Search Results (16,782)

Background/Objectives: Alzheimer’s disease (AD) is a progressive neurodegenerative disorder that affects the daily lives of older adults, impacting their cognitive abilities as well as speech and language communication. Early detection is crucial, as it enables timely intervention and helps improve the quality of life for those affected. While large language models (LLMs) have shown promise from spontaneous speech, most studies are unimodal and miss complementary signals across modalities. Methods: We present an LLM-powered multimodal cross-attention framework that integrates lexical (text), acoustic (speech), and visual (image) information for dementia detection using the ADReSSo 2021 picture-description dataset. Within this framework, text data are encoded using the ModernBERT, audio features are extracted using the wav2vec 2.0-base-960, and the Cookie Theft image is represented through the CLIP ViT-L/14. These embeddings are linearly projected to a shared space and then combined via Transformer-based cross-attention, yielding a fused vector for AD detection. Results: Our results show that the trimodal model achieved the best overall performance when paired with an SVC classifier, reaching an accuracy of 0.8732 and an F1 score of 0.8571, surpassing both the top-performing unimodal and bimodal configurations. For interpretability, a sensitivity analysis of modality contributions reveals that text plays the primary role, audio provides complementary improvements, and image offers modest yet stabilizing contextual support. Conclusions: These results highlight that the method of multimodal embedding fusion significantly influences performance: a cross-attention block achieves an effective balance between accuracy and simplicity, producing integrated representations that align well with interpretable downstream classifiers. Full article

Background: Hearing loss affects over 466 million individuals globally and is recognized as a major risk factor for Alzheimer’s disease, yet treatment personalization remains limited due to the complexity and diversity of underlying causes. Current diagnostic and therapeutic approaches lack standardized methods to accurately predict the most appropriate intervention for individual patients. The integration of medical ontologies with machine learning offers a promising solution for enhancing diagnostic accuracy and treatment personalization. Aim: Our study aimed to (i) develop a Systematized Nomenclature of Medicine—Clinical Terms (SNOMED CT)-aligned clinical ontology for hearing loss using Semantic Web Rule Language for automated reasoning; (ii) implement a Random Forest classifier trained on ontology-enriched patient data to classify hearing loss types (conductive, sensorineural, mixed, or normal); and (iii) predict optimal personalized treatments based on laterality, severity, audiometric thresholds, and medical history using real-world patient data. Methods: We developed a task ontology using Protégé 5.6.3 with Web Ontology Language (OWL), integrated SNOMED CT terminology alignment, and implemented Semantic Web Rule Language rules executed by the Pellet 2.2.0 reasoner. The framework was trained and evaluated on 3723 adult patients from the 2015–2016 National Health and Nutrition Examination Survey (NHANES) dataset with complete audiometric and clinical data. Random Forest models were developed using an 80–20 train-test split with stratified sampling and five-fold cross-validation. Performance was compared between K-Means clustering-based labeling and ontology-based semantic inference using accuracy, precision, recall, F1-score, and log loss metrics. Results: The ontology successfully generated semantic labels for all 3723 patients, enabling precise classification of hearing loss types, severity levels, and laterality. The Random Forest model with K-Means clustering achieved a test accuracy of 90.2% with a log loss of 0.2766 and a cross-validation mean accuracy of 91.22% (standard deviation 1.2%). Integration of ontology-based semantic enrichment significantly improved performance, achieving a test accuracy of 92.48% with a cross-validation mean accuracy of 92.80% (standard deviation 0.9%). F1-scores improved across all classes, with mixed hearing loss showing a notable increase from 0.86 to 0.92. Feature importance analysis identified audiometric thresholds, ontology-derived severity labels, and medical history as top predictors, enhancing clinical interpretability. Conclusions: This study demonstrates that combining SNOMED CT-aligned ontology with Random Forest classification achieves superior diagnostic accuracy and enables personalized treatment recommendations for hearing loss. The hybrid framework provides clinically interpretable decision support while ensuring semantic interoperability with electronic health records. Multi-institutional validation studies are necessary to assess generalizability across diverse populations before clinical deployment. Full article

