Search Results (359)

Automatic detection of stress in social media text holds promise for supporting digital mental health, but most existing Transformer-based approaches are opaque and computationally demanding. This work presents DR-Transformer, a Dual-Regularized Transformer that combines two complementary mechanisms: (i) a group sparsity penalty ($L_{2,1}/L_2$ elastic net) applied to the query and key projection matrices of every attention head, which encourages whole-row sparsity, producing more concentrated and inspectable attention patterns; (ii) a supervised contrastive loss on the [CLS] projection, which organizes the latent space according to the stress label. The architecture is intentionally lightweight (six layers, eight heads, 256-dim embeddings; ∼9.5 M parameters) and runs entirely on consumer-grade hardware (NVIDIA GTX 1660, 6 GB). Experiments on the publicly available Dreaddit dataset (binary stress classification, 2838 train/715 test segments) compare DR-Transformer against Logistic Regression, BiLSTM, a Standard Transformer of identical architecture, and MentalBERT. Across five seeded runs, DR-Transformer (Full) reaches $F_1=0.876$ (bootstrap 95% CI $0.852$–$0.898$), outperforming the Standard Transformer ($F_1=0.842$; McNemar $p<0.001$ with Bonferroni correction) and performing comparably to the much larger MentalBERT ($F_1=0.879$; $p=0.421$). Sparse regularization increases the fraction of near-zero attention weights (below $0.01$) from $0.215$ to $0.682$, while the supervised contrastive loss improves the silhouette score of [CLS] embeddings from $0.312$ to $0.483$. Dual regularization thus combines accuracy, efficiency, and structurally induced attention concentration in a single model which can be trained without specialized infrastructure. We use the term “interpretable” throughout in this restricted, structural sense—to refer to concentrated and inspectable attention—rather than in the sense of established causal or mechanistic faithfulness; this is only partially and indirectly supported by our token deletion analysis. Full article

Precise step length estimation (SLE) is a key necessity for not only navigation systems design but also gait health monitoring in neurological conditions. Among existing solutions, non-invasive inertial sensor-based approaches operating without dedicated infrastructure are more cost-effective. Many such methods, however, rely on bodily affixed inertial sensors rather than freely held smartphone sensors. Traditional signal processing approaches, on the other hand, offer varying accuracy across diverse gait patterns due to user parameter calibration. This study, thus, proposes a regression-based SLE framework employing eight regression algorithms: linear regression (LR), k-nearest neighbors, support vector machine, decision tree, elastic network, random forest, histogram-based gradient boosting (HGB) regressor, and artificial neural network (ANN). Their extensive and rigorous evaluations across varied window sizes, using a dataset collected in normal and fast walking modes with two device positions (hand-held and trouser-pocket) during three evaluation scenarios, demonstrate the HGB regressor’s outstanding performance, achieving the lowest mean absolute error (MAE) below 1 cm across four different contexts under leave-one-out cross-validation-based evaluation and three in the seen test evaluations. Moreover, the findings report the ANN’s exceptional generalization capacity over other models and the previous method IRT-SD-SLE in unseen test evaluations, with an MAE not exceeding $6.3$ cm. The extensive evaluations of training and testing times reveal the highest computational efficiency for LR, moderate efficiency for the HGB regressor, and the highest training cost for the ANN, indicating a clear trade-off between MAE and computational expense. Additionally, this study includes an insightful discussion on the performance results, including the trade-offs between accuracy and efficiency. Full article

Introduction. Oxidative stress plays a critical role in the pathogenesis of idiopathic pulmonary fibrosis (IPF), yet its prognostic significance remains unclear. This study investigated the association between plasma sulfhydryl (SH) group concentrations and thiobarbituric acid reactive substances (TBARS), systemic markers of oxidative stress, and mortality in patients with IPF. Materials and methods. Eighty-eight patients with newly diagnosed IPF were recruited between 2016 and 2023 for the purposes of the study. Plasma SH and TBARS were measured at baseline under standardized conditions and normalized to plasma protein content. Survival analyses were performed using Kaplan–Meier curves and Cox regression models, adjusting for lung function parameters and IPF stage. Results. Patients with lower SH concentrations had significantly higher mortality (log-rank p = 0.012). SH group concentrations, but not TBARS, were independently and negatively associated with survival in multivariate models adjusting for %TLC, %FVC, %DLCO, and IPF stage (HR: 0.606, 95% CI: 0.443–0.830, p = 0.0018). Conclusions. Low plasma SH concentrations, reflecting systemic redox imbalance, are independently associated with increased mortality in newly diagnosed IPF. SH quantification represents a promising prognostic biomarker in IPF. Full article

