Search Results (1,820)

Few-shot cross-domain bearing fault diagnosis is challenging because labeled fault samples are limited and signals collected by vibration sensors under different operating conditions often show significant distribution shifts. To improve bearing fault identification under limited-sample and cross-condition scenarios, this paper proposes a wavelet convolution (WC) and Brownian distance covariance (BDC)-based metric meta-learning framework, termed WCBDC. In this framework, the WC is inserted into the feature extraction process to capture multiscale time–frequency information from vibration signals. The BDC is then applied to model nonlinear inter-channel statistical dependencies and improve the discriminability of fault embeddings. The obtained feature embeddings are further organized within a prototypical-network-based classifier, in which category prototypes are estimated from support samples and query instances are assigned by prototype-distance comparison. The proposed method is evaluated on the Paderborn University (PU) and Beijing Jiaotong University (BJTU) bearing datasets under both 5-way 5-shot and 5-way 1-shot scenarios. On the PU dataset, WCBDC reaches average accuracies of 92.19% and 84.13%, while the corresponding results on the BJTU dataset are 77.24% and 62.57%. These results exceed those of representative meta-learning baselines, demonstrating that WCBDC provides improved diagnostic performance for sensor-based bearing fault recognition when labeled samples are scarce and operating conditions vary. Full article

Audio deepfake detection systems often achieve excellent internal validation performance but fail to generalize under real-world inference conditions involving synthetic speech generated with previously unseen AI tools. To address this limitation, this work proposes the Spectral Multi-Representation Fusion (SMRF) framework, which integrates multiple spectral representations and decision-level fusion strategies to improve robustness under cross-domain conditions. Additionally, a Stability-Aware Multi-Metric Selection (SAMMS) strategy is introduced to select architectures by jointly considering predictive performance and cross-representation stability. The proposed framework was evaluated using four spectral representations (log-magnitude spectrogram (LOG), Mel spectrogram (MEL), Discrete Wavelet Transform (DWT), and Constant-Q Transform (CQT)) combined with multiple convolutional architectures and complementary voting strategies. The experiments revealed that isolated models exhibiting validation metrics above 95% may still produce very poor synthetic-audio detection rates during external inference (even lower than 10%). In contrast, fusion-based strategies substantially improved robustness by exploiting complementary synthetic evidence across spectral domains. The results also demonstrated that both the voting strategy and the SAMMS stability parameter $\lambda$ strongly affect the final behavior of the system. In particular, hybrid fusion using One-Hard Voting with two architectures selected using $\lambda \ge 0.25$ achieved the best balance between synthetic-audio detection and real-audio preservation, outperforming individual models under cross-domain inference conditions, with detection rates close to 75% for both synthetic and real audio. These findings suggest that stability-aware fusion strategies constitute a promising direction for improving robustness in realistic audio deepfake detection scenarios. Full article

Variations in mechanical load constitute a dominant source of domain shift in data-driven fault diagnosis of rotating machinery, causing models trained under one operating condition to degrade sharply when deployed under another. This work addresses the problem through unsupervised domain adaptation (UDA)—which transfers diagnostic knowledge from a labeled source condition to an unlabeled target condition by aligning their feature distributions—and introduces Padé Approximant Neural Networks (PadéNets) as compact yet highly expressive feature extractors. One-dimensional PadéNet encoders are embedded into three established adaptation frameworks—Deep CORAL, Domain-Adversarial Neural Networks (DANNs), and Conditional Domain-Adversarial Networks (CDANs)—to learn load-invariant representations without any labeled target data. On the Case Western Reserve University benchmark, where the models operate directly on raw time-domain vibration signals, replacing conventional convolutional encoders with PadéNets consistently improves cross-load diagnostic accuracy, reaching up to 99.28% average target-domain accuracy at a low parameter count. To assess generalization to a more demanding setting, the CDAN–PadéNet configuration is further evaluated on frequency-domain representations of the Paderborn University dataset, where domain shift arises from simultaneous variation of load torque and radial force on bearings with real accelerated-lifetime damage, attaining 99.84% average accuracy across six cross-condition transfer tasks while requiring fewer parameters than competing methods. These results establish PadéNet-enhanced UDA as an accurate, broadly applicable approach for robust bearing fault diagnosis under varying operating conditions, with a reduced parameter count suited to resource-constrained embedded platforms. Full article