This paper examines the short-run transmission of monetary policy to bank credit granted to the non-financial corporate sector in Morocco, a bank-based emerging economy. Using monthly macro-financial data over the period of 2014–2024, the study estimates a reduced-form VAR model to analyze the dynamic interactions between the policy rate, bank credit, banks’ holdings of sovereign securities, credit risk indicators, and short-term market spreads. Impulse response functions and forecast error variance decompositions indicate that a one-standard-deviation monetary policy shock is associated with a small and short-lived response of non-financial corporations bank credit at a monthly horizon, accounting for only a limited share of its forecast error variance, while the same shock is more strongly reflected in market spreads and banks’ balance-sheet reallocations toward sovereign assets, alongside temporary movements in credit risk indicators. Overall, these results are consistent with a reduced-form transmission pattern in which monetary policy appears to affect bank credit primarily through indirect financial channels related to risk perception, portfolio reallocation, and balance-sheet management, rather than through immediate changes in aggregate credit volumes. This interpretation is conditional on the VAR specification and short-run horizon considered, and suggests an attenuation of the interest rate–credit channel in a bank-dominated emerging economy, rather than evidence of a structural breakdown of monetary transmission. Full article

The current research examines the impact of climate change disclosure (CCD) on firm value (FV) of Egyptian listed non-financial companies. The current research also investigates the moderating role of board gender diversity (BGD). The study sample incorporates Egyptian non-financial companies indexed in EGX 100 whose reports were available from 2018 to 2023. The final sample comprises 82 companies with 492 observations. Statistical analysis was conducted using a POLS and Fixed Effects Model, GMM, and the 2SLS method to address potential endogeneity and dynamic panel concerns. The results revealed a positive and significant impact of CCD on FV. Furthermore, BGD had a positive and significant moderating impact as BGD enhanced the relationship between CCD and FV. Moreover, the critical mass (CM) analysis of female representation revealed that the number of females on the board significantly moderates the CCD-FV relationship; as CM increases, the effect on the CCD-FV relationship becomes stronger. Although advanced panel techniques and instrumental variable approaches are used to mitigate identification concerns, the results should be interpreted in light of the observational nature of the data and the reliance on disclosure-based proxies. These findings are significant for governments, regulators, investors, and company leaders because the moderating role of BGD demonstrates how board governance affects firm value, particularly in emerging markets. This research adds to the academic discussion by emphasizing the beneficial effects of both BGD and CCD on FV, with a particular focus on developing economies. Full article

Dental conditions, including caries, periodontal disease, plaque accumulation, malocclusion, and oral mucosal abnormalities, remain highly prevalent worldwide. Early detection is crucial for preventing disease progression, simplifying treatment, and improving patient outcomes. Conventional diagnostic methods rely on subjective visual and tactile examinations, which are often inconsistent. Recent advances in deep learning (DL), particularly convolutional neural networks and vision transformers, enable automated, accurate detection of dental diseases from intraoral images captured via smartphones or dedicated imaging devices. DL-driven systems facilitate cost-effective virtual consultations, community screenings, and remote oral health monitoring. This narrative review was conducted following a structured search of PubMed, Scopus, Web of Science, Embase, and Google Scholar (October 2020–October 2025), which identified 74 eligible studies on intraoral photographic imaging-based DL systems, encompassing caries, gingival inflammation, plaque, malocclusion, and soft-tissue lesions. Most studies focused on caries, plaque, and periodontal disease using CNN and U-Net-based models, often reporting accuracies above 85% but with substantial performance drops in external validation. Despite promising results, clinical integration remains limited by challenges such as class imbalance, limited external validation, heterogeneous imaging protocols, and insufficient model interpretability. Emerging approaches, including self-supervised and federated learning, explainable artificial intelligence, multimodal data fusion, and smartphone-based diagnostics, offer potential solutions. Standardized imaging workflows, high-quality annotations, and robust clinical trials are essential to translate DL-based dental diagnostic systems into real-world practice. This narrative review aims to guide the development of reliable, equitable, and clinically deployable DL solutions for oral health assessment. Full article

