Search Results (6,094)

Background/Objectives: Molecular docking and molecular dynamics are accurate but computationally expensive, so the compounds entering them must be chosen well. The present study proposes CADT, a confidence-gated affinity–ADME-T docking-triage cascade that decides which compounds are worth docking. Methods: The gate combines an ensemble estimate of drug–target affinity with its epistemic uncertainty and an applicability-domain check. Predicted absorption, distribution, metabolism, excretion, and toxicity (ADME-T) developability is added as a soft flag. All components were trained on openly licensed Therapeutics Data Commons data. Ranking was assessed on the DAVIS and KIBA kinase panels and on BindingDB $K_d$, under three split protocols over five seeds. The routing decision was then examined against molecular docking, in which 407 compound–target pairs were docked into six withheld kinases. Results: A Morgan-fingerprint gradient-boosting model reached a concordance index of $0.866\pm 0.006$, with $0.813$ for unseen targets and $0.720$ for unseen drugs. Across eight ADME-T endpoints, the area under the ROC curve ranged from $0.65$ to $0.91$. On the cold-target split the cascade reduced the compounds sent to docking by $86\%$ while retaining $61\%$ of the true strong binders. Docking measured that reduction at $85\%$, and at an equal budget, the gate enriched true binders more than the docking score itself. Conclusions: A transparent pre-screen can prioritise compounds ahead of structure-based calculation at a fraction of its cost. However, the uncertainty and applicability-domain terms act as an abstention mechanism rather than an accuracy gain, and that abstention is not free. Full article

Left atrial appendage closure has emerged as an established nonpharmacological strategy for stroke prevention in selected patients with atrial fibrillation, particularly those in whom long-term oral anticoagulation is problematic. Its contemporary role, however, is defined by a central tension between procedural precision and clinical uncertainty. Advances in multimodality imaging, including transesophageal echocardiography, cardiac computed tomography, three-dimensional echocardiography, fusion imaging, and emerging computational tools, now allow highly accurate assessment of left atrial appendage anatomy, device sizing, intraprocedural guidance, and postprocedural surveillance. These techniques are essential to minimize peri-device leak, device-related thrombus, and other procedure-related complications in a structure characterized by marked interindividual variability. At the same time, the randomized evidence base remains nuanced. Early warfarin-era trials established the feasibility of LAAC, whereas contemporary comparisons with direct oral anticoagulants and best available medical therapy have yielded more heterogeneous results, reflecting differences in patient selection, comparator regimens, and endpoint design. Taken together, the available data support LAAC as a reasonable option in carefully selected patients at elevated bleeding risk or with limited tolerance for chronic anticoagulation, but not as a universal substitute for oral anticoagulant therapy. Future progress will depend on more refined phenotyping, standardized imaging pathways, and longer-term comparative data to better define which patients derive the greatest net clinical benefit from LAAC. Full article

Bacterial self-healing cement-based materials (BSHCMs) couple microbial mineralization with cement-based material design to autonomously seal cracks and potentially restore durability. However, their performance depends on a complex interaction among bacterial viability and physiological state, mineralization pathway, carrier and nutrient systems, calcium availability, matrix chemistry, crack characteristics, moisture, and exposure history. This review, supported by bibliometric mapping of 805 Scopus-indexed records, examines these interdependencies through the specific lens of artificial intelligence (AI) readiness. The mapping revealed six interconnected research themes spanning bacterial mineralization, sustainable cement-based systems, matrix chemistry and transport, encapsulation, durability, and machine learning (ML). AI evidence was classified as directly demonstrated in bacterial self-healing systems, transferable from adjacent concrete and structural health monitoring applications, or prospective. Although existing ML studies report high internal predictive performance, their engineering generalizability remains limited by heterogeneous and frequently literature-derived datasets, random train–test partitioning, potential feature leakage, synthetic data dependence, insufficient uncertainty reporting, and scarce independent laboratory or field validation. The analysis further shows that sustainability and economic benefits cannot be assumed from the biological nature of the technology but must be demonstrated through service life extension relative to the additional burdens of cultivation, nutrients, carriers, and processing. Advancing BSHCMs toward trustworthy AI-supported engineering, therefore, requires harmonized and machine-readable datasets, matched attribution controls, delayed cracking and realistic exposure experiments, explicit uncertainty and negative result reporting, study-grouped and external validation, and pilot- to field-scale testing. AI should consequently be regarded not as a substitute for biological healing, but as a decision support layer whose value depends fundamentally on the quality, traceability, and transferability of the underlying experimental evidence. Full article