Background: Our research group has previously demonstrated the protective effect of short-term treatment with esmolol on large artery remodeling. However, whether this beneficial effect persists after treatment withdrawal remains unknown. Therefore, the aim of the present study was to investigate whether regression of thoracic aorta remodeling after a short-term esmolol treatment persists following drug withdrawal. Methods: Adult male spontaneously hypertensive rats (SHRs) received either esmolol (300 µg/kg/min) or vehicle (saline solution) by continuous infusion for 48 h. Following treatment, animals were evaluated either immediately (SHR-E 48 h group) or after withdrawal periods of 7 days (SHR-E 7 d group) or 1 month (SHR-E 1 m group). Hemodynamic parameters, thoracic aorta geometry, extracellular matrix composition (elastin and collagen), and the passive mechanical response of the arterial wall (β parameter) were assessed in all animals. Results: Forty-eight hours of esmolol treatment significantly reduced blood pressure and attenuated thoracic aortic wall thickness, external diameter, cross-sectional area, collagen content and elastic fiber density. Additionally, passive mechanical testing showed a significant reduction in the β parameter, indicating improved arterial compliance after treatment. Remarkably, these structural and mechanical benefits persisted for up to one month after withdrawal, despite the return of blood pressure to hypertensive levels. Conclusions: Protective effect of esmolol on aortic remodeling is persistent after treatment withdrawal in SHRs, perhaps independent of blood pressure. Full article

Tomato leaf disease detection in natural environments is challenged by subtle early-stage symptoms, complex backgrounds, leaf occlusion, and scale variation, which can lead to missed detections, false detections, and unstable leaf localization. Meanwhile, practical agricultural applications impose higher requirements on model lightweightness and edge-deployment capability. To address these issues, this study proposes DFR-YOLOv12n, a lightweight tomato leaf disease detection model based on YOLOv12n that integrates detail-preserving downsampling, feature enhancement, and regression optimization. First, a multi-source dataset collected in natural environments was constructed and curated, covering eight categories: bacterial spot, early blight, late blight, leaf mold, mosaic virus disease, septoria leaf spot, yellow leaf curl virus disease, and healthy leaves. Second, SPDConv was introduced into key downsampling layers to preserve fine-grained disease-related visual cues. The A2C2f_DEConv module was incorporated into the P3 feature fusion branch to enhance leaf texture and disease-related appearance features under complex backgrounds. In addition, MPDIoU was adopted to optimize bounding box regression and improve whole-leaf localization under occlusion and background interference. The optimal model configuration was determined through insertion-position, module comparison, and ablation experiments. Compared with the baseline model, DFR-YOLOv12n increased Precision, Recall, and mAP@0.5 from 86.8%, 76.9%, and 86.5% to 88.1%, 81.7%, and 88.6%, respectively. Meanwhile, FLOPs decreased from 5.83 G to 5.27 G, the parameter count decreased from 2.51 M to 2.25 M, and the model size decreased from 5.22 MB to 4.71 MB. Furthermore, the model was successfully deployed and validated on the Jetson Nano platform, demonstrating its potential for edge applications. The results indicate that DFR-YOLOv12n achieves a favorable balance among detection accuracy, model complexity, and deployment feasibility, providing a reference for intelligent tomato leaf disease detection in natural environments. Full article