The notion of complex thinking has become established as an essential competency for understanding multidimensional social phenomena and for engaging with democratic processes in diverse contexts. This study examined within-individual changes associated with participation in an educational intervention aimed at developing complex thinking among university students in Bolivia, Chile, Colombia, and Mexico, and explored their implications for intercultural citizenship education. A quasi-experimental pretest–posttest design was employed, drawing on panel data and fixed-effects regression models to estimate intraindividual variation over time, complemented by an exploratory differential analysis by sex. The findings revealed statistically significant within-individual changes across the four evaluated subdimensions, as well as differentiated patterns by sex, with women showing higher relative changes in critical and innovative thinking and men showing higher relative changes in scientific reasoning. Interpreted in dialogue with existing literature, these observed changes in complex thinking are consistent with theoretical frameworks that conceptualize such competencies as relevant cognitive foundations for intercultural citizenship. Overall, the study provides empirically grounded insights into the role of complexity-oriented learning experiences in higher education in Latin America and outlines considerations for the design of more context-sensitive and equity-oriented educational initiatives. Full article

We study Maximum Entropy density estimation on continuous domains under finitely many moment constraints, formulated as the minimization of the Kullback–Leibler divergence with respect to a reference measure. To model uncertainty in empirical moments, constraints are relaxed through convex penalty functions, leading to an infinite-dimensional convex optimization problem over probability densities. The main contribution of this work is a rigorous convex-analytic treatment of such relaxed Maximum Entropy problems in a functional setting, without discretization or smoothness assumptions on the density. Using convex integral functionals and an extension of Fenchel duality, we show that, under mild and explicit qualification conditions, the infinite-dimensional primal problem admits a dual formulation involving only finitely many variables. This reduction can be interpreted as a continuous-domain instance of partially finite convex programming. The resulting dual problem yields explicit primal–dual optimality conditions and characterizes Maximum Entropy solutions in exponential form. The proposed framework unifies exact and relaxed moment constraints, including box and quadratic relaxations, within a single variational formulation, and provides a mathematically sound foundation for relaxed Maximum Entropy methods previously studied mainly in finite or discrete settings. A brief numerical illustration demonstrates the practical tractability of the approach. Full article

Reliable ultrasonic inspection of welded structures requires a quantitative understanding of how defect morphology and depth influence detectability. In this study, a simulation-based signal-response Probability of Detection (POD) framework is developed to investigate ultrasonic wave interaction with representative planar and volumetric weld defects. Two-dimensional finite-element shear-wave simulations were conducted to model wave propagation and scattering from planar flaws (toe and root cracks) and volumetric flaws (porosity) across defined inspection depth zones. Peak terminal voltage was used as a continuous response metric for regression-based POD analysis. The results demonstrate that defect morphology dominates the influence on ultrasonic detectability. Planar defects produced systematically higher signal responses than volumetric defects of comparable size, resulting in lower characteristic detection limits. The estimated a₉₀ value for planar flaws was 2.96 mm, compared to 5.64 mm for volumetric flaws under identical threshold conditions. Depth-dependent analyses further revealed morphology-specific behavior: planar defects exhibited consistently high detection probabilities across depth zones (POD > 0.98), whereas volumetric defects showed a reduction in detectability with depth, with POD decreasing from approximately 0.32 in shallow zones to 0.16 in deeper regions. The resulting POD trends are interpreted as comparative, trend-based indicators of morphology and depth-dependent ultrasonic detectability under idealized inspection conditions. These findings quantitatively demonstrate how ultrasonic detectability is governed by wave-defect interaction mechanisms associated with defect morphology and inspection depth. Full article

