Search Results (3,006)

Existing randomness conditioning methods face a fundamental dilemma: one must choose between heuristic post-processing, which is practical but lacks rigorous mathematical uniformity guarantees, and provable seeded extractors, which offer information-theoretic guarantees but require an independent perfectly uniform seed—thereby shifting, rather than eliminating, the underlying trust assumption. Neither option alone achieves provable uniformity from a raw physical source. We present a universal, mathematically certified conditioning framework that resolves this dilemma. For any NIST SP 800-90B ESV-certified entropy source, regardless of bias or implementation, our framework generates a provably uniform output stream without requiring any external seed or heuristic whitening. The core contribution is the Geometric Convergence Theorem (GCT), proving that the modular reduction of a negative binomial counting variable $N_p\sim \mathrm{NB}(m,p)$—where m denotes the required number of successes generated from fixed ESV entropy blocks via Bernoulli trials with success probability p—converges exponentially to uniformity over ${\mathbb{Z}}_R$, with spectral radius ${\rho }_{\mathrm{NB}}=p/\sqrt{p^2+4(1-p){\sin }^2(\pi /R)}<1$. A Practical Entropy Budgeting mechanism ensures information-theoretic entropy conservation via a fixed input–output ratio. In a large-scale validation generating 100 MB of output from a biased ESV source ($H_{\mathrm{in}}=3.32$ bits/byte), the framework achieved Shannon entropy $7.999998$ bits/byte and min-entropy $7.9936$ bits/byte, approaching the theoretical lower bound of $7.9949$ bits/byte to within $0.0013$ bits/byte, with ${\chi }^2=275.95$ (df = 255). This establishes the first seedless, provable, and platform-agnostic conditioning framework for certified entropy sources. Full article

Preschoolers’ externalizing problem behaviors have a profound negative impact on their social adjustment, both concurrently and in the future. Household chaos is a key environmental risk factor; however, the mechanisms underlying this link require further investigation. The current study aimed to examine associations between household chaos and externalizing problem behaviors in preschool children and the roles of callous–unemotional traits and maternal mindful parenting. The participants were 615 children aged 3–7 years (341 boys and 274 girls) and their mothers from Guangdong, China. Data on household chaos, maternal mindful parenting, children’s callous–unemotional traits, and externalizing problem behaviors were collected using mother-reported questionnaires. Results indicated that household chaos was significantly and positively related to externalizing problem behaviors in young children, with callous–unemotional traits acting as a mediator. Maternal mindful parenting was also found to moderate the association between callous–unemotional traits and externalizing problem behaviors. These findings emphasize the necessity of optimizing the home environment and providing mindful parenting training in preventing and intervening in early childhood socioemotional and behavioral problems. Full article

Multimodal time-series anomaly detection incorporates textual data regarding external events to provide contextual information for numerical fluctuations; this facilitates the interpretation of sudden fluctuations in complex systems—such as those caused by disasters, public events, or environmental changes—and enables the identification of numerical fluctuations that cannot be explained by the available context. However, existing methods typically treat text as an auxiliary feature for enhancement while still relying primarily on the degree of deviation in numerical sequences to detect anomalies; consequently, they struggle to distinguish between explainable fluctuations driven by genuine external events and fluctuations that are unsupported by the available event semantics. Moreover, redundant, mismatched, or weakly correlated text can disrupt the modeling of normal temporal patterns, potentially triggering cross-modal negative transfer. To address these issues, we propose SISAD, a multimodal time-series anomaly-detection method based on semantic interference suppression. First, the method employs trend decoupling and multiscale encoding to extract local temporal variation patterns, thereby mitigating the impact of low-frequency trends on the modeling of localized or fine-grained temporal variations. Second, it constructs a prototype-based semantic fusion module that integrates external event text, endogenous statistical text, and domain prior knowledge, enhancing the semantic interpretability of the current segment by retrieving learned historical prototypes. Finally, a semantic interference suppression module is designed to measure the degree of cross-modal conflict and adaptively adjust the text fusion intensity; this allows relevant text to contribute to reconstruction and discrimination while attenuating interference from unreliable text. Experiments conducted on seven datasets spanning three domains demonstrate that SISAD outperforms existing methods across most datasets and key metrics, showing improved performance in distinguishing fluctuations with relevant contextual support from those that cannot be explained by the available context. Ablation and text-perturbation experiments further support the contributions of prototype-based semantic fusion and semantic interference suppression with the evaluated settings. Full article