Optimal harvesting time is crucial for the quality and yield of extra virgin olive oil. Despite the widespread use of visual ripeness, maturity index, and pulp hardness, no study has compared their predictive ability for the quality and production parameters of extra virgin olive oil at industrial scale, thus motivating this work. Fifteen 300 kg olive samples (five replicates per harvesting date) were collected on 11 October, 29 October, and 25 November 2024 and processed within six hours at an industrial olive oil mill. One-way analysis of variance and Multiple Ordinary Least Squares regression were applied to evaluate harvest maturity indicators against primary quality, phenolic profile, yield, temperature, and sensory parameters. The former confirmed a significant main effect of harvesting time for most parameters, while the latter showed that visual ripeness systematically outperformed the maturity index and pulp hardness in seven out of eight models; pulp hardness was the sole exception, proving the strongest correlation with final oil temperature. These findings confirm that visual ripeness—a cost-free, immediately applicable parameter—seems to be the most strongly correlated indicator of the quality, phenolic content, and sensory attributes of extra virgin olive oil, while pulp hardness emerges as an innovative operative tool for optimizing temperature management in the production of extra virgin olive oil. Full article

Background: Idiopathic pulmonary fibrosis (IPF) and progressive pulmonary fibrosis (PPF) are defined and monitored without definite exercise-based criterion, despite exertional desaturation being among the earliest functional signs of disease and an established independent predictor of mortality in IPF. The 6 min walk test (6MWT) provides robust prognostic information in IPF, yet whether it can kinetically distinguish IPF from PPF under sustained submaximal loading has not been systematically examined. We hypothesised that second-by-second oximetry during the 6MWT and the 1 min (min) sit-to-stand test (1STST) would expose phenotype-specific kinetic signatures invisible to static endpoints, and that test modality would matter. Methods: Fifty-one patients with IPF, 12 with PPF, and 100 with non-IPF/non-PPF ILD (reference cohort) from the European ILD Registry (eurILDreg) participated in this pilot study and completed both tests with continuous 1 Hz SpO₂ and pulse-rate recording. Phenotype-specific desaturation and recovery slopes were derived from random-effects panel regression models. Multivariable linear regression tested whether IPF and PPF phenotypes carried kinetic and static-endpoint signatures independent of DLCO, age, sex, and BMI. Results: Despite substantially lower DLCO in fibrosing phenotypes (IPF 45 ± 17%, PPF 38 ± 11%, reference 55 ± 19%; p < 0.001), conventional exercise performance metrics did not differ significantly between groups (6MWD p = 0.099; 1STST repetitions p = 0.351). The 6MWT produced statistically indistinguishable per-second desaturation slopes in IPF and PPF (both ≈ −0.016%/s versus −0.011%/s in the reference), whereas the 1STST exposed a clear phenotype gradient (PPF −0.044%/s, IPF −0.027%/s, reference −0.016%/s; a 2.75-fold spread). PPF additionally showed a blunted chronotropic response during the 6MWT (PR slope +0.014 vs. +0.032 and +0.033 bpm/s in IPF and reference). Likelihood-ratio tests confirmed significant phenotype effects on seven of eight time-resolved trajectories (all p < 0.001). IPF was independently associated with greater cumulative oxygenation deficit during the 1STST (SpO₂ AUC β = +669, p = 0.022), indicating excess dynamic burden beyond what diffusion capacity predicts. Conclusions: In this pilot analysis, continuous high-resolution oximetry identified two phenotype-related signals that were robust to the principal confounders. IPF was independently associated with a greater cumulative oxygenation impairment during the 1STST after DLCO adjustment, and PPF showed a blunted chronotropic response during sustained walking that, being pulse-rate based, was unaffected by supplemental oxygen and was not attributable to pulmonary hypertension. A phenotype gradient in per-second desaturation measurements during the 1STST was also observed but should be interpreted as hypothesis-generating, as the small PPF subgroup (n = 12), its greater disease severity, and supplemental oxygen use in half of its patients preclude firm attribution to phenotype. Taken as exploratory, these observations support the prospective evaluation of kinetic exercise parameters as candidate monitoring components for PPF. Full article

