modal features, avoid feature confusion, and achieve effective modeling of multi-modal information. We propose converting the time-domain information of user check-ins into frequency-domain information through Fourier transformation, directly enhancing the low-frequency signals of users’ periodic behavior and suppressing occasional high-frequency noise, thereby greatly alleviating noise interference caused by the introduction of too much information. Additionally, we introduced contrastive learning loss to distinguish user behavior patterns and better model personalized preferences. Extensive experiments on two real-world datasets demonstrate that MBSCL-Net outperforms state-of-the-art (SOTA) methods. Full article

Predictive maintenance strategies in wind turbine operations have risen in popularity with the growth of renewable electricity demand. The capacity of the strategy to predict system health, especially for the wind turbine gearboxes, is critical in reducing wind turbine operation and maintenance cost. Driven by the emergence of the application of large language models (LLMs) in diverse domains, this work explores the potential of LLMs in the development of wind turbine gearbox prognosis. A comparative analysis is designed to investigate the capability of two state-of-the-art LLMs—GPT-4o and DeepSeek-V3—in proposing machine learning (ML) pipelines to classify gearbox conditions based on a labelled SCADA dataset. The LLMs were prompted with the context of the task and detailed information about the SCADA dataset investigated. The outputs generated by the LLMs were evaluated in terms of pipeline quality and prediction performance using the confusion metric. Baseline ML models were developed and fine-tuned as benchmarks using Python 3.12 libraries. Among the baseline models, the random forest and XGBoost models achieved the highest cross-validated average F1-scores. The results have shown that the ML pipeline proposed by DeepSeek-V3 was significantly better than both GPT-4o and baseline models in terms of data analytical scope and prediction accuracy. Full article

This article provides an in-depth review of the concepts of interpretability and explainability in machine learning, which are two essential pillars for developing transparent, responsible, and trustworthy artificial intelligence (AI) systems. As algorithms become increasingly complex and are deployed in sensitive domains, the need for interpretability has grown. However, the ongoing confusion between interpretability and explainability has hindered the adoption of clear methodological frameworks. To address this conceptual ambiguity, we draw on the formal distinction introduced by Dib, which rigorously separates interpretability from explainability. Based on this foundation, we propose a revised classification of explanatory approaches structured around three complementary axes: intrinsic vs. extrinsic, specific vs. agnostic, and local vs. global. Unlike many existing typologies that are limited to a single dichotomy, our framework provides a unified perspective that facilitates the understanding, comparison, and selection of methods according to their application context. We illustrate these elements through an experiment on the Breast Cancer dataset, where several models are analyzed: some through their intrinsically interpretable characteristics (logistic regression, decision tree) and others using post hoc explainability techniques such as treeinterpreter for random forests. Additionally, the LIME method is applied even to interpretable models to assess the relevance and robustness of the locally generated explanations. This contribution aims to structure the field of explainable AI (XAI) more rigorously, supporting a reasoned, contextualized, and operational use of explanatory methods. Full article

This study proposes a novel, scalable framework for the automated classification and synthesis of survey literature by integrating state-of-the-art Large Language Models (LLMs) with robust ensemble voting techniques. The framework consolidates predictions from three independent models—GPT-4, LLaMA 3.3, and Claude 3—to generate consensus-based classifications, thereby enhancing reliability and mitigating individual model biases. We demonstrate the generalizability of our approach through comprehensive evaluation on two distinct domains: Question Answering (QA) systems and Computer Vision (CV) survey literature, using a dataset of 1154 real papers extracted from arXiv. Comprehensive visual evaluation tools, including distribution charts, heatmaps, confusion matrices, and statistical validation metrics, are employed to rigorously assess model performance and inter-model agreement. The framework incorporates advanced statistical measures, including k-fold cross-validation, Fleiss’ kappa for inter-rater reliability, and chi-square tests for independence to validate classification robustness. Extensive experimental evaluations demonstrate that this ensemble approach achieves superior performance compared to individual models, with accuracy improvements of 10.0% over the best single model on QA literature and 10.9% on CV literature. Furthermore, comprehensive cost–benefit analysis reveals that our automated approach reduces manual literature synthesis time by 95% while maintaining high classification accuracy (F1-score: 0.89 for QA, 0.87 for CV), making it a practical solution for large-scale literature analysis. The methodology effectively uncovers emerging research trends and persistent challenges across domains, providing researchers with powerful tools for continuous literature monitoring and informed decision-making in rapidly evolving scientific fields. Full article

