Search Results (3,309)

Medical image classification models often struggle to generalise across heterogeneous clinical domains owing to variations in visual characteristics, acquisition conditions, and class imbalance. Existing studies largely address these challenges independently, with limited investigation into their combined impact on classification robustness. This paper proposes a lightweight dual-stream framework based on a pretrained ConvNeXt-Tiny backbone, integrating complementary global semantic and local structural feature representations with imbalance-aware optimisation. Rather than modifying the backbone, the framework enhances feature discrimination through dual-stream processing while incorporating class-balanced focal loss and stratified sampling for minority-class recognition. The framework is evaluated independently on four public datasets spanning dermatology, ophthalmology, and gastrointestinal endoscopy using five-fold cross-validation to assess architectural robustness across heterogeneous domains rather than cross-domain transfer of a single trained model. Experimental results show strong performance across all datasets, achieving AUC values above 0.92. Ablation studies confirm that dual-stream representation learning provides the primary performance gains, while imbalance-aware optimisation further improves robustness under severe class imbalance. These findings demonstrate an effective and practical solution for medical image classification across heterogeneous clinical imaging domains without requiring dataset-specific architectural modifications. Limitations regarding independent per-domain (rather than cross-domain) evaluation, baseline comparability, and incomplete quantitative calibration analysis are discussed explicitly and identified as directions for future work. Full article

Maritime decarbonisation requires transparent use of verified emissions data without overstating what regulatory datasets can explain. This study examines how EU Monitoring, Reporting, and Verification (MRV) data can support descriptive benchmarking of container ship distance-normalised fuel consumption. Of 14,147 2024 MRV records, 2097 container ships were retained. Fuel consumption per nautical mile (FC/nm) is treated as an absolute distance-normalised fuel-use indicator, not a cargo-adjusted efficiency metric. Four algorithms were assessed under target-related diagnostic and reduced specifications. Diagnostic results represent target-related reconstruction, not genuine prediction. Training-only hyperparameter tuning yielded reduced-specification test R² values of 0.8958 for Gradient Boosting and 0.8888 for XGBoost. Random-Forest attribution in the reduced specification was dominated by emissions-derived predictors. A strict non-emissions Random Forest using only sea time and certification performed worse than the mean-value baseline (R² = −0.2890). Comparison with MRV-reported fuel consumption per transport work (mass), available for 95.1% of vessels, showed a strong inverse association (ρ = −0.822), confirming that FC/nm quartiles are not cargo-adjusted efficiency rankings. The study shows that explainable machine learning can transparently diagnose information structure in MRV data, while demonstrating that these variables alone cannot support causal, voyage-level, cross-vessel efficiency, or policy interpretations. Full article

Artificial intelligence (AI) and machine learning (ML) are increasingly applied to road traffic congestion prediction, but heterogeneous outcomes, models, horizons, and evaluation practices limit comparability. The objective of this study was to synthesize methods, applications, validation, explainability, and reproducibility in AI/ML-based road traffic congestion prediction and forecasting. Following PRISMA 2020, Scopus, Web of Science Core Collection, and IEEE Xplore were searched through 5 July 2026 for English-language journal articles and full conference papers published from 2000 to 2026. Two external reviewers independently screened 734 unique records and assessed the retrieved full texts, while the author resolved disagreements against the predefined eligibility criteria. Study characteristics, prediction tasks, congestion indicators, model families, metrics, explainability, validation, and data/code availability were synthesized descriptively and narratively. Of 1131 records identified, 397 duplicates were removed and 734 records were screened. Full-text retrieval was sought for 339 reports; 195 could not be retrieved, 144 were assessed for eligibility, and 129 were included. Congestion level was the main prediction task, while traffic flow and speed were the most frequent indicators. Heterogeneity and the absence of verified numerical performance values precluded meta-analysis or model ranking. Explainability was limited, and external validation, transferability, and reproducibility were insufficiently documented. Progress requires standardized outcomes, transparent validation, reproducible workflows, explainable models, and independent testing across networks and cities. Full article