Accurate carbon price forecasting is essential for emission trading risk management and low-carbon investment decisions. In existing decomposition-prediction frameworks, secondary decomposition targets are typically selected based on statistical complexity rather than domain-informed causality, and standard Long Short-Term Memory (LSTM)-Transformer architectures discard the cell state that encodes long-term temporal memory. These limitations are particularly pronounced where energy-driven causal structures and regime-switching volatility coexist. This study proposes Causal State-driven Network (CaStNet), an intelligent forecasting framework with two core innovations. A Policy-Causality-guided Residual Secondary Decomposition (PCRSD) module replaces entropy-based criteria with Granger causality to select intrinsic mode functions (IMFs) exhibiting significant energy-carbon causal linkages for targeted variational mode decomposition (VMD). A Cell-State-Driven Dual-function Attention (CSDA) mechanism repurposes the LSTM cell state for simultaneously injecting long-term memory into the Transformer and employing the cell-state differential velocity as a volatility proxy to adaptively regulate Top-k attention sparsity. The Artificial Lemming Algorithm (ALA) globally co-optimizes decomposition dimensions and attention boundaries. A Shapley Additive exPlanations (SHAP)–Local Interpretable Model-agnostic Explanations (LIME) interpretability analysis reveals horizon-dependent driver transitions from short-term autoregressive momentum to long-term energy fundamentals, uncovering threshold nonlinearities in energy-carbon transmission channels. Validation on the Shanghai market (2013–2025) achieves point-forecast RMSE = 0.8326 and R² = 0.9777, outperforming all twelve benchmark models. Cross-market testing on the Hubei market yields R² = 0.9487, and expanding-window five-fold cross-validation on the Shanghai dataset yields mean R² = 0.9704, jointly confirming generalization robustness. Full article

Data has become a strategic production factor, but the institutional logic underlying data’s tradability, priceability, and governability remains insufficiently theorized. In response, this study develops a coevolutionary framework that connects conventional factor market theory with digital political economy, platform theory, and comparative institutional analysis. This study adopts a conceptual–analytical research design, integrating three research methods: theory synthesis, comparative institutional analysis, and policy-process interpretation. Through theoretical synthesis, institutional comparison, and policy-process interpretation, it analyzes the conditions under which data circulation becomes feasible, lawful, and economically sustainable. In addition, by combining transaction data, exchange listings, property rights registrations, network indicators, and regional policy variations, it formulates testable propositions and an empirical agenda. The study finds that data factor markets do not emerge automatically with digitalization; their formation requires three mutually reinforcing conditions: technologically reducing search, verification, privacy protection, and contract enforcement costs; institutionally realizing a modular definition of rights and establishing compliance boundaries; and market demand from firms, public agencies, and research organizations generating use-case-specific value. Meanwhile, this study revises the three-stage model of market evolution as a contingent and testable pathway—from administrative pilot allocation, through hybrid state–market professionalization, to ecosystem-based cross-domain circulation. It also clarifies a closed-loop dynamic mechanism consisting of external shocks, internal strategic feedback, and adaptive governance, which jointly shapes market boundaries, pricing rules, and competition patterns. Full article

This paper proposes a Machine Learning (ML) framework designed to attribute liability between drivers and pedestrians in traffic crashes. This study applies classification algorithms and interpretability techniques to analyze judicial rulings related to pedestrian crashes in Badajoz, Spain, from 2015 to 2024. The primary objective is to identify recurring crash patterns and determine liability levels for the parties involved. Several classification algorithms were evaluated, including Support Vector Machines (SVM), Neural Network (NN), Decision Trees (DT), Boosted Trees (BT), Naïve Bayes (NB), Random Forest (RF), K-Nearest Neighbors (K-NN), and Logistic Regression (LR). Among them, the quadratic-kernel SVM achieved the highest overall performance. To address the severe class imbalance of the data, stratified k-fold cross-validation and the Synthetic Minority Oversampling Technique (SMOTE) were applied to enhance the robustness and generalization capability of the model. A multiclass classification framework was implemented, and SHAP (SHapley Additive exPlanations) was integrated to improve interpretability by quantifying the contribution of each feature to the model’s predictions. The analysis identified critical factors that play a significant role in determining liability outcomes: driver license status, crash location, lighting conditions, reaction time, and the presence of drugs or alcohol. This research aims to contribute to the legal domain. While most existing studies have focused on predicting injury severity, few have addressed liability attribution. This is a multifactorial task that requires a comprehensive analysis of judicial decisions. The results demonstrate that machine learning-driven liability attribution can support judicial decision-making and provide valuable insights for the development of proactive urban traffic safety strategies. Full article