Background: Because an uncovered arch scaffold does not provide a covered seal across an unresected arch tear, we evaluated a source-adjudicated sequential eligibility framework separating study-defined stop gates from residual-risk modifiers. Methods: We retrospectively analyzed 266 patients with acute DeBakey type I dissection treated by hemiarch repair alone (No. = 160) or hemiarch plus AMDS (No. = 106). Blinded preoperative CT-scan re-review validated Gate 1 (visible arch tear), and operative-report review confirmed Gate 2 (tear directly observed at open inspection). Supra-aortic false-lumen communication was assessed only after both gates were negative. Results: Sixty-nine patients (25.9%) passed both study-defined gates (AMDS = 32; hemiarch = 37). Twelve-month CT-scan was available in 45/69 (65.2%; AMDS 24/32, hemiarch 21/37). Among imaged patients, complete zone 2–3 false-lumen exclusion occurred in 20/24 (83.3%) after AMDS and 0/21 after hemiarch (risk difference +83.3 percentage points, 95% CI +58.7 to +93.3; p < 0.001). DANE occurred in 4/32 (12.5%) and 16/37 (43.2%), respectively (p = 0.007). Under the adverse opposing missingness scenario, the exclusion risk difference narrowed to +19.3 percentage points (95% CI −4.2 to +39.9; p = 0.148). Conclusions: These findings are hypothesis-generating and apply to a selected post hoc subgroup; treatment comparisons are non-causal. Prospective external validation of the study-defined hierarchy remains required. Full article

Background/Objectives: Early recognition of children with severe physiological illness is essential in emergency care. The Emergency Department Pediatric Early Warning System score (ED-PEWS) has not been evaluated in Vietnam. We assessed how well ED-PEWS recorded at presentation identifies children who meet criteria for resuscitation-defined critical illness in a high-acuity pediatric emergency setting. Methods: This prospective observational diagnostic classification study was conducted from August to October 2022 in the emergency and resuscitation department of a national tertiary children’s hospital in Vietnam. Children aged 1 month to 15 years were enrolled according to a prespecified shift-based schedule. The outcome was resuscitation-defined critical illness within 24 h (cardiac arrest or bradycardia with poor perfusion requiring cardiopulmonary resuscitation). Because the timing of scoring relative to interventions and outcome was not recorded, the analysis was interpreted as concurrent classification rather than prediction. Results: Of 805 children (median age 19 months; 65.3% male), 156 (19.4%) met the outcome criteria. The median ED-PEWS was 32 (interquartile range, 24–43) in children with the outcome and 10 (8–12) in those without (p < 0.001). The area under the receiver operating characteristic curve was 0.952 (95% CI, 0.926–0.977). At the exploratory Youden-derived threshold of ED-PEWS ≥ 16, there were 143 true positives, 49 false positives, 13 false negatives, and 600 true negatives; sensitivity was 91.7% (95% CI, 86.2–95.5), specificity 92.4% (90.1–94.4), and negative predictive value 97.9% (96.4–98.9). Conclusions: In this selected high-acuity cohort, ED-PEWS showed strong concurrent discrimination for resuscitation-defined critical illness. Because the score and the outcome share physiological components and the threshold was derived and evaluated in the same dataset, these findings are exploratory and require multicenter external validation and comparison with existing triage systems. Full article