In the manufacturing supply chain, management reports often begin with concise messages that summarize key inventory insights. Traditionally, human analysts manually crafted these summary messages by sifting through complex data—a process that is both time-consuming and prone to inconsistency. In this research study, we present an AI-based system that automatically generates high-quality inventory insight summaries, referred to as “headline messages,” using real-world inventory data. The proposed system leverages lightweight natural language processing (NLP) and machine learning models to achieve accurate and efficient performance. Historical messages are first clustered using a sentence-translation MiniLM model that provides fast semantic embedding. This is used to derive key message categories and define structured input features for this purpose. Then, an explainable and low-complexity classifier trained to predict appropriate headline messages based on current inventory metrics using minimal computational resources. Through empirical experiments with real enterprise data, we demonstrate that this approach can reproduce expert-written headline messages with high accuracy while reducing report generation time from hours to minutes. This study makes three contributions. First, it introduces a lightweight approach that transforms inventory data into concise messages. Second, the proposed approach mitigates confusion by maintaining interpretability and fact-based control, and aligns wording with domain-specific terminology. Furthermore, it reports an industrial validation and deployment case study, demonstrating that the system can be integrated with enterprise data pipelines to generate large-scale weekly reports. These results demonstrate the application and technological innovation of combining small-scale language models with interpretable machine learning to provide insights. Full article

With the popularization of social media, information related to public health events has seen explosive growth online, making it essential to accurately identify informative tweets with decision-making and management value for public health emergency response and risk monitoring. However, existing methods often suffer performance degradation during cross-event transfer due to differences in data distribution, and research specifically targeting public health events remains limited. To address this, we propose the Label Confusion Domain Adversarial Network (LCDAN), which innovatively integrates label confusion with domain adaptation to enhance the detection of informative tweets across different public health events. First, LCDAN employs an adversarial domain adaptation model to learn cross-domain feature representation. Second, it dynamically evaluates the importance of different source domain samples to the target domain through label confusion to optimize the migration effect. Experiments were conducted on datasets related to COVID-19, Ebola disease, and Middle East Respiratory Syndrome public health events. The results demonstrate that LCDAN significantly outperforms existing methods across all tasks. This research provides an effective tool for information detection during public health emergencies, with substantial theoretical and practical implications. Full article

Jute, a vital lignocellulosic fiber crop with substantial industrial and ecological relevance, continues to suffer considerable yield and quality degradation due to pervasive foliar pathologies. Traditional diagnostic modalities reliant on manual field inspections are inherently constrained by subjectivity, diagnostic latency, and inadequate scalability across geographically distributed agrarian systems. To transcend these limitations, we propose DERIENet, a robust and scalable classification approach within a deep ensemble learning framework. It is meticulously engineered by integrating three high-performing convolutional neural networks—ResNet50, InceptionV3, and EfficientNetB0—along with regularization, batch normalization, and dropout strategies, to accurately classify jute leaf diseases such as Cercospora Leaf Spot, Golden Mosaic Virus, and healthy leaves. A key methodological contribution is the design of a novel augmentation pipeline, termed Geometric Localized Occlusion and Adaptive Rescaling (GLOAR), which dynamically modulates photometric and geometric distortions based on image entropy and luminance to synthetically upscale a limited dataset (920 images) into a significantly enriched and diverse dataset of 7800 samples, thereby mitigating overfitting and enhancing domain generalizability. Empirical evaluation, utilizing a comprehensive set of performance metrics—accuracy, precision, recall, F1-score, confusion matrices, and ROC curves—demonstrates that DERIENet achieves a state-of-the-art classification accuracy of 99.89%, with macro-averaged and weighted average precision, recall, and F1-score uniformly at 99.89%, and an AUC of 1.0 across all disease categories. The reliability of the model is validated by the confusion matrix, which shows that 899 out of 900 test images were correctly identified and that there was only one misclassification. Comparative evaluations of the various ensemble baselines, such as DenseNet201, MobileNetV2, and VGG16, and individual base learners demonstrate that DERIENet performs noticeably superior to all baseline models. It provides a highly interpretable, deployment-ready, and computationally efficient architecture that is ideal for integrating into edge or mobile platforms to facilitate in situ, real-time disease diagnostics in precision agriculture. Full article