We propose a mathematical framework for domain generalization in medical image segmentation built on dual-tree complex wavelet transform (DTCWT) and variational information theory. The core premise is that, under adequate spatial normalization and acquisition-style shifts, DTCWT phase components are more closely associated with anatomical structure, whereas amplitude components are more sensitive to domain-specific intensity and style variations. We formulate this as a local phase–magnitude complementarity premise and construct an information bottleneck that operates on structured subband representations. The framework provides several key theoretical results under explicit structural assumptions: an information bound showing when DTCWT amplitude subbands better isolate domain-related information than global Fourier representations; a variational information bottleneck encoder that compresses domain-specific amplitude information into low-dimensional latent codes; a triple constraint mechanism (domain supervision, KL compression, and orthogonality) that controls domain–task information leakage; and a predictive feature modulation scheme with $O\left(1\right)$ spatial complexity. We further analyze test-time adaptation via calibrated uncertainty, deriving a sufficient condition under which a two-pass inference strategy reduces the expected generalization gap. Finally, we include illustrative public-dataset checks on FeTS 2022 and BraTS 2023 to test the central phase–amplitude premise and the feasibility of DTCWT-front-end segmentation. All theorems are stated with their assumptions and verifiable conditions, offering a physically motivated approach to domain generalization in medical imaging. Full article

Background/Objectives: Osteoporosis is a major public health concern associated with increased fracture risk and reduced quality of life if not detected at an early stage. Automated analysis of knee X-ray images using artificial intelligence has shown promising potential for opportunistic osteoporosis screening. This study aims to develop and evaluate a multi-stage deep learning and machine learning framework for osteoporosis classification, with particular emphasis on external validation, calibration drift, and cross-domain generalization performance. Methods: Knee X-ray images were categorized into three classes: Normal, Osteopenia, and Osteoporosis. Deep features were extracted using pretrained convolutional neural networks, including ResNet18, EfficientNetB0, and DenseNet121. The extracted features were subsequently classified using multiple machine learning models, including Neural Network, Efficient Linear, Support Vector Machine, and Naive Bayes classifiers. Two data augmentation strategies were investigated: targeted minority-class augmentation and full 3× dataset expansion with class balancing. Model performance was evaluated using accuracy, precision, recall, F1-score, and AUC on internal validation, independent test sets, and external validation datasets. Additional analyses included reliability calibration assessment, isotonic recalibration, and class-prior boosting with cross-validated threshold optimization to address external domain shift. Results: EfficientNetB0 and DenseNet121 consistently outperformed ResNet18 across most evaluation metrics. Under the balanced augmentation strategy, EfficientNetB0 combined with Efficient Linear demonstrated strong and stable performance, while DenseNet121 paired with a Neural Network achieved the highest overall classification performance. External validation revealed a substantial discrepancy between AUC and threshold-based metrics, indicating the presence of calibration drift and class-prior mismatch across imaging domains. Reliability analysis showed severe probability collapse in the Osteopenia class during external testing. Post-hoc recalibration improved probability reliability, while class-prior boosting substantially increased Osteopenia sensitivity and improved balanced accuracy and macro F1-score under external validation conditions. Conclusions: The proposed framework demonstrates the feasibility of combining pretrained CNN-based deep feature extraction with machine learning classifiers for osteoporosis classification from knee X-ray images. The findings further highlight that maintaining model performance under external testing conditions may require not only strong feature extraction capability but also adaptive recalibration and deployment-aware threshold optimization to address calibration drift and cross-domain variability. While the results are encouraging, the present study should be considered a proof-of-concept investigation. Although the framework was evaluated using an independent public external dataset, further validation using larger and more diverse multi-center clinical cohorts is necessary to establish generalizability and clinical utility before routine clinical implementation can be considered. Full article