The use of supplementary cementitious materials such as fly ash can reduce environmental impacts and improve the sustainability of concrete construction. However, the nonlinear interactions among mixture design parameters make accurate prediction of concrete compressive strength challenging. In this study, TabPFN, a pre-trained foundation model for tabular data, was applied to predict the compressive strength of fly ash concrete and compared with tuned Random Forest, support vector regression, an artificial neural network, LightGBM, CatBoost, Ridge regression, and Abrams empirical regression. A dataset containing 1062 samples and eight mixture-level variables was used for model development and evaluation. Predictive performance was assessed using the coefficient of determination, mean absolute error, and root mean square error over 100 repeated random splits. The results showed that TabPFN achieved the best overall performance, with an average coefficient of determination of 0.9329, a mean absolute error of 3.2758 MPa, and a root mean square error of 4.6678 MPa. Compared with the strongest tuned gradient-boosting baseline, CatBoost, TabPFN reduced the mean absolute error and root mean square error by 0.8768 MPa and 0.8560 MPa, respectively. Furthermore, repeated-split conformal prediction demonstrated reliable uncertainty quantification, with an average prediction interval coverage probability of 0.9615 and a mean prediction interval width of 23.4554 MPa. SHAP analysis identified the water-to-cement ratio, mortar strength, and water-to-binder ratio as important variables, while additional multicollinearity and feature ablation analyses indicated that correlated ratio variables should be interpreted cautiously. The results indicate that TabPFN provides an accurate, robust, and uncertainty-aware framework for preliminary prediction of 28-day fly ash concrete compressive strength. Full article

Objectives: To investigate the relationship between intrapartum cardiotocography (CTG) findings and neonatal outcomes by evaluating the association between CTG tracings, umbilical cord blood gas parameters, and APGAR scores, and to assess the diagnostic performance of CTG in identifying fetuses at risk of neonatal acidemia. Methods: This prospective study included women who delivered at a tertiary teaching hospital between January 2017 and January 2018. Of 1100 women initially screened, 596 met the inclusion criteria and were analyzed. Among them, 162 underwent operative delivery for non-reassuring fetal status (NRFS), while 434 served as controls. Demographic, obstetric, and neonatal characteristics were compared. Multivariable logistic regression was performed to explore factors associated with umbilical arterial pH < 7.20. Results: Mean umbilical arterial pH was significantly lower in the NRFS group than in controls (7.30 ± 0.08 vs. 7.32 ± 0.07, p < 0.05). Umbilical arterial pH differed significantly across NICHD fetal heart rate categories, with the highest values observed in Category I tracings. Twenty-eight neonates (4.7%) had an umbilical arterial pH < 7.20. CTG-based identification of NRFS predicted umbilical arterial pH < 7.20 with a sensitivity of 42.9%, specificity of 73.6%, positive predictive value of 7.4%, and negative predictive value of 96.3%. In multivariable analysis, pregestational diabetes mellitus and preeclampsia were independently associated with umbilical arterial pH < 7.20. Conclusions: Although CTG remains an essential tool for intrapartum fetal surveillance, its low positive predictive value indicates that many fetuses classified as having NRFS do not have biochemical evidence of acidemia. CTG findings should therefore be interpreted together with the overall clinical context, including maternal risk factors such as pregestational diabetes mellitus and preeclampsia, rather than being used in isolation to guide intrapartum management. Full article

Accurately predicting the degree of hydration in cement systems blended with slag and fly ash is difficult when experimental data are limited. Empirical kinetic models typically need recalibration for each specific blend, while purely data-driven approaches often fail to maintain physical consistency. In this work, we present a physics-informed neural network for predicting isothermal hydration by incorporating the Avrami–Erofeev–Arrhenius ordinary differential equation into the training loss and using a composition-focused sub-network to relate blend proportions to blend-specific kinetic parameters. The model was trained and tested on 29,379 calorimetry data points from 77 multi-component cement systems gathered from two open-access datasets. For previously unseen test systems, it reached an $R^2$ of 0.9864 and a symmetric mean absolute percentage error (sMAPE) of 25.3%. Its overall sMAPE was lower than that of a neural network with the same architecture but without physics-based constraints (34.9%); stage-resolved analysis showed that the largest difference occurred during early hydration. Across eight random seeds, the physics constraint did not confer a data-efficiency or training-stability advantage over the same-architecture network. The distinguishing features of the PINN are instead its explicit ODE and monotonicity regularization and its composition-conditioned effective parameters. Although random forest regression produced lower pointwise error, it failed to maintain physically consistent hydration-rate behavior. Overall, the framework provides a composition-conditioned surrogate that incorporates kinetic regularization within the composition range investigated. Full article

