Search Results (550)

Knowledge tracing (KT) is a core technology in intelligent tutoring systems that predicts students’ future responses by analyzing their historical interaction sequences. Although existing deep learning-based KT models achieve high predictive accuracy, their “black-box” nature severely hinders practical deployment in educational scenarios. To balance predictive performance and interpretability, this paper proposes AMDKT, an interpretable dual-stream Transformer model grounded in the Student Proficiency–Item Competency Matching (SPIM) mechanism. The model employs two parallel Transformer branches to separately model the temporal evolution of student proficiency and the competency demands of each item and defines the discrepancy between their outputs as “proficiency surplus.” A non-negative regularization loss is further introduced to constrain the proficiency surplus to be non-negative for correctly answered samples, thereby embedding pedagogical rules into the optimization objective and ensuring that the model outputs conform to educational cognitive priors. Experiments on five public datasets show that AMDKT achieves AUC performance comparable to the state-of-the-art AKT model, and obtains statistically superior results against DKT, DKVMN, DEEP-IRT, and DIMKT on most datasets, with relatively weaker performance observed on the KDD Cup 2010 dataset. Ablation studies verify the effectiveness of each component, and visualization results demonstrate that AMDKT produces smooth, coherent, and interpretable student proficiency trajectories, providing a fine-grained tool for quantifying individual learning progress. Therefore, AMDKT offers a feasible solution for applications such as weak knowledge point localization, adaptive exercise recommendation, and learning risk warning. Full article

Background: Neoadjuvant systemic therapy (NST) enables tumor downsizing and increases the feasibility of breast-conserving surgery (BCS) in patients with early breast cancer (EBC). Reliable tumor localization after NST remains challenging, particularly in patients with a complete clinical response. If clips are not visible on ultrasound, stereotactic mammography-guided wire localization is required, which involves additional radiation exposure and may increase patient discomfort and procedural complexity. The 3D-shaped Tumark^® Vision clip may enable ultrasound-guided localization after NST. Methods: In this prospective multicenter registry study (NCT04468113), 324 patients with biopsy-proven EBC scheduled for NST and breast-conserving surgery were enrolled across 19 German centers. Clip placement was performed under ultrasound guidance prior to NST, and clip detectability was assessed at predefined time points during therapy (4–8, 9–12, and ≥13 weeks after treatment initiation). Detection rates, visualization, and preoperative localization methods were recorded. Non-detectable clips required stereotactic wire localization, whereas ultrasound-guided localization was performed for detectable clips. Results: The preoperative detection rate was 91.1%. Clip detectability was higher in patients with longer NST durations and partial response by imaging. Ultrasound-guided wire localization was feasible in 214 patients (83.9%); among these, 165 patients (64.7%) underwent localization without and 49 patients (19.2%) with post-procedural mammographic verification. Stereotactic localization was required in 41 patients (16.1%), primarily due to non-visualization of the clip on ultrasound. The accuracy of ultrasound for residual tumor assessment was moderate (72.0%), with a sensitivity of 70.7%, indicating limitations in its ability to reliably assess residual tumor extent as a standalone modality. Conclusions: The 3D nickel–titanium clip enables ultrasound-guided tumor localization after NST in the majority of patients with early breast cancer, although additional mammographic guidance remains necessary in a relevant proportion of cases. Full article