Identifying bioactive compounds in traditional Chinese medicine (TCM) requires computational prioritization to be connected with chemical and pharmacological evidence. This structured narrative and critical review examines machine learning applications in compound prediction, formula-compatibility analysis, multi-omics integration, and drug development and repurposing. Molecular fingerprints, graph and sequence representations, knowledge graphs, and generative models are discussed in relation to their intended endpoints and validation conditions. Studies are classified along two independent dimensions: application scope, distinguishing direct TCM applications from natural-product studies and general drug-discovery methods, and validation level, ranging from computational prediction to chemical, experimental, and clinical evidence. Three case analyses illustrate a TCM-specific graph-learning platform with chemical confirmation, a downstream animal and metabolomics study without machine learning screening, and a proposed causal framework assembled from separate studies. Across these applications, machine learning supports candidate ranking and hypothesis generation, but predicted targets, attention weights, and multi-omics associations do not independently establish target engagement, formula synergy, or therapeutic efficacy. Major limitations include chemical-domain shift, inconsistent database provenance, data leakage, sparse negative results, and incomplete experimental follow-up. Reliable advancement requires scaffold-aware or external evaluation, explicit applicability domains and uncertainty, traceable data processing, and constituent-resolved intervention studies. A minimum reporting checklist consolidates these requirements, while route-specific assessment distinguishes synthetic accessibility from biosynthetic feasibility. Together, these considerations provide a basis for evaluating screening claims and designing reproducible workflows toward experimentally supported TCM compound discovery. Full article

Partial shading deforms photovoltaic (PV) current–voltage (I-V) characteristics and complicates reliable condition monitoring. This study develops and stress-tests a compact, feature-based machine-learning framework using a public model-generated I-V dataset. The primary cohort contains 186 curves at 1000 W m⁻², balanced across uniform operation and five nominal shading severities (10%, 20%, 30%, 40%, and 50%) at 31 temperatures from 20 to 50 °C. A verified set of 37 electrical, normalized, maximum-power-point, and curve-shape predictors was evaluated with six classifiers under 30 repeated stratified temperature-grouped five-fold validations, followed by nested grouped confirmation. The selected 10-feature logistic-regression model achieved 1.000 accuracy, balanced accuracy, macro-F1, Matthews correlation coefficient, Cohen’s κ, and ROC-AUC for binary and six-class diagnosis; the temperature-group bootstrap 95% confidence interval for macro-F1 was 1.000–1.000. A single normalized maximum-power-point-voltage predictor also achieved macro-F1 = 1.000, showing that the controlled source domain is strongly and partly rule-separable. The clean high-C solution was highly sensitive to measurement disturbance: under a 5% combined bias, drift, Gaussian, and quantization stress test, mean macro-F1 fell to 0.395 for binary detection and 0.095 for six-class diagnosis, whereas noise-aware training remained 0.976 and 0.962, respectively. Normalized maximum-power-point voltage, fill factor, maximum-power-point voltage, and maximum-power-point power were the dominant features. A uniform-operation negative control across 500–1000 W m⁻² also exposed irradiance-domain sensitivity in the noise-aware model. The results establish reproducible controlled-domain discrimination and uncertainty sensitivity, but external measured-data validation across irradiance, modules, array configurations, and shadow geometries remains necessary before deployment. Full article

Background/Objectives: Acute appendicitis is the most common surgical emergency of childhood, yet its diagnosis remains difficult because presentations are atypical, inflammatory markers are nonspecific, and the consequences of error run in both directions, from negative appendectomy to missed perforation. Over the past two decades, a large body of work has tried to support this decision with clinical scores, and more recently with machine learning, deep learning, and large language models. Methods: This critical methodological review synthesises that literature across the whole care pathway rather than the single question of binary diagnosis, using a reproducible search and a transparent study-level appraisal, and it therefore covers severity stratification, prediction of non-operative treatment response, imaging stewardship, and the postoperative course. We evaluate this evidence through the lens of contemporary methodological standards, in particular the Transparent Reporting of a multivariable prediction model for Individual Prognosis Or Diagnosis, updated for artificial intelligence (TRIPOD+AI), and the Prediction model Risk Of Bias Assessment Tool, updated for artificial intelligence (PROBAST+AI). Results: We describe the main families of models, from the Alvarado and Pediatric Appendicitis scores to random forests, gradient boosting, convolutional networks applied to ultrasound, and emerging generative models. Reported performance is dominated by discrimination while calibration, clinical utility, external validation, fairness, and reproducibility are reported inconsistently and are often absent. We argue that the recurring pattern of very high reported accuracy reflects methodological fragility more than clinical readiness. Conclusions: We offer a practical checklist for the critical appraisal of appendicitis prediction studies together with a research agenda aimed at closing the gap between a high area under the curve and safe use at the bedside. Full article