Real-time lithology identification while drilling is widely applied in oil and gas exploration, development drilling, geo-steering, unconventional resource extraction, well logging, and environmental monitoring, enhancing efficiency and accuracy in subsurface operations. This study investigates the frequency characteristics of rock-drilling sounds generated during drilling operations and explores their potential for real-time lithology identification. Experiments were conducted using 8 mm and 14 mm drill bits at both high and low rotational speeds on four types of rock samples: sandstone, limestone, granite, and shaly sandstone. Sound signals were recorded both within the rock and in air using high-fidelity sensors. The results reveal distinct frequency patterns for each rock type, with sandstone exhibiting dominant low-frequency energy, limestone and granite showing broader frequency bands with strong high-frequency components, and shaly sandstone displaying a mix of low- and high-frequency energy. Quadratic polynomial regression models between the Vp or Vs and the peak frequencies of the four distinct rock samples are built, and the corresponding coefficients of determination are 0.9878 and 0.9799. The study also demonstrates that drilling parameters, such as drill bit diameter and revolutions per minute (RPM), significantly influence the frequency distribution of rock-drilling sounds, with larger drill bits and higher RPMs producing broader frequency bands and stronger high-frequency energy. Comparisons between in-rock and in-air recordings show that the latter captures richer high-frequency information, though the overall trends remain consistent. These findings provide an experimental foundation for using rock-breaking sounds as a potential tool for lithology identification during drilling operations. The study highlights the importance of considering rock heterogeneity and drilling conditions when interpreting acoustic data and suggests future work to validate the method in field conditions and integrate advanced data processing techniques. Full article

Detecting surface potsherds using low-altitude remote sensing is challenging due to severe class imbalance and limited training data. This study develops and validates a semi-automatic detection methodology that adapts threshold-optimized boosting classifiers (AdaBoost, XGBoost) to maximize ceramic detection recall under extreme class imbalance in the Western Megaris archeological landscape, Greece. Models were trained on only 15% of the available data to simulate realistic field conditions. Evaluation emphasized recall-oriented metrics (precision, recall, F1-score, AUC) for the minority class, addressing the accuracy paradox where high overall accuracy masks poor rare-class performance. Threshold optimization enabled AdaBoost and XGBoost to achieve substantially improved recall compared to baseline methods, with detection-to-ground-truth ratios of 2.5 and 3.2, respectively, reflecting deliberate prioritization of recall over precision for exploratory survey purposes. The results demonstrate that this methodological framework provides archeologically interpretable screening tools for identifying high-probability ceramic locations, supporting more efficient field survey design and heritage documentation workflows in Mediterranean landscapes. Full article

Heritage Building Information Modelling (HBIM) has emerged as a powerful methodology for documenting, analysing, and managing historic buildings. However, its pedagogical potential in teaching adaptive reuse and heritage-sensitive design remains underexplored, particularly in postgraduate architectural education. This study evaluates a pedagogical HBIM framework implemented in a master’s-level course, where students applied HBIM methodologies to propose adaptive reuse interventions for a historic building in Jeddah Historic District, Saudi Arabia. Student design projects were analysed to assess how HBIM informed documentation accuracy, heritage value interpretation, and design decision-making. In addition, a retrospective questionnaire was administered to former students to evaluate the long-term educational effectiveness of the HBIM-based methodology, focusing on learning quality, design comprehension, and professional preparedness. The results indicate that HBIM significantly enhanced students’ understanding of historic fabric, improved their ability to propose context-sensitive reuse strategies, and supported more informed and evidence-based design decisions. Survey findings further demonstrate the high perceived value of HBIM in architectural education, particularly in linking theoretical knowledge with real-world heritage challenges. This research contributes a validated educational framework for integrating HBIM into adaptive reuse curricula and provides evidence-based insights applicable to heritage education and professional practice. Full article

Ultrasonic guided wave sensors are widely employed for structural health monitoring; yet, their signal interpretation reliability is frequently compromised in extreme environments where coupled temperature and stress induce significant nonlinear drifts in dispersion characteristics. To overcome the computational bottleneck of conventional numerical methods that hinders real-time sensor calibration, this paper proposes a Dispersion Multi-Head Attention Network (D-MHAN) that directly maps eleven raw physical parameters to full-band dispersion responses. By adopting a non-normalized input strategy to internalize underlying physical laws, the model enables robust out-of-distribution extrapolation even when material properties exceed the training manifold. It was validated against a high-fidelity dataset spanning temperatures from −250 °C to 100 °C and stresses from 0 to 150 MPa generated by the Semi-Analytical Finite Element (SAFE) method. The proposed D-MHAN achieves a Pearson correlation coefficient of 0.9999 and provides computational speeds approximately 30 and 168 times faster than SAFE and the Wave Finite Element Method (WFEM). The model’s practical utility is further corroborated by cryogenic experiments on an aerospace storage tank. This work establishes a critical foundation for real-time parameter sensitivity analysis and environmental effect compensation in practical ultrasonic sensing applications. Full article