Blast-induced traumatic brain injury is a prevalent and underreported condition, particularly among military service members, for whom effective prophylactic interventions are lacking. Nano-pulsed laser therapy (NPLT) is a non-invasive neuromodulatory approach that delivers short pulses of near-infrared light to generate optoacoustic effects within cerebral tissue and has previously demonstrated therapeutic benefit following TBI. In this study, we evaluated whether a single pre-exposure application of NPLT could confer protection against neurological, cognitive, and cellular sequelae of mild blast injury. Adult male Sprague-Dawley rats were randomized to receive NPLT or Sham treatment 24 h prior to either Sham or mild blast exposure using the Advanced Blast Simulator. Neurological reflexes and vestibulomotor function were assessed on post-injury days (PIDs) 1–5, while cognitive performance was evaluated using the Morris Water Maze on PIDs 13–17. Histological analyses of microglia, astrocytes, and myelination were performed on PID 17. A single mild blast did not significantly alter gross neurological function but was associated with deficits in fine motor coordination and cognitive performance. Pre-exposure NPLT modestly attenuated blast-associated fine motor dysfunction, with a significant improvement compared with TBI on PID 4. In the Morris Water Maze, TBI animals exhibited significantly increased latency compared with Sham on PIDs 13 and 17, whereas NPLT + TBI animals did not significantly differ from Sham across the testing period, consistent with preservation of cognitive performance. Histological responses were regionally heterogeneous: NPLT alone produced distinct glial alterations, while NPLT + TBI animals exhibited a mixture of treatment- and injury-associated responses rather than uniform normalization to uninjured controls. NPLT did not prevent localized blast-associated reductions in corpus callosum myelin staining. In naive animals, NPLT significantly increased hippocampal brain-derived neurotrophic factor (BDNF) mRNA expression 24 h after treatment. A single pre-injury application of NPLT was associated with functional resilience following mild blast exposure despite persistent and regionally heterogeneous histopathological alterations. Increased hippocampal BDNF 24 h after NPLT, together with region-specific glial changes following NPLT in the absence of injury, demonstrates that a single treatment produces sustained molecular and cellular effects before blast exposure. These findings are consistent with the hypothesis that prophylactic NPLT establishes an altered pre-injury biological state that may modify the subsequent response to blast and support further investigation of NPLT as a prophylactic strategy and of the mechanisms underlying NPLT-associated preconditioning. Full article

Towed hydrophone arrays are widely employed for underwater target detection and direction-of-arrival (DOA) estimation. However, array shape distortion induced by ocean currents, internal waves, and platform maneuvers severely degrades beamforming performance and DOA estimation accuracy. In this paper, a novel Bayesian fusion framework is proposed to achieve robust array shape estimation. Specifically, based on the time-delay estimates derived from the phase differences of line-spectrum components in a pre-processing step, the array geometry is first reconstructed via a piecewise straight-line fitting method. Concurrently, an existing hidden Markov model (HMM)-based method is adopted to estimate the inter-segment deviation angles, in which the smoothness of the array shape is enforced through the state-transition probabilities. The proposed framework then treats these two preliminary estimates as observations from distinct sources and incorporates a smoothness prior within a maximum a posteriori (MAP) formulation that admits a non-iterative closed-form solution to enforce physical continuity constraints on the array geometry. By fusing these complementary estimates, the proposed method simultaneously preserves local sensitivity to fine-scale bends and maintains global consistency of the array shape. Both simulation and lake-trial experiments validate the effectiveness of the proposed method, reducing the array shape estimation error by more than 30% relative to representative existing methods. Moreover, by relying solely on the received acoustic data, the method lowers the dependence on auxiliary sensors and the associated system cost. Full article

Accurate non-contact human height estimation is vital for large-scale growth monitoring in schools but remains challenging for monocular RGB sensors due to scale ambiguity and keypoint jitter. This study proposes a robust temporal SegPose framework for high-precision height measurement in unconstrained outdoor playground environments. We develop a multi-task deep learning model using a MobileNetV4 backbone and a novel Height-Aware Boundary Refinement (HABR) module, which utilizes nose-spatial priors to refine cranial vertex localization. To resolve scale issues, a three-way orthogonal calibration system is established using existing playground marking lines and goalposts to dynamically estimate ground plane metric factors. A linear Kalman filter is integrated to smooth keypoint trajectories, suppressing gait-induced oscillations and reducing high-frequency jitter by 64.92%. Validated on a dataset of 95 volunteers (53 males, 42 females) at distances of 6–12 m, the proposed system achieves a mean absolute error (MAE) of 1.42 cm and a mean absolute percentage error (MAPE) of 0.84%, significantly outperforming recent Transformer-based state-of-the-art methods. The framework operates at 42.7 FPS, ensuring real-time performance while adhering to a privacy-preserving protocol that decouples biometric records from individual identities. These results demonstrate that our framework effectively overcomes boundary ambiguity and distance-dependent resolution loss, providing a reliable, efficient, and ethical solution for automated physical health assessments in educational settings. Full article