Objective: This study evaluated the diagnostic value of a combined SHOX2 and RASSF1A methylation assay (LungMe^®) in differentiating malignant from benign pleural and peritoneal effusions. Methods: A total of 696 patients with pleural or peritoneal effusions were prospectively enrolled. Effusion sediment samples were analyzed by liquid-based cytology (TCT), cell block histology, and methylation-specific PCR for both SHOX2 and RASSF1A. The final diagnosis was confirmed by clinical, imaging, and pathological follow-up. Results: Lung cancer was the predominant cause of malignant pleural effusions (71.96%), while malignant peritoneal effusions mainly resulted from ovarian, gastric, and liver cancers. The optimal ΔCt cutoffs for SHOX2 and RASSF1A were ΔCt ≤ 9 and ΔCt ≤ 12, respectively. The combined methylation assay achieved a sensitivity of 79.1% and specificity of 90.4% for malignant effusions. When combined with cytology, the overall sensitivity increased from 34.7% (TCT alone) to 91.7%, with a positive predictive value (PPV) of 93.9% and a negative predictive value (NPV) of 86.7%. The assay showed high sensitivity across multiple tumor types, including 100% for esophageal and breast cancers, and ≥80% for lung cancer, gastric cancer, lymphoma, and cholangiocarcinoma. Conclusions: The combined SHOX2 and RASSF1A methylation assay has high diagnostic value for benign–malignant discrimination of pleural and peritoneal effusions, improving diagnostic sensitivity when combined with conventional cytology. Thus, within the context of this single-center study, the LungMe^® assay demonstrates promise in expanding pan-cancer diagnostics, but will require further multicenter external validation and prospective clinical studies before routine clinical adoption. Full article

Purpose: To characterize short-term outer retinal changes in silicone oil-filled eyes after primary macula-involving rhegmatogenous retinal detachment (RRD) repair using en face and cross-sectional optical coherence tomography (OCT), and to examine the association of outer retinal features, including foveal ellipsoid zone rosettes (FEZRs), with 12-week postoperative best-corrected visual acuity (BCVA). Methods: This prospective observational study included 35 patients with primary macula-involving RRD who underwent pars plana vitrectomy with silicone oil tamponade with follow-up at 3, 6, and 12 weeks. BCVA, external limiting membrane (ELM)/ellipsoid zone (EZ) integrity, persistent subretinal fluid (PSRF), outer retinal folds (ORFs), and FEZR were evaluated. Results: Mean BCVA improved from 1.96 logMAR at baseline to 0.64 logMAR at 12 weeks, with progressive restoration of foveal ELM/EZ integrity. En face OCT showed high agreement with cross-sectional OCT for PSRF and ORF detection. PSRF was detected in 25.7%, ORFs in 5.7%, and FEZR in 22.9% of eyes; FEZR were more frequent in eyes with poor 12-week visual outcomes. Conclusion: Cross-sectional and en face OCT provided complementary information on early outer retinal remodeling after RRD repair with silicone oil tamponade. Progressive ELM/EZ restoration was associated with better 12-week BCVA, whereas FEZR presence was associated with worse 12-week BCVA. These findings support further prospective evaluation of FEZR as a potential early negative prognostic imaging marker. Full article

