Search Results (47,281)

Deep learning has been applied in various fields and has been proven to provide good results for classification tasks. However, there is limited understanding of a deep learning model’s decisions, so deep learning is commonly described as a black box. Applying deep learning for critical applications such as medical diagnostic process introduces trust issues. For the deep learning model to be trusted by the medical practitioners, the methods employed by the deep learning model must be seen to be aligned with the diagnostic process employed by the medical practitioners. Explainable methods such as Grad-CAM can be applied to improve the explainability of the deep learning models by providing an visual interpretation of the deep learning classification result decision process. In this study, two deep learning models, VGG16 and ResNet50 are trained using three training methods, one with randomly initialized weights, and two transfer learning methods, which are feature extraction and fine-tuning, to classify the spinal abnormalities based on X-ray images. The classification metrics results are compared and further analyses using Grad-CAM heatmaps are included. The models also evaluated using a stratified five-fold cross-validation, results revealed some disparity between the model’s accuracy and clinical relevance. The randomly initialized VGG16 obtained a classification accuracy of 93.79% but does not focus on clinically relevant regions. On the other hand, not only do the fine-tuned ResNet50 and VGG16 obtain high accuracies of 98.22% and 99.12%, but the heatmaps show that the models focus on more relevant regions. A comparison of the two models shows that the heatmaps produced by the fine-tuned ResNet50 are in more agreement with the clinical view than the fine-tuned VGG16. This study provides a useful reference for interpreting a deep learning-based classification result using explainable method particularly in spine abnormalities analysis with Grad-CAM. Full article

Background/Objectives: HBeAg-positive chronic hepatitis B (CHB) patients with very high viral replication are often clinically considered a homogeneous, low-risk population. However, substantial biochemical, virological, and fibrosis-related heterogeneity may exist. This study aimed to characterize this heterogeneity in treatment-naive, HBeAg-positive CHB patients with ultra-high viral loads (HBV DNA ≥ 7.0 log₁₀ IU/mL). Furthermore, we sought to identify predictors of significant fibrosis and detect clinically relevant discordant phenotypes, such as silent disease progression despite normal alanine aminotransferase (ALT) levels. Methods: This single-center, retrospective, cross-sectional study analyzed consecutively screened eligible patients. A liver stiffness measurement (LSM, kPa) and controlled attenuation parameter (CAP, dB/m) were obtained via transient elastography. Significant fibrosis was defined as an LSM ≥ 7.0 kPa. Statistical evaluations included Spearman’s correlation, multivariable regression, ALT-LSM stratification, and K-means clustering. Results: Among 413 included patients, age and aspartate aminotransferase (AST) emerged as independent risk factors for significant fibrosis, whereas log10 HBV DNA and log10 HBsAg were independent negative predictors. Patients with HBsAg ≥ 25,000 IU/mL exhibited significantly lower LSM values than those with lower HBsAg levels. Notably, 18.4% of patients with strictly normal ALT (≤40 U/L) presented with an LSM ≥ 7.0 kPa, indicating silent progression. Cluster analysis further identified two distinct patient phenotypes characterized by differing age, ALT, viral load, and fibrosis profiles. Conclusions: An ultra-high viral load in HBeAg-positive CHB does not guarantee a uniformly benign clinical state. By quantifying biochemical, virological, and fibrotic heterogeneity, this study highlights a critical subgroup with silent fibrosis progression that risks being overlooked by ALT-based assessments alone. Full article

Urban air quality assessment increasingly relies on spatial interpolation to complement fixed monitoring networks; however, the reliability of geostatistical methods depends strongly on temporal conditions and pollutant characteristics. Despite extensive application, limited attention has been paid to how kriging performance varies across hours of the day and months of the year, particularly when contrasting primary pollutants driven by local emissions with secondary pollutants formed through atmospheric chemistry. This study evaluates the temporal sensitivity of Ordinary Kriging (OK) for mapping urban air pollutants in the Mexico City Metropolitan Area. Using hourly observations from the official air quality monitoring network (2021), we analyze ozone (O₃), a secondary pollutant, and sulfur dioxide (SO₂), a primary pollutant, under representative diurnal and monthly scenarios. Variogram model selection and predictive performance are assessed through leave-one-out cross-validation and external hold-out validation across multiple temporal blocks and months. Results indicate that kriging performance is highly sensitive to both hour of day and month. For O₃, smoother Gaussian variogram structures perform best during peak photochemical conditions, producing coherent regional concentration fields with gradual spatial gradients. In contrast, SO₂ exhibits stronger local variability and sharper spatial gradients, favoring exponential variogram models, particularly under stable morning atmospheric conditions associated with primary emission accumulation. Sensitivity analyses further reveal that no single variogram model is universally optimal and that interpolation accuracy depends more on temporal stratification and pollutant behavior than on variogram form alone. These findings demonstrate that geostatistical interpolation is a valuable tool for urban air quality assessment only when temporal sensitivity and pollutant-specific dynamics are explicitly incorporated. The proposed framework provides practical guidance for the responsible use of interpolated air quality maps, supports sustainable urban monitoring strategies, and contributes to more reliable exposure assessment in megacities with limited sensor coverage. Full article