Accurate fault location in low-observability active distribution networks is hindered by uncertain inter-node relationships, underutilized early transients, and insufficient global context. To address these challenges, this paper proposes an adaptive Mamba-driven spatiotemporal graph neural network (AM-STGNN). It provides a unified task-driven spatiotemporal representation framework that progressively integrates fault-propagation modeling, global dependency modeling, transient-state learning, and topology-aware discriminative enhancement. Specifically, a prior-guided adaptive implicit topology is first learned to characterize task-dependent electrical coupling relationships among sparse observation nodes. Based on the resulting topology, topology-conditioned multi-order feature propagation and a dual-axis linear-attention module based on the spatiotemporal graph transformer (STGformer) are employed to capture local and global spatiotemporal dependencies. The resulting global spatiotemporal representation is subsequently processed by a Mamba selective state-space encoder to model input-dependent temporal evolution and emphasize informative fault transients. Finally, element-wise gated fusion, topology-aware differential output, and a margin constraint are employed to integrate the STGformer and Mamba representations and enhance the separability of adjacent faulted line sections with similar response characteristics. Extensive experiments demonstrate the effectiveness of AM-STGNN, while additional evaluations confirm its applicability to larger-scale networks, strongly phase-unbalanced conditions, and field-measured operating backgrounds. Robustness tests under individual and multi-level joint disturbances further demonstrate the practical relevance of the proposed architecture. Compared with the baseline models, AM-STGNN achieves consistent improvements in the macro-averaged F1 score (Macro-F1), exact accuracy, and one-hop accuracy under the clean IEEE 123-node condition. More importantly, it maintains clear performance advantages under identical mild, moderate, and severe joint disturbances, demonstrating improved robustness and practical relevance under simulated non-ideal operating conditions. Full article

Background/Objectives: Accurate prognostic stratification after nephrectomy for clear cell renal carcinoma (ccRCC) remains challenging. Traditional models often lack the intuitive clinical application or the ability to handle non-linear interactions between variables. We aimed to develop and internally validate a decision tree-based model to predict cancer-specific survival in patients with ccRCC following nephrectomy. Methods: We analyzed 79,526 patients with ccRCC who underwent nephrectomy from the SEER database (2012–2018). Patients were randomized into development (2/3) and validation (1/3) cohorts. A conditional inference tree was constructed to predict cancer-specific survival. Multiple imputation by chained equations was used to handle missing data. Discriminatory ability was assessed using the C-index. Net clinical benefit was evaluated with decision curve analysis. The model was evaluated using CHARMS and PROBAST. Results: A decision tree with 15 risk groups is presented, further classified into high-, intermediate-, and low-risk categories according to observed median survival. The final predictors were tumor localization, tumor grade, TNM stage, age, and sarcomatoid differentiation. The model demonstrated excellent discriminatory performance, with a C-index of 0.846 (95% CI: 0.834–0.847). PROBAST assessment showed low risk of bias and low concern regarding applicability. Conclusions: The use of decision trees provides an interpretable alternative to conventional regression-based models. Three main risk categories and 15 subgroups are proposed based on tumor localization, tumor grade, TNM stage, age, and sarcomatoid differentiation. Our model demonstrates good applicability and a low risk of bias according to PROBAST guidelines; however, external validation in independent cohorts is required prior to clinical implementation. Full article