Cinematic rendering (CR) is a physically based three-dimensional (3D) computed tomography (CT) visualization technique that has been applied chiefly to the external surface morphology of vascular structures. Intraluminal CR has previously been described for the cardiac chambers as black-blood cinematic rendering (BBCR), in which the opacity assigned to the contrast-enhanced blood pool is suppressed. Here, we describe the adaptation of that principle to the pulmonary vasculature as a negative-contrast intraluminal preset, in which the opacified blood pool is rendered semi-transparent so that the vessel lumen is displayed as a low-signal channel delineated by the enhancing vessel wall. We report the acquisition and rendering conditions under which the appearance was obtained in a single illustrative case performed for chest pain and showing no pulmonary vascular pathology, together with the rendering platform, base preset, and lighting configuration; the attenuation-band settings that define the appearance were tuned for the individual examination rather than applied as a fixed parameter set, were not recorded, and cannot be supplied as numerical values, so a stepwise construction procedure is given in their place. Intermediate rather than maximal compartment-specific opacification was required for the transfer function to separate lumen from wall; the case presented was acquired in an early pulmonary arterial phase, and a late split-bolus acquisition is an alternative route to the same condition that was not used here. In the case presented, the lumen of segmental and up to third-order subsegmental pulmonary arterial and venous branches was visually demonstrated on qualitative assessment by a single reader, and the preset remained stable when applied unchanged across the reconstructed 40–80% R–R window for four-dimensional display, which is a partial rather than a complete cardiac cycle reconstruction. Preset construction was less successful at the central pulmonary arteries and veins, where dense opacification and proximity to the cardiac blood pool reduced the contrast available to the transfer function. This is a single-case technical description. It demonstrates that the appearance can be produced under defined conditions and sets out the procedure by which it was constructed; it does not establish that the exact parameter values are recoverable or transferable, and it does not establish diagnostic performance, which will require prospective comparison against conventional reformats in adequately powered studies. Full article

Sustainable tourism development depends on decoupling tourism growth from carbon emissions, yet whether tourism specialization follows an Environmental Kuznets Curve (EKC) in carbon emission intensity, and whether environmental fiscal expenditure can reshape that relationship, remains unsettled. Drawing on panel data for 30 Chinese provinces over the period 2007–2022, this study estimates two-way fixed-effects and spatial Durbin models. An inverted U-shaped relationship emerges as a point estimate, but it is not statistically robust: a formal Lind–Mehlum test does not reject the absence of an inverted U, and the relationship is identifiable only in specific subsamples. This pattern is not driven by the COVID-19 period: excluding the pandemic years (2020–2022) preserves both the sign pattern and the location of the turning point. Its turning point lies far beyond the level of tourism specialization that most provinces actually reach, so a spontaneous green transition cannot be assumed. Environmental fiscal expenditure flattens rather than steepens the curve, and the inverted U remains identifiable only where expenditure is at or below its average; in provinces with large overnight-visitor volumes, expenditure renders an otherwise undetectable tourism–carbon relationship estimable, a pattern we term statistical activation. Spillover effects from neighbouring provinces are negative in sign under a contiguity matrix, but there they are imprecisely estimated and the total effect cannot be distinguished from zero, whereas under a dense economic-distance matrix both the indirect effect and the total effect become positive and statistically significant; the direction of the externality is therefore weight-matrix dependent rather than settled. Heterogeneity in the pattern of coefficients is evident across regional endowments and governance intensity. The findings reframe sustainable tourism policy: waiting for an endogenous turning point is not a viable decarbonization strategy, whereas proactive and spatially coordinated fiscal intervention is. Expressed as carbon intensity per unit of tourism output, environmental performance becomes a measurable, monitorable basis for setting destination-specific sustainability targets. Full article

Durability assessments of textile-reinforced mortar (TRM) are often fragmented across exposure conditions, test scales, and observation times. An existing durability campaign on glass- and basalt-textile TRM was reorganized into a hierarchical, condition-specific state comprising matrix, textile pull-out, and single-lap bond responses, preserving the different pull-out definitions the two textile systems require. A frozen, target-blind linear transition rule was assessed by withheld 3000 h reconstruction across 43 prediction records corresponding to 31 unique physical targets, the difference arising because matrix observations are shared by both textile branches. Single-lap retention was reconstructed with the lowest errors, at median absolute percentage errors of 3.2% for glass and 6.8% for basalt, whereas the basalt pull-out coordinates reached 84.7% and 175.9% and produced four physically inadmissible negative extrapolations, indicating structural mismatch with non-monotonic bond trajectories. Benchmarked against persistence baselines using identical information, the rule was approximately unbiased but did not achieve a lower absolute error than carrying the 5000 h state backwards. The 5000 h extrapolation was shown to be algebraically determined by the reconstruction residuals, the signed errors differing by an exact factor of minus two, so the two temporal exercises are not independent confirmation. Cross-fitted sequential updating did not improve the withheld single-lap estimate under either training design, the fitted coefficient was indistinguishable from zero in every fold, and propagated uncertainty increased throughout. One of fifteen cross-scale associations survived multiplicity correction. The contribution is therefore architectural and evidential, identifying where state-transition assumptions are informative, mixed, or structurally unsupported, rather than evidence of external predictive validity. Full article