Impurity profiling is of significant analytical and regulatory importance, particularly in the context of lifecycle quality management. A robust chaotropic chromatography method was developed for the determination of olanzapine and its two oxidative degradation products in tablets, in accordance with the ICH Q14 guideline and the principles of Analytical Quality by Design (AQbD). Risk assessment was performed using a combination of the Ishikawa diagram, CNX (Control, Noise and eXperimental) classification, and Failure Mode and Effect Analysis (FMEA). This multistep evaluation identified the critical analytical procedure parameters (APPs) as the acetonitrile content in the mobile phase, the concentration of perchloric acid in the aqueous phase, and the pH of the aqueous phase. These APPs were studied using an experimental design approach to model their effects on key analytical procedure attributes and to compute a multidimensional design space. Robust optimization supported by Monte Carlo simulations ensured compliance with predefined acceptance criteria with a probability of at least 95%. Method validation demonstrated adequate selectivity, limits of quantification of 0.75 µg/mL and 0.5 µg/mL for impurities B and D, linearity with correlation coefficients ≥0.990, accuracy of 98–102% for olanzapine and 70–130% for impurities, and repeatability with RSD ≤2% for the assay and ≤10% for impurities. The method was successfully applied to commercial tablet analysis. Full article

The constant growth of IP data traffic, driven by sustained annual increases surpassing 26%, is pushing current optical transport infrastructures towards their capacity limits. Since the deployment of new fiber cables is economically demanding, ultra-wideband transmission is emerging as a promising cost-effective solution, enabled by multi-band amplifiers and transceivers spanning the entire low-loss window of standard single-mode fibers. In this scenario, an accurate modeling of the frequency-dependent fiber parameters is essential to reliably model optical signal propagation. In particular, the combined impact of attenuation variations with frequency and inter-channel stimulated Raman scattering (SRS) fundamentally shapes the power evolution of wide wavelength division multiplexing (WDM) combs and directly affects nonlinear interference (NLI) generation, as well as the amount of ASE noise. In this work, we review a set of analytical approximations, based on phenomenological approaches, for frequency-dependent attenuation and Raman scattering gain, and analyze their impact on achieving an effective balance between computational efficiency and physical fidelity. Through extensive analyses performed with the open-source software GNPy (version 2.12, Telecom Infra Project) on an optical line system exploring multi-band scenarios spanning C+L+S, C+L+E, and U-to-E transmission, we demonstrate that the proposed approximations reproduce the reference SRS power evolution and NLI profiles with root mean square errors (RMSEs) consistently below 0.03 dB, and down to the 10⁻³–10⁻² dB range for the most accurate configurations. Although the current implementation does not yet provide a direct reduction in computational time, the proposed framework lays the groundwork for future developments toward closed-form or semi-analytical solutions, enabling more efficient modeling and optimization of ultra-wideband optical transmission. Full article

Background: Gastric cancer (GC) remains a global health challenge, with high mortality rates often linked to late-stage diagnosis. Novel, non-invasive biomarkers are urgently needed to improve the detection and prognosis of this malignant pathology. This study aimed to evaluate the diagnostic and prognostic utility of serum Cluster of Differentiation 276 (CD276) and Dickkopf Related Protein 3 (DKK3) in patients with GC. Methods: In this case–control study, serum levels of CD276 and DKK3 were quantified in 40 GC patients and 40 age-matched healthy controls. The diagnostic performance of each marker and their combination was assessed using Receiver Operating Characteristic (ROC) curve analysis. Correlations between biomarker levels and clinicopathological features were evaluated using Spearman’s correlation. The Kaplan–Meier method and the Cox Proportional Hazards Regression Model were used to assess survival. Results: Serum CD276 levels were found to be significantly elevated in GC patients compared to healthy controls (median 60.06 vs. 18.71 units, p < 0.001). Conversely, serum DKK3 levels were significantly suppressed in the GC group (median 92.47 vs. 121.02 units, p < 0.001). In ROC analysis, CD276 demonstrated excellent diagnostic accuracy as a standalone biomarker (AUC: 0.836). DKK3 showed independent diagnostic value (AUC: 0.792), but adding DKK3 to CD276 did not provide statistically significant incremental benefit (DeLong’s p = 0.443). Survival analysis was underpowered due to limited events and short follow-up duration. Conclusions: In patients with predominantly locally advanced gastric cancer, CD276 can be a primary diagnostic marker, and the addition of DKK3 does not demonstrate a statistically significant improvement but may provide complementary information. Performance in early-stage disease requires validation in future studies. The opposing dysregulation of these markers, reflecting immune checkpoint activation (CD276) and tumor suppressor loss (DKK3), provides a robust and synergistic noninvasive signature. To assess the prognostic value of these two markers, studies involving a larger number of patients and a longer follow-up period are needed. Full article