Against the background of AI penetration in C2C second-hand commerce, AI chatbots have emerged as vital operational tools for individual sellers. Prior technology adoption literature predominantly emphasizes positive acceptance drivers yet overlooks the simultaneous existence of benefit and risk perceptions, and cannot fully explain individual sellers’ paradoxical adoption intentions. This study integrates the Unified Theory of Acceptance and Use of Technology (UTAUT) and Technology Threat Avoidance Theory (TTAT) to construct an integrated dual mediation model, adopting perceived effortlessness and perceived risk as parallel mediators. Online questionnaire surveys were distributed to individual sellers operating on Chinese C2C second-hand trading platforms. After screening and data cleaning, a final valid sample of 261 respondents was retained. Partial least-squares structural equation modelling (PLS-SEM) was employed to empirically test the proposed research model. The results demonstrate that personal factors (perceived busyness, desire for control, AI acceptance) and situational factors (product standardization, social influence, platform safeguards) jointly shape individual sellers’ dual perceptions. These antecedents exert differentiated impacts on the two mediating constructs, which subsequently predict individual sellers’ intentions to adopt AI chatbots. This research clarifies the balancing mechanism of positive and negative perceptions in technology decision-making. The framework reconciles conflicting psychological evaluations of intelligent systems, deepens the understanding of paradoxical adoption behaviors, and offers practical implications for optimizing intelligent services within C2C second-hand ecosystems. Full article

As artificial intelligence systems increasingly serve as a gateway for access to information, economic opportunity, and civic life, higher education programs training AI developers must evolve to prepare practitioners who are not only technically proficient in building AI solutions but also socially accountable in their work. This paper proposes a conceptual framework for integrating explainable AI (XAI) and social accountability into data science, computer science, and information science curricula. Based on accountability theory, information science, and recent XAI research, the proposed framework is organized around four interrelated pillars: answerability, responsibility, enforcement, and reflexivity. These pillars are further situated within technical, social, organizational, and political dimensions of XAI implementation, with particular focus on how XAI techniques such as LIME, SHAP, model cards, and counterfactual explanations can be operationalized as instruments of meaningful accountability. This paper then proposes a multi-level governance framework that links interpretability methods to institutional oversight, regulatory literacy, and participatory design, illustrated through a concrete scenario grounded in graduate data science education. Together, these elements represent a new pedagogical approach that can equip future AI developers to design and deploy AI systems that are accurate as well as transparent, justifiable, and responsive to the communities they serve. Full article

Background/Objectives: Artificial intelligence is increasingly used in medicine, particularly in orthopedics, to support medical image interpretation, diagnosis, clinical decision-making, and treatment selection. Its integration into clinical practice depends largely on doctors’ perceptions and attitudes toward this technology. This study aimed to identify the main factors associated with Romanian orthopedic doctors’ attitudes toward AI and their intention to use it in clinical practice. Methods: The study included 166 doctors specializing in Orthopedics and Traumatology in Romania. Data were collected through an online questionnaire using non-probability snowball sampling. The conceptual model examined the associations between five factors and doctors’ attitudes toward AI: perceived usefulness, perceived ease of use, trust in AI, digital competencies, and perceived advantages. The relationship between attitude and intention to use AI was also analyzed. The model was tested using structural equation modeling with WarpPLS 8.0. Results: All factors were positively and significantly associated with doctors’ attitudes toward AI. Perceived usefulness showed the largest association with attitude (β = 0.27), followed by digital competencies (β = 0.22). The model explained 83% of the variance in attitude. Attitude was positively associated with intention to use AI (β = 0.65), with the model explaining 43% of the variance in intention. The model showed a GoF of 0.707 and an AFVIF of 2.645. Conclusions: The findings may support healthcare managers in developing digitalization strategies, structured digital competency training, and clear institutional guidelines for the safe and effective use of AI in clinical practice. Full article

Breast cancer remains a major global health concern, making accurate lesion assessment essential for effective clinical decision making. Deep learning has shown promising performance in breast ultrasound analysis, yet models evaluated on data from the same source may not generalize reliably to images acquired using different scanners and acquisition settings. This study therefore examines cross-dataset generalization and the factors that can improve it. Five lightweight YOLO instance-segmentation architectures (YOLOv8n, YOLO11n, YOLO11s, YOLO26n, and YOLO26s) were trained using a patient-grouped BUS-BRA protocol and a single training seed, and evaluated internally on held-out data and externally on BUS-UCLM, which served as the single target dataset. Ultrasound-specific preprocessing, test-time augmentation (TTA), and Optuna-selected configurations were assessed across 40 paired internal–external comparisons. Cross-dataset evaluation demonstrated consistent lesion delineation on BUS-UCLM, with matched Dice ranging from 0.861 to 0.875 and matched IoU from 0.764 to 0.786 across the five architectures. Preprocessing improved mask mAP@50–95 across all ten checkpoints on both datasets, while TTA improved detection-adjusted Dice despite having little effect on mAP@50–95. Optuna tuning improved external mAP@50–95 across all five architectures despite limited internal gains. Grad-CAM++ showed predominantly lesion-centered attention across both datasets, while ONNX Runtime deployment achieved 5.65–13.62 FPS on CPU. These findings highlight the importance of external validation, detection-aware evaluation, and efficient deployment for reliable breast ultrasound segmentation. Full article