Under China’s dual-carbon strategic goals, large-scale public centralized heating plays a critical role in energy conservation. However, traditional manual open-loop control suffers from high response latency. Furthermore, existing unidirectional predictive methods lack dynamic feedback correction mechanisms. To address these issues, this study proposes an intelligent dual-system collaborative control architecture specifically designed for public heating systems. This architecture utilizes a cloud–edge device framework. It establishes a two-way linkage between heating equipment and the cloud decision platform. Consequently, it constructs an integrated regulation framework encompassing forward decision generation, reverse state verification, and dynamic feedback correction. Specifically, the forward module utilizes a Mamba-structured state-space model to generate data-driven boiler operation strategies. Meanwhile, the reverse module employs a Temporal Convolutional Network with Monte Carlo Dropout (TCN-MC Dropout). This probabilistic network enables state inversion evaluation with reliable uncertainty prediction intervals. These two modules are deeply coupled through an adaptive feedback correction mechanism. Together, they significantly improve system stability and operational robustness under complex thermal disturbances. Specifically, the proposed architecture achieves a room temperature compliance rate exceeding 96% and restricts temperature fluctuations to within $\pm 0.75 °\mathrm{C}$. Simultaneously, it reduces boiler energy consumption by over 16.4%. This solution has been successfully deployed in the heating network at Shaanxi Normal University as a representative real-world case study. Ultimately, it provides a practical technical reference for the intelligent upgrading and low-carbon transformation of public centralized heating systems. Full article

Background: Artificial intelligence (AI) is increasingly being integrated into cardiovascular medicine, with potential applications across image analysis, procedural planning, risk stratification, decision support, and workflow optimization. However, its adoption in interventional cardiology remains heterogeneous and may be influenced by several factors. We aimed to conduct a nationwide survey to assess attitudes towards AI among Italian interventional cardiologists. Methods: We conducted a nationwide, cross-sectional, web-based survey of Italian interventional cardiologists. A structured questionnaire collected information on professional characteristics, familiarity with and current use of AI, perceived clinical applications, expected benefits, trust, implementation barriers, and training needs. Conditional branching was used to obtain additional details from respondents who reported current use of AI-based tools, while all responses were collected voluntarily and analyzed in anonymized, aggregate form. Categorical variables and Likert-scale responses were summarized using descriptive statistics, with exploratory comparisons performed across prespecified professional and institutional subgroups. Results: Among 129 respondents, 70.5% reported at least moderate familiarity with AI and 77.5% reported some current use, although only 60.5% reported regular or occasional professional use, and applications were concentrated mainly in research, education, and information synthesis rather than direct procedural support. Nearly half (48.1%) expected AI to become standard in many procedures within 5 years, while 69.0% anticipated either routine use or particular value in complex cases. Attitudes were broadly favorable, with 85.3% agreeing that AI could improve diagnostic and procedural precision, 86.8% expressing strong interest in future use, and 76.7% stating that AI should support rather than replace physician judgment. The leading barriers were medico-legal uncertainty (45.0%), poor integration with existing clinical systems (34.9%), and cultural resistance or operator distrust (29.5%), whereas preservation of physician control was the most frequently cited requirement for adoption (58.9%). Greater AI familiarity was independently associated with current AI use (p < 0.001) and good or high trust (p < 0.001). Compared with no prior training, one and multiple AI training experiences were independently associated with good or high familiarity (both p < 0.05). Conclusions: Italian interventional cardiologists showed substantial exposure to AI, strong interest in future adoption, and generally favorable expectations regarding its contribution to diagnostic precision, workflow, and procedural support. Acceptance remained conditional on physician oversight, stronger clinical validation, reliable interoperability, and clear medico-legal governance, and previous AI-focused education appeared independently associated with greater familiarity. Full article