This study presents a hybrid geodetic validation framework for assessing the geometric consistency of the axis of a steel lattice transmission tower determined from terrestrial laser scanning (TLS) data using an independently established tachymetric reference. Unlike previous investigations that focused on the influence of TLS scanner characteristics, registration strategies, or internal consistency of TLS-derived axes, the proposed approach introduces an external geodetic reference, enabling direct external assessment of TLS-based geometric axis estimation. The reference axis was determined at fourteen height levels, while the TLS axis was estimated from horizontal cross-sections of a point cloud acquired from multiple scanning stations and registered using a cloud-to-cloud method without control points. To enable direct comparison, both datasets were transformed into a common reference system using a seven-parameter Helmert transformation. The transformation was applied solely to remove differences between the independent local coordinate systems prior to the geometric comparison. Axis consistency was evaluated using residual vectors and three-dimensional distances between corresponding points. The mean deviation was 0.031 m, the RMS value was 0.033 m, and the maximum deviation reached 0.078 m. Larger discrepancies occurred predominantly in the upper sections of the structure, in a pattern consistent with the combined influence of TLS registration uncertainty, non-uniform point-cloud coverage, and local geometric conditions. A comparison of TLS axis estimators (centroid, LS-R regression, and PCA) showed that PCA produced an RMS value close to that of the centroid estimator, whereas LS-R produced a higher RMS value; the maximum deviation was lowest for the centroid estimator and highest for PCA. Regression analysis revealed a statistically significant linear trend in the X direction (p = 0.019), indicating residual systematic geometric drift after coordinate-system integration. The obtained discrepancies should be interpreted in the context of a rapid engineering TLS workflow performed without registration targets or a control network, rather than as the intrinsic accuracy of the TLS instrument itself. The proposed hybrid validation framework provides an objective quality-control methodology for evaluating TLS-derived geometric axes against independent geodetic observations and may support reliability assessment of TLS-based inventories and deformation monitoring of slender engineering structures. Full article

Under complex environmental disturbances and visual measurement noise, globally coherent components, damage-induced local anomalies, and random disturbances are coupled in full-field vertical displacement responses. This coupling limits conventional low-rank–sparse decomposition because of its lack of mechanics-based constraints and can obscure weak damage-induced anomalies. To address this issue, this study proposes a physics-regularized low-rank–sparse damage identification method incorporating a physics prior derived from curvature-strain-energy perturbation. The method first extracts deflection curvature from the full-field displacement responses of the healthy and damaged states. A normalized physical evidence field is then constructed from the curvature-energy difference through Gaussian spatial regularization and mapped into spatially varying sparsity weights to modulate anomaly separation. Subsequently, the Physics-Regularized Differential Damage Index (PRDDI) is constructed from the difference in physics-regularized sparse anomaly intensity between the two states for damage localization and severity-sensitive characterization. The proposed method is primarily intended for beam-like structures satisfying the small-deformation bending assumption. For more complex structures, such as continuous beams, frames, plates, and shells, the corresponding mechanics-based physical evidence and spatial neighborhood relationships can be extended according to their load-transfer mechanisms and spatial geometries. Experimental and numerical results show that the peak-to-background ratio of the physics-regularized sparse anomaly field reaches approximately 2.77 times that of conventional robust principal component analysis (RPCA), while the background level is reduced by approximately 60%, and spurious peaks in non-damaged regions are markedly suppressed. For local stiffness reductions of 5–30%, the PRDDI localization error remains within 0–1 spatial measurement points. Both the peak value and local integrated area within the damaged region increase consistently with the degree of stiffness reduction, with coefficients of determination $R^2$ exceeding 0.99 and Spearman rank correlation coefficients of 1.00. For representative dual-damage cases, the proposed method maintains good dual-peak resolution. Under 10 dB noise, the complete dual-damage detection rate is approximately 87%, while the missed-detection rate for weak damage is approximately 10%. The physics prior derived from curvature-strain-energy perturbation improves consistency with structural mechanics, spatial separability, and the identification reliability of local damage anomaly extraction under complex measurement conditions. By exploiting spatially continuous, vision-based full-field displacement measurements, the proposed method can identify local damage regions in bridges and characterize variations in damage severity, providing a basis for subsequent detailed inspection and condition assessment. Full article