Accurate runoff forecasting serves as a fundamental basis for the scientific management of water resources and flood and drought risk mitigation. Owing to the nonlinearity, non-stationarity, and multi-scale temporal characteristics of runoff series, existing deep learning models still exhibit notable limitations in capturing long-term trends, responding to abrupt hydrological events, and ensuring model interpretability. The original Transformer relies on global self-attention, whose computational complexity increases quadratically with sequence length; it also has limited capacity to capture short-term local temporal dependencies such as rainfall–runoff relationships, and lacks prior constraints tailored to hydrological processes. To address these challenges, this study proposes a collaborative forecasting framework that integrates a gated convolutional Transformer (GCTrans) with an improved black kite algorithm (IBKA), enabling accurate, stable, and interpretable daily-scale runoff prediction. The GCTrans model consists of three customized modules: convolution-enhanced positional encoding (CEPE), which combines learnable positional encoding with local causal convolution to strengthen the temporal association of adjacent rainfall–runoff events, thereby providing a hydrologically meaningful positional reference for the attention mechanism; gated convolutional attention (GCA), which adopts a dual-path parallel architecture comprising global self-attention and local causal convolution to adaptively fuse long-term seasonal patterns with short-term storm-induced variations, thus capturing both baseflow evolution and flood peak responses; and a temporal gated output layer (TGOL), which performs adaptive feature weighting along the temporal dimension to selectively enhance the contribution of critical driving periods associated with extreme flood events, thereby improving the flood peak prediction accuracy. In addition, an improved black kite algorithm (IBKA) was developed by incorporating Tent chaotic initialization to enhance initial population diversity and introducing cosine adaptive inertia weights to dynamically balance global exploration and local exploitation, effectively alleviating premature convergence in high-dimensional hyperparameter spaces. Validation using data from the ME-Inland snowmelt-dominated watershed and the OR-Coastal storm-driven coastal watershed in the United States demonstrated that the GCTrans model consistently outperformed benchmark models including TCN, LSTM, Transformer, and Informer. After synergistic optimization with IBKA, both prediction accuracy and stability were further improved. SHAP-based interpretability analysis revealed that the model’s feature response patterns are statistically consistent with the rainfall–runoff generation mechanisms of the study basins: temperature-related drivers dominate in the inland watershed, while precipitation plays a dominant role in the coastal watershed, confirming the hydrological plausibility of the model’s decision-making logic. The integrated framework—encompassing model architecture, optimization algorithm, and interpretability—offers a valuable methodological reference for deep learning-based runoff forecasting in complex hydrological settings. Full article

This study investigates the development of an AI-driven document authentication framework for Anti-Money Laundering (AML) and Know Your Customer (KYC) verification environments. Conventional manual inspection and rule-based verification techniques often fail to detect sophisticated forged identity documents containing subtle visual or semantic manipulations. To address this limitation, the proposed framework combines handcrafted forensic feature extraction, OCR-driven semantic analysis, rule-based semantic field extraction and Random Forest classification to identify inconsistencies within identity documents captured under realistic mobile imaging conditions. Experimental evaluation was conducted using selected MIDV-2020 identity document subsets consisting of Albanian identity cards, Latvian passports, and Slovakian identity cards. The proposed framework achieved a recall rate of 92.31% and an overall accuracy of 84.85% on the held-out test set, while maintaining interpretable forensic feature analysis suitable for regulated AML/KYC environments. The results demonstrate that lightweight and explainable machine learning approaches can provide effective forged-document detection without requiring computationally intensive deep learning architectures. Full article