Background/Objectives: Contemporary patient-specific subperiosteal implants have re-emerged as graftless solutions for oral and maxillofacial rehabilitation, driven by advances in digital planning, CAD/CAM workflows, additive manufacturing, and biomaterial engineering. Their indications have progressively expanded from severely atrophic edentulous jaws to segmental defects, single-tooth replacement, congenital craniofacial anomalies, salvage situations, and oncologic reconstruction. This scoping review aimed to map the current evidence on modern patient-specific subperiosteal implants, focusing on indications, workflow, design principles, materials, outcomes, complications, and maintenance. Methods: A scoping review was conducted according to PRISMA-ScR principles to identify clinical studies, case series, case reports, systematic and scoping reviews, technical notes, finite element analyses, in vitro studies, and relevant translational investigations dealing with contemporary custom-made or CAD/CAM subperiosteal implants. The evidence was narratively synthesized according to clinical indication and thematic domains, including full-arch rehabilitation, sectional and single-tooth applications, congenital and post-oncologic defects, rescue indications, biomechanics, material selection, surface response, prosthetic protocols, and complication management. No quantitative meta-analysis was performed because of the scoping design and the substantial heterogeneity of study types, indications, implant systems, outcome definitions, and follow-up durations. Results: The final evidence map included 116 records, of which 56 were unique human clinical records with extractable denominators and 60 were biomechanical, in vitro, surface-biology, review, consensus, historical, or conceptual records. Of the 56 unique clinical records, 49 were mapped within the six indication-level clinical sections, while seven were retained as cross-cutting clinical evidence addressing patient-reported outcomes, design-related complications, bone apposition, anchorage strategy, comparative graftless rehabilitation, or reconstructive/prosthetic principles. The six indication-level sections included 52 clinical-record assignments: 15 for full-arch rehabilitation, 13 for segmental or sectional rehabilitation, one for single-tooth rehabilitation, four for congenital or craniofacial indications, 13 for post-oncologic or post-ablative reconstruction, and six for rescue or salvage indications. Because three records addressed more than one indication, these counts represent indication-level assignments rather than mutually exclusive clinical records. Reported survival in most short- to mid-term clinical series was generally high, commonly ranging from 90% to 100%, although lower values of 70–80% were reported in selected longer-term cohorts and survival clearly overestimated clinical success in some studies. Expanding applications include posterior mandibular and maxillary defects, lateral incisor agenesis, cleft-related or syndromic deformities, maxillectomy reconstruction, obturator support, and hybrid rehabilitation with endosseous implants; however, evidence for the indications at the extremes of this spectrum—single-tooth replacement and primary oncologic reconstruction—remains limited to small, largely single-group case series and reports. Soft-tissue events, including dehiscence, mucositis, recession, and framework exposure, were the dominant complications and showed wide variability, with reported recession/exposure rates ranging from approximately 10% in some sectional and full-arch series to as high as 65% in bilateral maxillary cohorts; their clinical significance varied from asymptomatic stable findings to progressive inflammatory complications requiring revision. Conclusions: Patient-specific subperiosteal implants represent a promising and increasingly versatile reconstructive option; however, the present findings should be interpreted as evidence mapping rather than as definitive comparative evidence. Their clinical use should remain highly selective, prosthetically driven, and supported by meticulous planning, rigid fixation, soft-tissue management, and structured maintenance. Standardized success criteria, longer follow-up, and comparative prospective studies are required. Full article

Extended-range three-dimensional (3D) single-molecule localization microscopy (SMLM) and single-particle tracking (SPT) require precise emitter localization across cellular-scale axial distances. However, long-rangeengineered point-spread functions (PSFs) spread photons over wider camera footprints, lowering the signal-to-noise ratio (SNR) and localization precision. We numerically evaluated a physics-guided deep learning workflow for 3D localization over a 10.0 µm axial range using simulated electron-multiplying charge-coupled device (EMCCD) images. The workflow combines an analytical secondary-astigmatism phase mask, frequency-domain cross-filtering, a cross-filtering generative adversarial network (CFGAN), and coarse-to-fine fitting. The optical model and engineered PSF provide physical signal priors, cross-filtering preserves directional Fourier-domain energy, and CFGAN suppresses residual structured noise before model-based localization. In low-SNR simulations, lateral, axial, and radial root-mean-squared localization errors (RMSEs) decreased from 54.11, 96.12, and 112.79 nm without denoising to 31.14, 39.06, and 50.12 nm after CFGAN denoising—close to Cramér–Rao lower-bound (CRLB) references of 34.39, 38.94, and 51.95 nm. High-SNR RMSE values were 8.78, 12.00, and 14.96 nm, comparable to CRLB references of 10.36, 11.71, and 15.64 nm. These simulations suggest that physics-guided restoration can improve extended-range 3D SMLM precision, while experimental validation remains necessary. Full article