This study investigates directional causality between Bitcoin and gold across different market conditions. Rather than relying on mean-based dependence, we examine how causal effects vary across return quantiles, investment horizons, and market regimes. To address this question, we apply a Causal–Frequency–Quantile–Regime (CFQR) framework. The approach combines frequency-domain Granger causality, quantile-based non-causality tests, and endogenous regime classification within a unified setting. Macroeconomic controls are included to reduce omitted variable bias. Statistical inference relies on bootstrap procedures with false discovery rate correction to account for multiple testing. Using daily data from 2013 to 2025, we find that the full-sample directional dominance between Bitcoin and gold is generally weak after multiple testing adjustments. However, under stress regimes, the causal relationship of gold to Bitcoin becomes more pronounced at longer investment horizons. Under normal conditions, causal effects remain unstable and fragmented. Economic effects are modest. Variance-based hedging gains are limited, while downside risk measures show moderate improvement during stress periods. Overall, the evidence suggests that gold does not serve as a universal hedge for Bitcoin, but may exert conditional informational influence during high-uncertainty states. The CFQR framework provides a structured way to identify such state-dependent causal patterns. Full article

Horizontal well drilling is the mainstream technology for developing deep oil and gas resources. Engineering practice has demonstrated that hydraulic oscillators can solve the problem of the backing pressure of pipe strings and improve drilling efficiency. However, the design of excitation parameters for hydraulic oscillators is currently largely based on idealized friction models and does not fully consider the nonlinear characteristics of friction between the drill string and the formation, resulting in a lack of quantitative basis for parameter selection under different operating conditions. A series of laboratory friction tests was conducted to systematically characterize the dependence of interfacial friction behavior on sliding velocity across different combinations of drill string materials, drilling fluid systems, and rock lithologies. Based on the experimentally determined velocity–friction relationships, a drill string dynamic model incorporating a hydraulic oscillator was developed in which nonlinear frictional effects at the interface were explicitly represented. Using this modeling framework, parametric simulations were carried out to examine how variations in excitation amplitude and excitation frequency influence drag reduction performance under diverse operating conditions. The simulation results indicate that the contribution of drill string material to overall drag reduction effectiveness is comparatively limited, whereas drilling fluid type plays a dominant regulatory role. Oil-based drilling fluids significantly enhance drag reduction performance relative to water-based systems and exhibit greater responsiveness to adjustments in excitation parameters. Rock lithology exerts a pronounced influence on the effectiveness of drag reduction. When water-based drilling fluids are used, the overall performance ranks from highest to lowest as limestone, shale, and sandstone. In contrast, under oil-based drilling fluid conditions, the relative ordering shifts to shale, followed by sandstone, and then limestone. Excitation amplitude is the dominant parameter in enhancing drag reduction capability, and in most cases, its incremental effect exceeds that of excitation frequency; however, under certain specific operating conditions, increasing the excitation frequency can provide additional drag reduction benefits. Based on the above findings, a hydraulic oscillator excitation parameter design method was proposed that matches drilling conditions and formation characteristics by distinguishing between different drilling fluid environments and lithologies, with amplitude as the primary control parameter and frequency as a supplementary parameter. This method provides a theoretical foundation for the design of output parameters of hydraulic oscillators operating under diverse working conditions. Full article