This paper examines the predictive power of credit default swap (CDS) spread correlations in the context of the 2023 Credit Suisse failure. Using a two-window design, we separate a 50-week pre-crisis period from the final two-week “jump” window and evaluate whether historical market-implied dependence anticipated cross-sectional crisis repricing across Global Systemically Important Banks (G-SIBs). We find that the magnitude of each bank’s crisis-period CDS jump is significantly related to its prior co-movement with Credit Suisse across the full sample ($r=0.80$, $p<0.001$, $n=15$), indicating that contagion followed a structured dependence pattern rather than an undifferentiated panic dynamic. The relationship holds across both regional cohorts, with the European G-SIB group displaying a considerably tighter fit ($r=0.96$, $p<0.001$, $n=8$) than the non-European group ($r=0.84$, $p=0.019$, $n=7$), consistent with geographic and institutional proximity to Credit Suisse amplifying the contagion channel. Additional empirical outputs, including stepwise-regression diagnostics and placebo/event-time checks, support the interpretation that the estimated relationship contains an economically meaningful signal while remaining partly event-driven in short horizons. Overall, the evidence suggests that rolling CDS dependence regimes may serve as a useful leading indicator for identifying institutions most likely to face disproportionate repricing pressure during a localized systemic shock. These findings are drawn from a single crisis episode and 15 peer institutions; they should be read as preliminary evidence of a potentially useful mechanism rather than as the basis for an operational early-warning system, and replication across additional crises and institutional settings is required before broader generalization. Full article

Non-uniform illumination in tunnel environments severely degrades image quality, posing substantial challenges to visual monitoring and intelligent transportation systems. While histogram equalization (HE) remains prevalent due to its computational simplicity, its non-linear pixel transformations frequently induce over-enhancement, artifacts, and structural distortions. This paper proposes Prior-Guided Histogram Equalization (PGHE), a lightweight enhancement framework that integrates conventional HE with Retinex-based illumination priors. Within the Retinex decomposition paradigm, PGHE constructs a contrast illumination map from the ratio between the HE-enhanced image and the original input. A Prior Correction Module (PCM) subsequently refines this map via relative total variation regularization, thereby restoring spatial coherence and alleviating local discontinuities introduced by HE. The corrected map is then applied to the original image to obtain the final enhanced result. Extensive evaluation on the LOL low-light benchmarks and a proprietary tunnel dataset comprising 247 real-world frames shows that PGHE offers favorable trade-offs among contrast enhancement, structural fidelity, and brightness preservation: it is particularly strong in brightness preservation and Entropy, while its PSNR/SSIM on LOL and its NIQE on the tunnel dataset are comparable to, but not always the best among, the compared methods. Furthermore, the proposed PCM functions as a plug-in module that improves existing HE variants with measurable gains in Structural Similarity and perceived naturalness at a modest cost in Absolute Mean Brightness Error. Full article

Rift Valley fever virus (RVFV) is a zoonotic, vector-borne disease affecting livestock and humans. The consumption of raw unpasteurized milk has been epidemiologically linked to RVFV exposure in humans and recognized as a potentially important non-vector route for RVFV transmission. This study aims to investigate the potential of RVFV transmission to humans through the consumption of infected raw milk. For this purpose, we conducted agent-based stochastic simulations to assess the potential for RVFV transmission through ingestion of contaminated raw milk. We created synthetic populations of cows, humans, and mosquitoes representing local farm dynamics in Kisumu, Kenya. A total of 1000 farms were simulated with binomial and Poisson distributions utilized to generate cow and human populations, respectively. Given our prior field data, we varied the probability of consuming raw milk from 1% to 80%. We also varied the probability of infectious raw milk causing RVFV infection at 10% and 40%. Results indicate that when the probability of infectious raw milk causing an infection is set at 10%, the cumulative incidence of RVFV infections in humans through raw milk consumption varies between 0.4 and 13% during the inter-epidemic periods. When the probability of infectious raw milk causing an infection is set at 40%, the percentage of cumulative milk infections ranges from 3.7 to 28%. Findings highlight the great potential for RVFV infection and spread through infectious milk consumption and the need for further evaluation of milk as a potential pathway of RVFV transmission to humans. Raw milk consumption in RVFV endemic settings remains an ongoing public health threat. Full article