Artificial intelligence (AI) is increasingly embedded in organizational decision-making, requiring employees not only to use AI-generated recommendations but also to evaluate their quality and determine when reliance is appropriate. Although established research examines behavioral reliance on algorithmic and AI advice, fewer studies have approached performance under AI advice as an individual-differences assessment problem integrating psychometric structure, cognitive correlates, process indicators, and criterion-related evidence. This study developed and initially validated a CHC-informed, performance-based assessment of cognitive performance under AI advice using 24 organizational decision scenarios. The assessment was designed around three closely related content/performance dimensions—AI error detection, evidence integration, and cognitive control and adaptive reliance—while also capturing confidence, response time, and reliance behavior. The validation sample comprised 780 employed adults in Türkiye. Psychometric analyses included confirmatory factor analysis, multidimensional item response theory, response-time analyses, scenario-level logistic regression, measurement invariance, differential item functioning, and internal cross-validation. Results indicated a dominant general cognitive-performance component together with additional structure corresponding to the three theoretically specified dimensions. Assessment performance was positively associated with established cognitive measures, including ICAR-16 reasoning performance, working memory, processing speed, and attentional control, whereas associations with AI-related self-reports were generally weaker. Dimension-aligned analyses supported the expected associations of ICAR-16 with AI error detection and working memory with evidence integration. Performance was also moderately associated with concurrently assessed organizational decision quality (r = 0.408) and explained additional variance in this criterion beyond demographic and work characteristics, AI experience, conventional cognitive-performance measures, and AI-related self-reports (ΔR² = 0.103, p < .001). In contrast, several hypothesized scenario-specific associations involving AI confidence, time pressure, interruptions, and resistance to confidently inaccurate advice were not supported. Overall, the findings provide initial evidence for a performance-based approach to assessing how employees evaluate and respond to AI advice, while indicating that the proposed scenario-specific mechanisms and group-comparability findings require further replication before consequential applications are considered. Full article

This study benchmarks four classifiers—Random Forest (RF), Classification and Regression Trees (CART), Support Vector Machine (SVM), and Gradient Tree Boosting (GTB)—on an identical 16-feature KOMPSAT-3/5 and Sentinel-1/2 stack harmonized to a common 2.8 m grid within Google Earth Engine. RF and GTB reached overall accuracies of 92.4% and 92.8%, separating them from the unconstrained CART under family-wise correction and from the linear-kernel SVM under false-discovery-rate control. Class-specific SHAP was cross-interpreted against permutation importance, feature-correlation analysis, and a leave-one-sensor-out (LOSO) ablation, with SHAP and permutation-importance rankings in broad agreement (Spearman ρ = 0.72–0.92). The central contribution is to characterize two conditions—strong within-sensor collinearity and class-conditional information concentration—under which feature-level attribution and sensor-level necessity diverge. In the Bare Land class, no individual Sentinel-2 SWIR band ranks among the top three by SHAP, yet withholding the Sentinel-2 group produces the largest class-wise degradation observed (ΔF1 = 0.102 for RF and 0.130 for GTB), consistent with attribution dilution across the collinear SWIR1–SWIR2 pair (r = 0.97). Because none of the sixteen ablation contrasts survives multiplicity correction with a sample size of 250 (n = 250), these contrasts are reported as effect-size estimates rather than confirmatory tests. Full article

Closed-set evaluation is the most common method in Internet of Things (IoT) sensor network botnet attack detection research. In this setting, classifiers are tested only on attack families encountered during training. As a result, they often fail to detect genuinely new (zero-day) malware in real-world environments. This study evaluates a calibrated two-tier detection pipeline combining an XGBoost gradient-boosted tree classifier for known-class prediction with a Monte Carlo (MC) dropout multilayer perceptron (MLP) as an independent uncertainty estimator, under conditions closer to real deployment than most prior evaluations. The uncertainty model is used to flag unfamiliar traffic that may represent zero-day traffic. The classifier is trained and calibrated on the Bot-IoT dataset and evaluated for zero-day generalization on N-BaIoT, which contains the previously unseen Mirai and BASHLITE malware families. During data preparation, a severe train/test duplication problem was identified. Around 97% of a naive split was found to be duplicates and was removed before splitting. The base classifier achieved 99.4% accuracy in known-class detection. However, the confidence scores were negatively correlated with zero-day traffic, producing an area under the receiver operating characteristic curve (AUROC) of 0.32. This indicates the overconfident misclassification of unseen attacks. The proposed uncertainty gate improved zero-day discrimination to an AUROC of 0.66, with catch rates reaching 13.75% for Mirai. An explainable AI analysis using Shapley Additive Explanations (SHAP) linked this gap to packet size statistics. To evaluate practical feasibility, the complete framework was deployed on an embedded Raspberry Pi Compute Module 5 (CM5) edge device. The full uncertainty-gated pipeline achieved mean inference latency of 15.03 ms per sample. Overall, these findings demonstrate that uncertainty-aware gating can improve zero-day robustness over conventional closed-set classification in resource-constrained IoT sensing environments. Full article