Oxygen therapy is one of the most widely used interventions in critical care, yet it remains poorly individualized. Recent trials and meta-analysis suggest no mortality difference between conservative and liberal oxygen strategies, reinforcing the perception that dose does not matter within usual ranges. From this perspective, we argue that this apparent neutrality may largely reflect methodological and conceptual limitations, although true clinical equivalence in some patient populations remains plausible and cannot be excluded based on current evidence. Heterogeneous populations, overlapping oxygenation targets, and the absence of exposure metrics (time in hyperoxia, time in hypoxemia, and cumulative partial pressure of arterial oxygen/peripheral oxygen saturation curves) dilute phenotype-specific signals and force distinct physiological responses into a single pooled estimate. We propose a conceptual model in which oxygen behaves as a dose-dependent, time-dependent drug with phenotype-specific therapeutic windows, particularly in chronic hypercapnia, traumatic brain injury, sepsis, and early versus late acute respiratory distress syndrome. Building on this model, we outline a methodological agenda for precision oxygen trials: defining interventions by actual exposure, pre-specifying pathophysiological subgroups, adopting patient-centered core outcome sets, and using adaptive, target-range designs and individual patient data meta-analyses. For contemporary guidelines and research, the key question is no longer whether conservative or liberal oxygen therapy is superior on average, but how to match the right oxygenation range to the right intensive care unit phenotype at the right time. Moving from population-averaged comparisons to exposure-aware, phenotype-oriented strategies is essential if oxygen therapy is to become a truly precision intervention in critical care. Full article

This study investigates the impact of board attributes (board size, board independence, gender diversity, and nationality diversity) on corporate board diligence through employing panel data of listed firms in Muscat Securities Market from 2014 to 2024. Through the application of multiple regression analysis, the paper determines predictors for board diligence and offers an agency theory-based and resource dependence theory-based perspective on this construct. The findings reveal positive relations between board independence and board diligence, which suggests that the independent director has monitoring function. On the other hand, board size and nationality diversity are negatively related to diligence levels indicating a lack of coordination and communication. However, board gender diversity does not seem statistically related to board diligence. Several robustness tests, such as lagged independent variables, fixed industry effects, alternative estimation techniques, and instrumental variable approach, support the validity of our findings. This research helps investors and policymakers to better understand the extent to which board structure is related to meeting activity and director engagement in emerging markets. The study contributes to the literature on board diligence in emerging markets and evidence the impact of gender and nationality diversity on corporate board performance in Oman. Full article

Population aging is widely regarded as a major fiscal risk for European welfare states and a central challenge to long-term fiscal sustainability. The article critically reexamines the deterministic assumption by assessing whether the fiscal implications of demographic aging in the European Union (EU) are mechanically driven or conditioned by policy context and institutional capacity. Using panel data for the EU-27 over the period 2014–2024, the study employs a two-way fixed-effects framework and interaction models to examine the relationship between demographic aging and key fiscal outcomes, including public pension expenditures, total social protection spending, and the general government balance. Furthermore, the analysis examines whether indicators associated with the silver economy, such as employment at older ages and digital inclusion, condition the fiscal effects of aging within countries over time. The results suggest that demographic aging does not exhibit a statistically significant association with pension or social protection expenditures once institutional heterogeneity and common shocks are controlled. In contrast to deterministic expectations, aging is positively associated with general government balance, suggesting the presence of policy-mediated fiscal adjustment dynamics rather than automatic fiscal deterioration. Interaction estimates further indicate that digital inclusion among older cohorts conditions the relationship between demographic aging and fiscal balance, while silver economy indicators do not display robust standalone fiscal effects. These findings should be interpreted as evidence of policy-mediated adjustment dynamics rather than as causal estimates of demographic effects. Building on these findings, the article advances a conceptual interpretation of the aging–fiscal nexus in which demographic pressures interact with institutional adaptation and policy capacity. Fiscal sustainability under demographic aging emerges as a policy-mediated outcome that may reflect broader institutional and governance contexts, rather than demographic structure alone. While governance quality is not directly estimated as an observable variable, the analysis interprets fiscal outcomes within a governance-conditioned institutional framework that emphasizes policy mediation rather than deterministic demographic effects. The findings contribute to ongoing debates on fiscal sustainability in aging societies by demonstrating that fiscal outcomes in the European Union are best understood as institutionally conditioned and policy-mediated rather than mechanically driven by demographic structure alone. Full article