Background/Objectives: The heterogenous nature of renal cell carcinomas (RCCs) is increasingly recognized. The purpose of this proof-of-concept pilot study was to evaluate correlations between multiparametric MRI (mpMRI)-derived and histopathological parameters in RCCs from spatially matched regions on both MRI and pathological examination to support targeted biopsy planning. Methods: In this prospective single-center pilot study, patients with solid renal tumors ≥2 cm undergoing nephrectomy were prospectively enrolled. Each patient underwent preoperative 3.0T-mpMRI including T2-weighted and pre-/post-contrast T1-weighted sequences, chemical-shift imaging, IVIM-DWI, and T1/T2*/R2 mapping. Tumor regions were defined jointly by a pathologist and radiologist, and identical regions of interest were assessed for each tumor region across all sequences to gain quantitative mpMRI-derived parameters. Histopathology provided quantitative regional fractions of viable tumor, fibrosis, hemorrhage, and cystic/necrotic components. Spearman’s rank correlations and univariable linear regression assessed associations between mpMRI and histopathological parameters on a regional level. Results: Across 49 tumor regions in eight patients (65.3% clear cell, 34.7% papillary RCCs), the mean viable tumor fraction was 80.9% (SD 17.6). The viable tumor fraction showed inverse correlations with nephrographic and delayed phase signal intensity changes (rho = −0.59/rho = −0.51), T1 values (rho = −0.56), true diffusion coefficient D (rho = −0.47), and ADC (rho = −0.45), and a positive correlation with R² times (rho = 0.55). Delayed and nephrographic phase signal intensity changes (R² = 0.41/R² = 0.39) were the strongest single exploratory imaging correlates of viable tumor fraction. Conclusions: These findings support the feasibility of quantitative mpMRI parameters to capture regional intratumoral heterogeneity in RCCs, thereby highlighting regions with high viable tumor burden, which may help to refine the imaging-based assessment of RCCs in the future. Full article

Accurate prediction of surface roughness and cutting forces in milling aluminum alloys remains challenging under data-scarce conditions, where limited experimental data restricts the application of conventional machine learning models. This study addresses this gap by developing a systematic machine learning framework using 108 milling experiments (repeated to 216 tests) on aluminum alloys AA2024-T351 and AA6061-T6. Five primary machining inputs—material type, spindle speed, feed rate, depth of cut, and tool coating—were used. Through feature engineering, 35 interaction features were generated to capture non-linear relationships. A two-step preprocessing strategy was applied: Winsorization at the 5th and 95th percentiles to handle outliers, followed by hybrid scaling combining RobustScaler and MinMaxScaler. Eight machine learning algorithms, including XGBoost, NGBoost, LightGBM, CatBoost, Random Forest, MLP, SVR, and Least Squares Boosting, were developed and hyperparameter-optimized using the Optuna framework with Tree-structured Parzen Estimator. Models were evaluated using R², MAE, and RMSE on a 70/15/15 train–validation–test split. Results demonstrate that XGBoost achieved the highest predictive accuracy for surface roughness (Ra) (R² = 0.99829) and for resultant cutting force (FN) (R² = 0.997). Feed rate was identified as the dominant machining parameter, accounting for 87.7% of the total importance in predicting surface roughness. SHAP analysis confirmed that engineered interaction features—particularly Feed_Coating and Material_Feed—carry strong physical relevance. Additionally, NGBoost enabled probabilistic regression, providing uncertainty estimates. The proposed framework proves highly effective for multi-output prediction in machining under limited data, offering a robust, interpretable, and industry-ready solution for quality control in aluminum alloy milling operations. Full article

Literature-based compilation indicates that there have been 165 earthquakes with moment magnitude M_w ≥ 7.0 in the period 1980–2024 for which identifiable source-parameter estimates are available. Despite the relative scarcity of such events globally, this dataset is sufficiently large to perform a robust rupture analysis. Using ordinary least-squares regression, we characterized the relationships between M_w and eight source parameters: rupture length (L), rupture width (W), rupture area (S), maximum slip (D_max), average slip ($\stackrel{- -}{D}$), average rupture velocity (V_r), source-averaged static stress drop (Δσ), and rupture duration (T_R). The logarithms of rupture length, width, area, maximum slip, and average slip generally showed statistically significant linear relationships with M_w. In contrast, average rupture velocity and static stress drop showed weak and mostly statistically insignificant relationships with M_w. Rupture duration was more strongly correlated with M_w for reverse/thrust- and normal-fault earthquakes than for strike-slip earthquakes. Strike-slip earthquakes had a mean length-to-width ratio of 3.97, compared with 2.05 and 2.43 for reverse/thrust- and normal-fault earthquakes, respectively. For the 25 events with both maximum- and average-slip estimates, the ratio D_max/$\stackrel{- -}{D}$ ranged mainly from 1.5 to 3.5, with a mean of 3.18, indicating spatially heterogeneous slip distributions. Because the dataset was compiled from studies using different source models and inversion methods, the resulting equations should be interpreted as dataset-specific empirical relationships. Full article