Background: As intelligent systems increasingly operate in high-risk environments, understanding how they perceive and respond to hazards is critical for ensuring safety. Methods: In this study, we conduct a comparative analysis of 60 real-world accident reports, 30 from process control systems (PCSs) and 30 from autonomous vehicles (AVs), to examine differences in risk triggers, perception paradigms, and interaction failures between humans and artificial intelligence (AI). Results: Our findings reveal that PCS risks are predominantly internal to the system and detectable through deterministic, rule-based mechanisms, whereas AVs’ risks are externally driven and managed via probabilistic, multi-modal sensor fusion. More importantly, despite these architectural differences, both domains exhibit recurring human–AI interaction failures, including over-reliance on automation, mode confusion, and delayed intervention. In the case of PCSs, these failures are historically tied to human–automation interaction; this article extrapolates these patterns to anticipate potential human–AI interaction challenges as AI adaptation increases. Conclusions: This study highlights the need for a hybrid risk perception framework and improved human-centered design to enhance situational awareness and responsiveness. While AI has not yet been implemented in PCS incident studies, this work interprets human–automation failures in these cases as indicative of potential challenges in human–AI interaction that may arise in future AI-integrated process systems. Implications extend to developing safer intelligent systems across industrial and transportation sectors. Full article

Gear condition monitoring is predominantly executed through the utilization of acceleration sensors positioned on the housing. However, recent advancements have identified measuring the instantaneous angular speed as a compelling alternative as it shortens the transmission path and therefore provides high-quality rotational angle information that can be used to increase damage prediction accuracy, particularly under transient operating conditions. Additionally, there are a variety of methodologies for integrating sensors into gears, which underscores the necessity for high-quality condition data. However, it should be noted that a significant amount of effort is required to successfully integrate these sensors into the rotating system. This publication uses a gear wheel sensor that employs the gear itself as a material measure to acquire rotational angle data and to deduce the damage condition. A magnetoresistive sensor is integrated into the gearbox housing radially facing a ferromagnetic gear and measures the rotational angle by the gear teeth. Various artificial tooth flank damages are applied to the pinion. The rotational angle is measured with the gear sensor, and the damage state is classified with a random forest classifier using established evaluations in the time and frequency domains. The tests are conducted under stationary operating conditions at an array of speed and torque levels. Additionally, they are performed under transient operating conditions, employing speed ramps at constant torque. The results of the classification are evaluated by means of classification accuracy and confusion matrices and compared with those obtained via a classic encoder at the pinion shaft and an acceleration sensor at the gearbox housing. Full article

This paper presents a real-time system for the detection and classification of facial micro-expressions, evaluated on the CASME II dataset. Micro-expressions are brief and subtle indicators of genuine emotions, posing significant challenges for automatic recognition due to their low intensity, short duration, and inter-subject variability. To address these challenges, the proposed system integrates advanced computer vision techniques, rule-based classification grounded in the Facial Action Coding System, and artificial intelligence components. The architecture employs MediaPipe for facial landmark tracking and action unit extraction, expert rules to resolve common emotional confusions, and deep learning modules for optimized classification. Experimental validation demonstrated a classification accuracy of 93.30% on CASME II, highlighting the effectiveness of the hybrid design. The system also incorporates mechanisms for amplifying weak signals and adapting to new subjects through continuous knowledge updates. These results confirm the advantages of combining domain expertise with AI-driven reasoning to improve micro-expression recognition. The proposed methodology has practical implications for various fields, including clinical psychology, security, marketing, and human-computer interaction, where the accurate interpretation of emotional micro-signals is essential. Full article

This study investigates a multi-scenario combustion state diagnosis system based on acoustic feature extraction techniques. In this study, the voiceprint technology is applied to combustion condition monitoring for the first time, and an integrated approach for monitoring and diagnosis is proposed by combining multiple acoustic features, such as acoustic pattern features, step index P, and frequency-domain monitoring. In this study, a premixed hydrogen combustion test bed was built to simulate common combustion faults, and the corresponding acoustic features were collected and extracted. In this study, step index P and acoustic features are used for parallel diagnostic analysis, and CNN, ANN, and BP models are used to train the four states of flameout, flameback, thermoacoustic oscillation, and stable combustion, and the training diagnostic performance of each model is compared and analyzed using a confusion matrix. It is found that CNN has the strongest classification ability, can accurately distinguish the four states, has the lowest misclassification rate, has very strong generalization ability, and has a diagnostic accuracy of 93.49%. The classification accuracy of ANN is not as good as that of CNN, and there are local fluctuations during the training process. The BP neural network has a slower convergence speed and a high error rate in recognizing the flameback and thermoacoustic oscillations. In summary, the combustion state diagnosis system based on CNN model combined with acoustic features has optimal performance, and the combination of step index P and frequency-domain monitoring in the flameback diagnosis can improve the accuracy of combustion state identification and safety control level, which provides an important theoretical basis and practical reference in the field of combustion state diagnosis and is of profound significance to ensure the safe and efficient operation of the combustion process. Full article