Engineering ethics education faces the challenge of moving students from abstract principles toward context-sensitive decision-making. This exploratory mixed-methods study examined associations between participation in a board-game-based strategy and patterns consistent with situated ethical awareness in first-semester engineering courses at Pontificia Universidad Javeriana (Bogotá/Cali) and Universidad de los Andes (Bogotá). The 90-min intervention adapted Sheriff of Nottingham to an engineering entrepreneurship context and combined asymmetric roles, incentives, negotiation, uncertainty, consequences, and structured debriefing. Across 17 course offerings, 616 pre-intervention and 413 post-intervention questionnaires were analyzed as unmatched samples. In 2025-1, aggregate post-intervention scores were higher across all six measures (global score: 7.40 to 8.79), with the largest difference in academic integrity. In 2025-2, Rights (+6.9 percentage points) and Common Good (+6.1 pp) gained first-priority prominence, while the perceived role of ethics increased modestly. Qualitative responses after the intervention more often incorporated consequences, effects on others, professional responsibility, and individual–collective tensions. In 2025-2, 90.4% agreed or strongly agreed that the game helped them reflect differently on ethical dilemmas. These convergent patterns show early ethical sensitization and contextualization rather than stable ethical competence; the unmatched design, absence of a comparison group, and exploratory instruments preclude causal or individual-level claims. Full article

The preparation of reliable input data for groundwater flow modeling remains a persistent bottleneck in data-scarce regions, where parameter estimation relies heavily on subjective expert judgment. This study introduces ALAES, a novel object-oriented expert system that codifies formal and heuristic knowledge to guide hydrogeologists through the entire pre-modeling workflow. The knowledge base was developed through structured interviews with 20 international experts and formalized using the KOD methodology within the Kappa-PC shell. The system comprises 258 production rules, 114 classes, and 1136 instances. Validation on the data-scarce Rhis-Nekor aquifer in Morocco showed that ALAES recommended MODFLOW and diagnosed modeling feasibility as challenging. A comparative assessment revealed substantial improvements over a baseline model developed without ALAES guidance: spatial resolution increased by a factor of four in critical zones, steady-state water balance consistency improved from 87% to 94%, and mean absolute errors were reduced by over 50% under ±20% perturbations. The system guided parameter estimation, reducing porosity uncertainty by over 40%, and achieved strong transient calibration (R² = 0.99). Three future management scenarios were evaluated, enabling formulation of a recommended exploitation strategy. By bridging the gap between data availability and modeling requirements, ALAES provides an explicit, reproducible decision-support tool for sustainable groundwater management in data-limited environments worldwide. Full article

Cape Verdean immigrants are a growing yet understudied population in the United States (U.S.), with limited data on HPV and cervical cancer prevention knowledge. This descriptive qualitative study explored awareness, knowledge, beliefs, and perceptions related to HPV and cervical cancer prevention among Cape Verdean immigrant mothers of children aged 11–17 residing in Massachusetts, a state in the northeastern U.S. A qualitative descriptive design was used. Semi-structured interviews were conducted with 27 Cape Verdean immigrant mothers of adolescents aged 11–17 residing in the study area. Interviews were transcribed, translated, and analyzed thematically using inductive and deductive approaches guided by the literature and emergent narratives. MAXQDA was used for data management, and coding reliability was established through independent review by bilingual researchers. Participants demonstrated varying levels of awareness, ranging from comprehensive understanding to limited knowledge. Two major themes emerged: (1) knowledge and beliefs about cervical cancer and (2) knowledge and beliefs about HPV. Findings revealed fragmented knowledge, misconceptions about screening and transmission, and gendered perceptions of HPV as a “women’s issue.” While many mothers valued prevention and recognized HPV’s link to cervical cancer, uncertainties remained regarding vaccination, screening recommendations, and vaccine effectiveness. Findings highlight the need for culturally and linguistically tailored interventions addressing HPV misinformation, strengthening parental confidence, and supporting Cape Verdean mothers as informed health decision-makers to improve HPV vaccination and cervical cancer prevention. Full article