Investment in risky financial assets plays a crucial role in individual wealth accumulation and broader financial market development. However, existing research has primarily emphasized financial literacy while giving limited attention to behavioral mechanisms that may weaken its influence on investment behavior. In particular, hyperbolic discounting, reflecting time-inconsistent preferences that favor immediate rewards over long-term gains, may constrain the effective translation of financial knowledge into forward-looking financial decisions. Against this background, this study examines whether hyperbolic discounting mediates the association between financial literacy and investment in risky assets using large-scale survey data from Japan’s Money and Life survey. Employing regression-based mediation analysis within a cross-sectional framework, the results indicate that financial literacy is strongly and positively associated with risky asset investment, while hyperbolic discounting exerts a statistically significant but economically small mediating effect that slightly attenuates this relationship. The findings suggest that cognitive financial capability remains the dominant driver of participation in risky financial markets, whereas present-biased preferences play a secondary behavioral role. These results provide important implications for investors, educators, and policymakers by highlighting that policies aimed at improving financial literacy are likely to yield substantial investment benefits, while complementary interventions addressing behavioral biases may offer additional, though more modest, gains in promoting long-term, forward-looking financial decision-making. Full article

The continuous increase in the penetration rate of renewable energy has posed severe challenges to the flexibility of power systems, especially in coastal and island areas where local power supply is insufficient while electricity demand keeps growing. Focusing on the regional water–energy nexus system (WENS), this paper fully taps into the flexibility potential of seawater desalination plants (SWDPs) as adjustable loads, and proposes a bi-level optimal dispatch model. First, the operational characteristics of reverse osmosis (RO) seawater desalination loads are analyzed, and an operational model encompassing water intake equipment, high-pressure pumps, clear water tanks and product water tanks is established. Second, a dispatch framework for the regional WENS incorporating SWDP is designed, on the basis of which a bi-level optimal dispatch model is constructed: the upper-level model takes maximizing wind power accommodation and minimizing wind power output fluctuation as the objectives, so as to determine the wind power output and the charging/discharging strategy of supercapacitors; constrained by the decisions made by the upper-level model, the lower-level model comprehensively takes into account the operation cost of thermal power units (TPUs), the wind curtailment penalty cost of the system, the operation cost of energy storage systems and the operation cost of SWDP, and thus establishes an optimization model with the goal of minimizing the comprehensive operation cost of the system. Finally, a comparative analysis is carried out under different scenarios. The results show that compared with the optimal scheduling scheme in which the seawater desalination load does not participate in regulation, the proposed method can reduce the wind curtailment rate by 43.71%, the energy consumption cost of the seawater desalination load by 50.98%, and the total system operation cost by 22.51%, thus providing a feasible approach for the collaborative optimization of water–energy systems in coastal areas. Full article

The fast growth of edge–cloud computing infrastructures has increased the cybersecurity burden even as it has substantially amplified the energy use and carbon footprint of intrusion detection systems (IDSs). In order to overcome this challenge, this paper suggests GreenShield, which is a framework of low-carbon cybersecurity involving lightweight cryptography, deep learning that is energy efficient, and carbon conscious system optimization across distributed edges and in cloud setup. GreenShield employs a hierarchical federated learning architecture with integrated knowledge distillation and a carbon-aware scheduling controller that dynamically adjusts security response execution based on threat intensity and renewable energy availability. As extensive experiments on the UNSW-NB15 and CIC-IDS2017 datasets show, GreenShield attains 98.73% detection accuracy and is 67.4% more energy efficient than traditional deeplearning-based IDSs. Further, the suggested system reduces the operational carbon emissions up to 97.6%, which is equivalent to a reduction of around 2.8 kg CO2-equivalent/per hour in a typical edge-deployment situation, yet it does not undermine the performance of the detection. These findings suggest that GreenShield can be one of the meaningful alternatives in providing viable and scalable sustainable cybersecurity that supports carbon-conscious security workflows in the future edge–cloud computing architecture. Full article

As remarked by Boltzmann, the Second Law of Thermodynamics is notable for the fact that it is readily proved using elementary statistical arguments, but becomes harder and harder to verify the more precise the microscopic description of a system. In this article, we investigate one particular realization of the 2nd Law, namely Joule heating in a wire under electrical bias. We analyze the production of entropy in an exactly solvable model of a quantum wire wherein the conserved flow of entropy under unitary quantum evolution is taken into account using an exact formula for the entropy current of a system of independent quantum particles. In this exact microscopic description of the quantum dynamics, the entropy production due to Joule heating does not arise automatically. Instead, we show that the expected entropy production is realized in the limit of a large number of local measurements by a series of floating thermoelectric probes along the length of the wire, which inject entropy into the system as a result of the information obtained via their continuous measurements of the system. The decoherence resulting from inelastic processes introduced by the local measurements is essential to the phenomenon of entropy production due to Joule heating, and would be expected to arise due to inelastic scattering in real systems of interacting particles. Full article