Interfacial engineering can modify the electrical response of hafnium zirconium oxide capacitors, but separating polarization dynamics from circuit charging remains essential. We compare approximately 10 nm thick Hf_0.5Zr_0.5O₂ (HZO) capacitors with otherwise identical capacitors containing an additional approximately 1 nm thick ZrO₂ bottom interlayer (hereafter HZO–ZrO₂ capacitors). Polarization–voltage measurements, cross-sectional scanning transmission electron microscopy, capacitance–voltage characterization, and switching/non-switching pulse measurements were combined across electrode areas of 25–10,000 μm². Both stacks exhibit ferroelectric hysteresis and butterfly-shaped capacitance curves. At 2.5 V and 100 kHz, the HZO–ZrO₂ capacitors have 10–15% higher capacitance. Their measured switching-onset delays are nevertheless lower at areas ≥ 400 μm²: 2.1, 5.6, and 23.0 ns at 400, 2500, and 10,000 μm², compared with 3.2, 7.7, and 30.9 ns for HZO; the largest relative reduction is 34.4%, observed at 400 μm², where the delay decreases from 3.2 ns to 2.1 ns. At 25 and 100 μm², the measured differences (0.1 ns) are within the ≈0.1 ns timing resolution of the single-record measurement and are reported as descriptive values. Dividing the measured delay by the capacitance yields a quantity with resistance units that serves as a useful diagnostic for comparing devices, but it cannot be equated with a physical contact resistance. In particular, simply setting the delay equal to R × C with the nominal 100 Ω external resistance leads to an inconsistency for the largest HZO–ZrO₂ capacitors, where it would imply a negative additional resistance. These results associate ZrO₂ insertion with reduced operational delay while establishing the calibration requirements for interpreting its physical origin. Full article

Entity matching (EM) aims to determine whether two tuples from heterogeneous data sources refer to the same real-world entity and serves as a fundamental task in data integration, knowledge graph construction, and intelligent information retrieval. However, existing EM methods still suffer from three major limitations: insufficient semantic representation for heterogeneous tuples, difficulty in constructing reliable supervision signals in zero-shot settings, and inadequate collaboration among multiple semantic views, which together result in limited robustness and unstable performance. To address these limitations, we propose a novel self-supervised entity matching framework based on multi-semantic collaboration (MSSC-EM). MSSC-EM consists of three tightly coupled modules: RAG-based Information Augmentation (RIA), Automatic Data Augmentation (ADA), and Collaborative Learning of Multi-Semantic Features (CL-MSF). In MSSC-EM, first, RIA employs Retrieval-Augmented Generation (RAG) to mine implicit contextual semantics from an external knowledge corpus and enrich the original tuple representations. Second, ADA constructs pseudo-positive and negative labels through an Incremental Knowledge Graph Construction (IKGC) strategy and enhances their reliability via confidence-aware Monte Carlo Dropout (MC-Dropout) filtering, thereby providing high-quality supervision signals without manual annotation. Finally, CL-MSF collaboratively learns digital, structural, and relational semantic features to improve the robustness and discriminative capability of tuple representations. Extensive experimental results on eight real-world EM benchmarks demonstrate that, under the unsupervised evaluation setting, MSSC-EM achieves the best or tied-best performance among the compared unsupervised methods. Under the supervised evaluation setting, MSSC-EM further outperforms or matches the compared self-supervised methods on most datasets and achieves competitive results against the compared supervised baselines. Full article