Solar thermal calorimeters serve as foundational reference instruments for quantifying absorbed thermal power, evaluating optical-to-thermal conversion efficiency, and validating computational models in concentrated solar thermal (CST) research. This review synthesizes recent numerical, thermo-hydraulic, and experimental advancements, evaluating the evolution from basic single-tube configurations to high-confinement cavity calorimeters (achieving apparent absorptances $>0.99$) and flat-plate architecture enhanced with impinging jets and internal fin arrays. The selection of working fluids is examined; water accounts for approximately $87\%$ of reported implementations due to low property uncertainty, whereas synthetic oils and gaseous coolants expand operating temperature ranges at the expense of thermochemical degradation and parasitic pumping penalties. Furthermore, critical thermo-hydraulic challenges induced by extreme non-uniform heat fluxes are identified, including structural thermal bowing, parallel-channel flow maldistribution, recirculation traps, and buoyancy-driven instabilities occurring at Richardson numbers $\mathrm{Ri}\gg 10$. Metrological constraints related to solar reflection interference in non-contact thermometry and calibration drift in photometric target arrays are also critically addressed. Finally, key strategic research directions are outlined, emphasizing high-temperature advanced materials, active flow equalization, real-time multi-physics digital twins, and standardized dynamic testing metrology. Full article

To improve prompt utilization and boundary perception in SAM-based interactive medical image segmentation, this paper proposes ASAM2-UNet, a U-shaped visual foundation model built upon SAM2-UNet. Unlike methods that treat prompts as auxiliary spatial inputs, ASAM2-UNet explicitly incorporates prompt-derived priors into feature reasoning through an Interactive Prompt-Guided Focal Attention module. Specifically, a user-provided spatial prompt is converted into an explicit attention prior that modulates focal self-attention, allowing the network to emphasize target-relevant regions while retaining efficient local–global contextual modeling. In addition, a Contextual Semantic Information Complement module integrates multi-scale decoder features with uncertainty-aware and structure-aware cues derived from the initial prediction to refine ambiguous lesion boundaries. By integrating cross-layer spatial attention, local–global spatial fusion, and semantic-aware feature aggregation, the proposed module enhances the discrimination of ambiguous edges and complex foreground–background regions. Extensive experiments on five public polyp segmentation datasets demonstrate the effectiveness of the proposed method. Full article

Background: Thesis writing is a complex academic and emotional process that requires learners to develop research, writing, and communication skills while navigating through uncertainty, evolving research challenges, and gaining greater levels of scholar independence. Although scaffolding approaches are widely used in higher education, a need for conceptually grounded supervisory frameworks that integrate research skill development with learner wellbeing is becoming necessary. Purpose: This paper presents a reflective, conceptually grounded account of a logically delineated supervisory practice, described as a Staircase Framework. Developed over a decade, this practice supports thesis writing and defense in an academic forum through structured, experiential learning. Approach: Drawing primarily on experiential learning and transformative learning, with complementary perspectives from constructivism and systems thinking, the Staircase Framework organizes thesis development into progressive and interconnected scholarly activities, beginning with weekly reports and extending through posters, presentations, journal manuscripts, and thesis completion. The paper provides illustrative applications and reflective insights by integrating supervisory and learner reflections and illustrative scholarly outputs to describe the framework’s development and application, without claiming empirical effectiveness. Results: Implemented between 2015 and 2025 in a Minority-Serving Institution in the southeastern United States, the proposed framework provided a structured approach to graduate supervision that emphasized incremental learning, continuous feedback, peer mentoring, and scholarly dissemination. Supervisory observations and learner reflections suggest that these features were perceived as supporting research capability development, organization, confidence, scholarly engagement, and management of uncertainty through the thesis process. These observations are presented as reflective evidence that informs hypotheses for future empirical investigation rather than as evidence of causal effectiveness. Conclusions: The Staircase Framework offers a conceptually grounded model for organizing graduate thesis supervision by integrating educational theory with structured research practice, while concurrently emphasizing learner independence. This framework provides a practical approach that may support learner development and wellbeing, while generating a foundation for future prospective studies to evaluate its educational and psychological outcomes rigorously. Full article

Hybrid renewable energy integration (HREI) has recently emerged as a promising solution to promote energy sustainability and achieve carbon emission reduction in rural regions. However, how to evaluate HREI projects in an effective and comprehensive manner remains an urgent challenge, due to substantial uncertainties and the involvement of multiple stakeholders with heterogeneous interests. To address this issue, this paper applies a structured BOCR-based fuzzy MCDM framework to assess the multidimensional characteristics of HREI projects from the perspectives of benefits, opportunities, costs, and risks (BOCR). First, an indicator system for HREI project evaluation is established considering BOCR factors, comprising four dimensions and sixteen factors that reflect key concerns of relevant stakeholders. Second, triangular fuzzy number is employed to capture the inherent uncertainty and vagueness in expert evaluation information, while the Best–Worst Method is used to determine the indicator weights and represent their relative importance. Subsequently, the Preference Ranking Organization Method for Enrichment Evaluations II is introduced to evaluate the comprehensive performance of HREI projects and provide a complete ranking of alternative HREI projects. The results indicate that alternative A2 in the Henan province, which integrates wind turbines, rooftop photovoltaic system, battery energy storage, and vehicle-to-grid interaction with electric vehicles, exhibits the highest evaluated performance among the selected alternatives under the BOCR-based evaluation framework. Furthermore, sensitivity analysis and comparative consistency analysis show that the ranking remains stable under the tested weight-variation scenarios and alternative MCDM procedures. Full article