The generation of guided trajectories for autonomous vehicles in multi-vehicle interaction scenarios remains challenging because surrounding vehicles continuously adapt their actions under uncertainty. This paper proposes a game-theoretic reinforcement learning (GT-RL) framework that combines a posterior-weighted rolling-horizon local game with proximal policy optimization (PPO) trajectory refinement. The revision makes the incomplete-information cost explicit through a normalized Bayesian posterior, a posterior expected cost, and a certainty-equivalent numerical approximation used by the SQP best-response solver. Using the supplied run-level logs (500 runs per method and scenario), GT-RL achieved mean safety scores of 94.1% (95% bootstrap CI: 93.91–94.28) in the intersection scenario and 91.5% (91.33–91.66) in the highway-merging scenario, with zero collisions in both sets of 500 runs. Paired comparisons with DQN, PPO, MPC-only, and potential-field baselines were significant after Holm correction (p < 0.001) for safety, traversal time, and comfort; DQN and PPO were faster in some cases but had lower safety and comfort. At the intersection, the rule-based method had higher safety but required 4.2 s more traversal time and had a 7.8-point lower comfort score than GT-RL. The findings support game-theoretic reasoning as a strategic prior for local interaction-aware trajectory generation. Claims are restricted to the tested low-to-moderate-speed, non-limit-handling simulations; high-fidelity vehicle dynamics and empirical large-scale timings remain areas of future study. Full article

Star trackers commonly suffer from star point trailing during stellar imaging under dynamic observation conditions. Traditional non-blind deconvolution methods rely on a known Point Spread Function (PSF), whereas blind deconvolution approaches are plagued by a complex solution space and a high tendency to fall into local optima. These drawbacks make it difficult to meet the requirements of high-precision star centroid extraction. To address these challenges, this paper proposes a novel blind restoration method based on physical model guidance and alternating iterative optimization. Firstly, the parameters of the blurred PSF are blindly estimated using image moment analysis, and a motion blur physical model with controllable direction and length is constructed. Secondly, the iterative ideal physical model is embedded as a strong prior into curvature filtering to achieve guided denoising, which effectively suppresses noise while maintaining the original trailing structure. Finally, a dual-layer alternating optimization framework grounded in the ideal physical model was developed. The inner layer employs the Richardson–Lucy (RL) algorithm integrated with intelligent convergence criteria for high-precision image restoration. The outer layer utilizes a gradient descent algorithm equipped with a confidence-based full step-length strategy to optimize PSF parameters. This architecture establishes a self-correcting closed-loop mechanism characterized by iterative image restoration–model refinement cycles. The simulation results demonstrate that the proposed method generally maintains the star centroiding error below 0.1 pixel without prior knowledge of the PSF, with a maximum observed error of 0.105 pixel under the most challenging high-background condition. Its performance is close to non-blind restoration and significantly outperforms traditional blind deconvolution algorithms. It provides an effective solution for high-precision star centroiding under dynamic conditions. Full article

Electroencephalography (EEG)-based emotion recognition is an important biosensing technique for affective brain-computer interfaces (BCIs), mental-state assessment, and physiological monitoring. Existing methods often rely on a single spectral descriptor or regular two-dimensional brain maps, which makes it difficult to jointly model local spatial representations, global spatial relations, and temporal dynamics. This paper proposes GLSTNet, a global-local spatial relations and temporal dynamics network for EEG emotion recognition. EEG trials are divided into short windows, from which multi-band spectral features are extracted and arranged into compact spatial maps. The local spatial encoder (LSE) learns local spatial and spatial–spectral representations from these compact multi-band spatial maps. The global spatial-relation encoder (GSRE) models long-range spatial relations between non-adjacent electrodes using a Pearson correlation prior and a learnable residual adjacency matrix. After local and global representations are integrated through gated fusion, the temporal dynamics encoder (TDE) models consecutive EEG windows using a gated recurrent unit with temporal attention. Comprehensive validation is conducted on two public EEG emotion datasets, the Database for Emotion Analysis using Physiological Signals (DEAP) and the SJTU Emotion EEG Dataset (SEED). In the subject-dependence setting, GLSTNet achieves 93.50 ± 3.22% accuracy for valence and 93.79 ± 3.64% accuracy for arousal on DEAP, and 92.48 ± 3.30% accuracy on SEED. In the subject-independence setting with target-subject calibration, GLSTNet obtains 75.61 ± 6.19% and 79.57 ± 5.99% accuracy for DEAP valence and arousal, respectively, and 88.22 ± 4.70% accuracy on SEED. These results indicate that integrating global-local spatial relations with temporal dynamics provides an effective representation strategy for EEG-based emotion recognition. Full article