The Internet of Vehicles (IoV) connects vehicles, roadside infrastructure, and cloud or edge platforms to support safer and more efficient transportation. This connectivity also exposes vehicular networks to message spoofing, denial-of-service, and man-in-the-middle attacks, and to threats that target the machine-learning defenses themselves, such as model poisoning, gradient inversion, and adversarial examples. Machine learning and deep learning are now the dominant defensive tools, and the resulting literature is large and fragmented. This survey reviews that literature using a documented database search and stated inclusion criteria. The retained studies are organized into three categories, each defined by the dominant security-design problem it addresses. The categories cover classical detection and prevention, decentralized security based on federated learning and blockchain, and the assurance directions of Large Language Model (LLM)-driven detection, adversarial robustness, and explainable artificial intelligence. Rather than pooling reported scores, the survey records the evaluation setting of each study. This shows that de-duplication, class balancing, and metric aggregation account for much of the apparent variation in reported performance. A study-level assessment finds that no reviewed study demonstrates more than one of the three assurance capabilities, and no detector evaluated on vehicular traffic has been tested against adversarial perturbation. The survey concludes with a cross-category comparison, an EU AI Act alignment assessment, and recommendations for future research. Full article

Steel rails are critical components of industrial dynamic transportation systems, and their in-service fault diagnosis demands reliable discrimination of multiple defect categories under multi-mode Lamb-wave dispersion and sample-level physical-parameter drift. The rail surface is modelled by a thin metal plate instrumented with two EMAT probes (the standard laboratory surrogate for in-service rail inspection), and the proposed architecture is evaluated on this rail-equivalent plate geometry. Purely data-driven one-dimensional classifiers plateau near 80% test accuracy on a 25,000-sample simulated EMAT A-scan benchmark, while conventional physics-informed neural networks (PINNs) that inject the physical prior only at the loss-function level fail to break this ceiling. We propose a hybrid data–physics residual network, the proposed EMAT-PINN, that couples a convolutional backbone with a logit-orthogonal four-head architecture tying each defect class—hole, crack, corrosion, weld—to one simulator-derived physical quantity (reflected energy, S0/A0 ratio, arrival time, or dispersion shift) via a bias-free additive projection of the class logit. The bias-free construction guarantees that deleting or zeroing any head collapses the affected class logit to the shared baseline, so the remaining heads cannot reroute around the missing head—a structural non-replaceability that supports explainable fault diagnosis. Combined with a four-term physics regression loss (${\lambda }_{\mathrm{phys}}=2.0$), the resulting proposed EMAT-PINN attains 95.05% test accuracy at only 0.195 M parameters, with every knockout ablation dropping the model below the $80\%$ threshold commonly referenced as a practical acceptance benchmark. This per-class mapping provides an auditable link between the model’s internal representation and the physical scattering mechanism behind each decision, directly supporting explainable fault diagnosis in industrial deployment. Full article

Geopolitical uncertainty may affect financial markets, but its incremental value for forecasting emerging-market stock returns remains unclear. Using monthly data from January 1990 to July 2026, this study compares ARIMA-GARCH and ARIMAX-GARCH benchmarks with Random Forest, XGBoost, LightGBM, and a zero-return benchmark across 1-, 3-, 6-, and 12-month horizons. Forecasts are generated with a target- and predictor-leakage-safe expanding-window design: training targets never exceed the forecast origin, and no realized future predictor values are used. Econometric forecasts are conditional on fixed ARIMA and ARIMAX orders selected during full-sample diagnostics. Among the estimated models, XGBoost achieves the lowest RMSE and MAE at three months, Random Forest has the lowest RMSE at one and six months, and ARIMA-GARCH performs best at twelve months. Nevertheless, the zero-return benchmark records the lowest RMSE at every horizon, while Diebold–Mariano tests generally do not reject equal predictive accuracy, and the Model Confidence Set retains multiple competitive models. Rolling SHAP analysis ranks geopolitical risk first among 16 predictors in the three-month XGBoost model, accounting for 14.38% of aggregate mean absolute attribution. Thus, geopolitical risk provides model-specific short-horizon conditioning information, but its standalone accuracy gain is modest, statistically insignificant, and absent at twelve months. Full article