Biochar water treatment stands at a decisive crossroads, where the promise of large-scale application meets the reality of laboratory trial-and-error. This study contends that the fundamental bottleneck to progress lies in the field’s persistent reliance on empirical experimentation and black-box data models. We therefore propose a conceptual research paradigm that aims to deeply integrate physics-informed machine learning (PIML) with life cycle assessment (LCA). The novelty of this framework lies in three dimensions: (i) the bidirectional information flow between PIML and LCA, enabling simultaneous material design and sustainability assessment; (ii) the embedding of fundamental physical laws (adsorption isotherms, kinetics, thermodynamics) directly into learning architectures to ensure physical consistency; and (iii) the extension to a water–energy–soil–food closed-loop system for holistic resource management. While the individual components of this framework have been demonstrated in other domains, their integrated application to biochar water treatment remains in early development stages. This perspective outlines potential pathways and identifies critical research gaps that must be addressed to realize this vision. The focus is on charting future directions rather than reporting established achievements. Through critical evaluation, we assess current integrated models under small-sample constraints and explicitly pinpoint explainability and cross-scale generalization as two indispensable gaps that industrial deployment demands be bridged. Building on this foundation, we outline a blueprint for a closed-loop system coupling water, energy, soil, and food, and present a three-phase roadmap for future research. This study seeks to offer a constructive perspective with the hope of supporting biochar technology toward more sustainable implementation. Full article

Background: Problematic smartphone use may threaten student well-being, especially among nursing students who rely on smartphones for academic and clinical activities. This study estimated potential problematic smartphone use among Greek nursing students, examined its association with quality of life (QoL), and explored whether contextual factors modified these associations. Methods: A cross-sectional study was conducted among 331 nursing students in Greece from September to November 2025. Participants completed an anonymous online questionnaire including sociodemographic data, the Smartphone Addiction Scale–Short Version (SAS-SV), and the World Health Organization Quality of Life–BREF (WHOQOL-BREF). Results: The mean SAS-SV score was 29.30 ± 9.69, and 18.9% of students screened positive for potential problematic smartphone use. Mean overall QoL and general health satisfaction were 3.80 ± 0.78 and 3.97 ± 0.88, respectively. Higher SAS-SV scores were associated with lower physical, psychological, and environmental QoL, but not with social QoL. Years of study moderated only the association with environmental QoL. Conclusions: Problematic smartphone use was associated with poorer physical, psychological, and environmental QoL among Greek nursing students. These domain-specific findings extend existing evidence and support integrating digital well-being, self-regulation, and sleep-hygiene strategies into nursing education and student-support services. Full article

Background: Lung transplant recipients often live for years with residual respiratory symptoms and psychological distress, but the pathways through which these factors affect quality of life (QoL) are not fully understood. We examined how transplant-specific pulmonary symptom burden and psychological distress relate to generic and transplant-specific QoL in long-term lung transplant recipients. Methods: In this cross-sectional study, 76 adult lung transplant recipients from a single center completed the Lung Transplant Quality of Life (LT-QoL) questionnaire, EQ-5D-5L, SF-36, St George’s Respiratory Questionnaire (SGRQ), and Hospital Anxiety and Depression Scale (HADS). A composite psychological distress index was derived from HADS-Anxiety, HADS-Depression, and the LT-QoL Anxiety/Depression and Health Distress subscales. Associations were examined using Pearson correlations, hierarchical linear regression (adjusting for age, sex, and time since transplant), and statistical mediation models examining psychological distress as a potential mediator between pulmonary symptoms and QoL outcomes. Results: Pulmonary symptom burden (LT-QoL Pulmonary Symptoms) was in the low–moderate range yet showed robust correlations with poorer generic, transplant-specific, and respiratory-specific QoL (|r| up to 0.82). The psychological distress index demonstrated good internal consistency (α = 0.84) and was strongly associated with worse EQ-5D, SF-36, and LT-QoL General QoL scores. In regression models, pulmonary symptoms and psychological distress independently predicted SF-36 overall QoL (R² = 0.55), whereas psychological distress was the stronger predictor of the EQ-5D Index Value. Statistical mediation analyses were consistent with partial mediation of the association between pulmonary symptoms and SF-36 and the EQ-5D Index Value, while effects on the EQ-VAS and LT-QoL General QoL were largely direct. Conclusions: Even modest pulmonary symptom burden and psychological distress are tightly linked to QoL years after lung transplantation. Routine follow-up should include brief assessment of both domains, and integrated care models that combine optimization of pulmonary status with targeted psychological support may be needed to preserve long-term QoL in lung transplant recipients. Full article