Floating plastic debris in urban canals is a management-relevant precursor to downstream microplastic pollution. This structured narrative review synthesizes conventional field surveys, camera-, UAV-, and satellite-based image analysis and AI-assisted image-based monitoring. Emphasis is placed on Southeast Asian engineered canals, where monsoon-driven flow, tides, gates, turbidity, glare, occlusion, and organic debris challenge continuous observation. Conventional surveys provide verifiable composition data but limited temporal coverage. Camera systems increase observation frequency, while deep learning can automate detection and tracking; however, reported performance depends strongly on the dataset, site, target size, and validation design. Published studies show substantial losses under cross-site transfer and condition-specific gains from preprocessing rather than a universal accuracy threshold. The synthesis therefore develops a decision-oriented framework linking camera calibration, conditional preprocessing, site-separated validation, uncertainty reporting, and hydrological data to operational triggers for cleanup or interception. Current evidence supports monitoring and pilot decision support, while broader autonomous operation requires further field validation. Priorities include transparent evidence reporting, shared Southeast Asian datasets, standardized metrics and environmental descriptors, cross-site testing, and life-cycle evaluation of deployment cost and maintenance. Full article

Seismic pier-top pounding in high-speed railway bridges transfers short-duration girder-restraint contact forces into bridge piers, coupling local contact damage with global vibration and possible base yielding. This critical review evaluates how evidence and modelling strategies can be transferred from local contact mechanics to pier response and, ultimately, to whole-bridge seismic demand. The literature is synthesized across experimental and refined numerical characterization, reduced-order contact–structure modelling, response-equivalent-pulse construction, and nonlinear whole-bridge analysis. A qualitative evidence-confidence grading is introduced to distinguish the strength and transferability of the available evidence based on study independence, evidence type, and configuration similarity. The primary scope is high-speed railway bridges, while the underlying contact–structure modelling principles are transferable to conventional railway and highway bridges with comparable pier-top restraints, subject to bridge-specific calibration. Conventional spring-dashpot models are computationally efficient but sensitive to contact stiffness, damping, restitution, and damage assumptions, whereas refined finite-element models resolve local response at substantially greater computational cost. Static, impulse-equivalent, and prescribed pulse representations can reduce analysis effort, but agreement in force or impulse alone does not ensure equivalence in pier displacement, base moment, plastic rotation, or residual demand. Demand-oriented pulses can reproduce selected component-level responses within a calibrated applicability domain, while response-triggered loading remains a conditional system-level reduction requiring reliable event logic, state updating, and independent benchmark validation. Future research should prioritize realistic restraint tests, identifiable parameter ranges, multi-demand validation, uncertainty quantification and damage-updatable repeated-impact models. These advances can provide a mechanics-based basis for performance-oriented restraint assessment, while practical design application requires consistency with code-based seismic restraint provisions and post-earthquake track-system serviceability criteria. Full article

Financial econometrics has expanded rapidly in recent decades, giving researchers many new tools for empirical analysis. Yet empirical findings often remain sensitive to modeling choices, sample construction, and specification decisions, suggesting that many disagreements arise not from estimation methods alone but from deeper limitations in what data can reveal. This survey organizes financial econometrics around three issues that determine the credibility of empirical inference: identification, dependence, and uncertainty. Identification concerns whether economic quantities such as causal effects, risk premia, and structural parameters can be credibly recovered from observable data. Dependence recognizes that assets, firms, and markets are interconnected, reducing the amount of independent information contained in financial data. Uncertainty captures not only sampling variation but also model misspecification, measurement error, and competing explanations. We review how these problems arise in asset pricing, corporate finance, ESG, and risk management and discuss implications for empirical design and inference. Full article