Individual biochemical indicators are insufficient for comprehensive tomato food flavor quality assessment, necessitating multi-parameter models of the core soluble taste matrix. We hypothesized that age stratification of trained sensory assessors would expose differential biochemical variable importance profiles in flavor quality prediction. Accordingly, this study aimed to: (1) construct and compare multiple regression models linking eight biochemical indicators to sensory scores, (2) identify key quality drivers via feature selection, and (3) examine whether age stratification alters the identified sensory drivers. Eight baseline taste indicators across 62 tomato cultivars were evaluated by 30 age-stratified trained sensory panelists (<40 and ≥40 years), using cross-validation to ensure model robustness against small-sample constraints. Partial least squares regression (PLSR), support vector regression (SVR), random forest (RF), and Boruta were applied. Random forest achieved the best performance (R² = 0.82). In the full panel model, key variables were fructose, total free amino acids, and vitamin C. After age stratification, the under-40 group retained these variables, whereas the ≥40 group replaced vitamin C with soluble solids. Fructose and total free amino acids were consistently robust drivers, while total acidity remained least important. Deploying the RF–Boruta framework within an age-stratified context provides a structured analytical framework for investigating flavor perception from biochemical data. These findings suggest that fructose and total free amino acids represent highly robust candidate indicators for flavor quality prediction, while age-stratified variances suggest the utility of integrating demographic-specific metrics into precision breeding frameworks. Full article

Background/Objectives: Optical coherence tomography (OCT) and OCT angiography (OCT-A) provide detailed measurements of retinal structure and vasculature; however, age-related differences in how these parameters correlate with one another remain poorly understood. We hypothesized that vascular–structural integration in the macula is more pronounced in adults than in children. Our aim was to characterize correlation patterns in pediatric and adult populations to inform the development of age-specific clinical interpretation guidelines. Methods: This prospective cross-sectional observational study enrolled 37 healthy children (age 1–17 years) and 28 healthy adults (age 18–65 years). Eyes with ocular or systemic conditions affecting the retina or prior intraocular surgery were excluded. Standardized OCT and OCT-A acquisition protocols provided structural and vascular measures. Univariable correlation analyses applied a stringent threshold (p < 0.001) to identify robust associations. Significant univariable results were entered into multivariable regression models adjusting for age, gender, intraocular pressure, and axial length. A Group-wise Linkage Proportion quantified the percentage of potential significant correlations among eight predefined anatomical parameter groups. Results: Ninety univariable correlations met p < 0.001. Fourteen correlations were shared across age groups, notably foveal avascular zone metrics and vessel density, showing very large negative correlations (r = −0.70 to −0.87). The pediatric cohort displayed 40 unique correlations, primarily linking optic nerve head flow indices to retinal nerve fiber layer thickness. Adults exhibited 36 unique correlations, dominated by macular vascular–thickness coupling concentrated in the parafoveal region. After multivariable adjustment, 52 of 90 associations remained significant. Adult-specific associations lost significance more frequently (58%) than pediatric-specific associations (43%), whereas correlations shared across both groups showed complete stability (100%). The Group-wise Linkage Proportion indicated pronounced macular vascular–structural coupling in adults (48.4%) versus near absence in children (1.2%). Conclusions: Retinal parameter correlation patterns show fundamental differences between pediatric and adult eyes. While optic nerve head-macular thickness relationships remain consistent across ages, adults exhibit mature, localized integration of macular vascular and structural parameters absent in children. These findings suggest that pediatric and adult OCT/OCT-A measurements may benefit from separate reference standards, although prospective validation is required before clinical implementation. Full article