Industrial defect generation has long been challenged by the scarcity of real anomaly samples and the imbalance of defect categories, particularly in complex industrial scenarios involving transparent containers. Taking vials as an example, glass reflection, specular highlights, and fine-grained defects make continuous defect acquisition difficult, thereby making the realism and controllability of augmented samples critical to downstream detection performance. Although existing diffusion-based generation methods can improve synthetic image quality, they often require additional training or lightweight fine-tuning, which limits their efficiency in sample-limited industrial scenarios. To address this issue, this paper builds upon the TF-IDG framework and proposes a training-free industrial defect generation method based on multi-domain consistency and geometric-semantic constraints. To alleviate the unnatural texture details, boundary transitions, and background blending commonly observed in generated defects, a multi-domain consistency constraint is introduced to enhance generation realism from both frequency-domain structures and cross-domain contextual representations, thereby improving anomaly texture expression and overall visual coherence. To further mitigate unstable defect contours, spatial deviation, and structural mismatch with target objects, a geometric-semantic constraint is designed to regulate the generation process through elastic shape constraints and semantic region-anchored attention, enhancing the rationality of defect morphology evolution and spatial localization. Experimental results on both the MVTec AD dataset and a self-built vial defect dataset demonstrate that the proposed method outperforms comparative approaches. Specifically, when YOLOv11 is used as the downstream detector, the mAP@50 on the MVTec AD dataset and the self-built vial defect dataset is improved from 88.5% and 98.0% for the TF-IDG baseline to 89.6% and 98.8%, respectively. Full article

Accurate quantification of ecological losses in estuarine environments is often hindered by the mismatch between coarse-resolution biological surveys and fine-scale physical disturbances from engineering activities. While numerical models can simulate high-resolution environmental shifts, the inherent sparsity of ecological monitoring points limits the precision of spatial impact assessments. This study develops an integrated spatial-downscaling framework to transform sparse monitoring data into a high-resolution spatial continuum. A three-tiered modeling approach was used: first, the estuarine domain was partitioned into five eco-hydrodynamic zones using an entropy-weighted Support Vector Machine (SVM); second, localized chained Generalized Additive Models (GAMs) were established within each zone using MIKE-simulated hydrodynamic and water-quality data as proxy drivers; and third, these localized response functions were propagated across the study area to quantify multi-trophic biomass and economic losses. The framework revealed substantial spatial non-stationarity. Dredging operations locally altered the estuarine hydrodynamic regime. In northern channels, decreases in flow velocity were statistically associated with phytoplankton biomass to decline by 5.0% to 23.42%. Conversely, southern velocity increases enhanced water exchange and plankton growth. Using silt curtains as a mitigation strategy reduced the loss of phytoplankton by 11.4% and zooplankton by 9.6%. As a result, the total economic loss decreased from 26.54 million CNY to 25.34 million CNY, equivalent to a 4.5% reduction in economic loss. These results indicate that the proposed downscaling method can generate spatially explicit biological estimates. By offering a systematic pathway for impact evaluation and compensation in data-limited coastal regions, this framework supports more ecologically sustainable dredging operations. Nevertheless, the framework remains dependent on the representativeness of sparse monitoring stations, and future applications should integrate cross-estuary validation to improve transferability and uncertainty control. Full article