The strengthening of reinforced concrete (RC) beams requires repair systems that can enhance strength, stiffness, and energy dissipation without significantly increasing self-weight or compromising durability. This study explores the structural response of RC beams strengthened using an integrated shear–flexure system combining near-surface-mounted carbon fiber-reinforced polymer (NSM-CFRP) ropes and steel-reinforced geopolymer overlays in the compression zone. Monotonic three-point bending tests were performed on two RC beam specimens, one unstrengthened control and one strengthened beam, to obtain preliminary observations of load–deflection behavior, stiffness, ductility, and energy absorption. The strengthened specimen exhibited increases in ultimate load (28.6%), stiffness (13.6%), and energy absorption (7.65%) relative to the control beam, suggesting the potential for effective composite action between the CFRP ropes and geopolymer material. A three-dimensional nonlinear finite element model was developed using ATENA to support interpretation of the experimental response, incorporating detailed constitutive models for concrete, steel reinforcement, and CFRP ropes. The numerical predictions showed reasonable agreement with the experimental results. Within the limitations of the test matrix, the results indicate that the proposed dual strengthening system may offer a viable and sustainable approach for enhancing the shear–flexural performance of RC beams. Full article

Background: The translation of antimicrobial resistance (AMR) data and evidence into policy remains limited in many low- and middle-income countries (LMICs) in Asia and Africa, despite expanded investments being made in AMR surveillance and research. This is due to fragmented governance, weak knowledge translation capacity, and insufficient Results: Participant surveys showed improved confidence and capability to synthesize, interpret, and apply AMR evidence to inform policy. The four countries highlighted persistent constraints such as sectoral silos, limited institutional ownership, and gaps in evidence-use systems, but reported enhanced cross-sectoral collaboration and a clearer, replicable process for EBP development. Methodology: During Phase II of the Fleming Fund-resourced Regional AMR Data Analysis for Advocacy, Response, and Policy (RADAAR) project, a structured, hybrid, evidence-to-policy (E2P) capacity-strengthening model was implemented in Bhutan, Ghana, Kenya, and Lao People’s Democratic Republic, combining online and in-person training, targeted mentorship, and policy engagement. Each country developed a country-specific evidence brief for policy (EBP) and conducted policy dialogues to facilitate stronger decision maker involvement. Conclusions: RADAAR’s approach strengthened the foundational capacity for evidence-informed policymaking and demonstrated the value of institutionalized knowledge translation mechanisms. Sustained investment in E2P systems is essential to bridge the AMR E2P gap and ultimately support AMR prevention and control. Full article

Large language models (LLMs) are increasingly deployed as information systems that evolve over time, where managing internal knowledge—acquisition, retention, and removal—becomes essential. In practice, these processes are primarily realized through continual learning and machine unlearning mechanisms. Despite this, these two mechanisms are often studied in isolation, limiting both interpretability and controllability. In this work, we present a parameter-efficient knowledge management framework where continual learning and machine unlearning—despite employing distinct task-specific objectives—are integrated through a shared retention-controlled parameter evolution mechanism. We ground these structural constraints in a drift-aware design principle: under a model smoothness assumption, we establish a formal upper bound showing that Kullback–Leibler (KL) divergence on retained knowledge is controlled by the magnitude and direction of parameter updates, providing a principled rationale for combining Low-Rank Adaptation (LoRA) freezing, sparse masking, and orthogonal gradient projection into a unified constraint system. Experiments on the Task of Fictitious Unlearning (TOFU) benchmark and real-world benchmarks demonstrate effective knowledge acquisition, selective removal, and robust retention across sequential tasks with strong overall performance and stability. This work provides a practical parameter-efficient recipe and a drift-aware design principle validated on controlled interleaved benchmarks, offering insights toward reliable knowledge management in evolving deployment scenarios. Full article