Search Results (2,708)

Financial time series are heterogeneous, nonstationary, and dispersed across institutions that cannot share raw data. While federated learning enables collaborative modeling under privacy constraints, fixed architectures struggle to accommodate cross-market drift and device-resource diversity; conversely, existing neural architecture search techniques presume centralized data and typically ignore communication, latency, and privacy budgets. This paper introduces FedRegNAS, a regime-aware federated NAS framework that jointly optimizes forecasting accuracy, communication cost, and on-device latency under user-level $(\epsilon ,\delta )$-differential privacy. FedRegNAS trains a shared temporal supernet composed of candidate operators (dilated temporal convolutions, gated recurrent units, and attention blocks) with regime-conditioned gating and lightweight market-aware personalization. Clients perform differentiable architecture updates locally via Gumbel-Softmax and mirror descent; the server aggregates architecture distributions through Dirichlet barycenters with participation-weighted trust, while model weights are combined by adaptive, staleness-robust federated averaging. A risk-sensitive objective emphasizes downside errors and integrates transaction-cost-aware profit terms. We further inject calibrated noise into architecture gradients to decouple privacy leakage from weight updates and schedule search-to-train phases to reduce communication. Across three real-world equity datasets, FedRegNAS improves directional accuracy by 3–7 percentage points and Sharpe ratio by 18–32%. Ablations highlight the importance of regime gating and barycentric aggregation, and analyses outline convergence of the architecture mirror-descent under standard smoothness assumptions. FedRegNAS yields adaptive, privacy-aware architectures that translate into materially better trading-relevant forecasts without centralizing data. Full article

This study introduces a novel AI-powered Business Intelligence Dashboard System (AIBIDS) designed to detect and visualize calendar-based anomalies in cryptocurrency returns. Focusing on Bitcoin as a case study, the system integrates unsupervised machine learning algorithms to identify periods of abnormal market behavior across multiple temporal resolutions. The proposed system leverages a star-schema OLAP data warehouse, enabling real-time anomaly detection, dynamic visualization, and drill-down exploration of market irregularities. Empirical results confirm the presence of pronounced calendar effects in Bitcoin returns, such as heightened anomalies during Q1 and Q4, and reveal model-specific sensitivities to local versus global volatility. Our novel platform offers a practical, scalable innovation for investors, analysts, and regulators seeking to monitor cryptocurrency markets more effectively, and contributes to the emerging FinTech literature on AI-driven anomaly detection and behavioral market dynamics. Full article

Background: Dysphagia from aging, neurodegeneration, structural anomalies, or cognitive decline harms quality of life. The videofluoroscopic swallowing study (VFSS) is the diagnostic gold standard but manual interpretation is labor-intensive and costly, motivating automation. Methods: We introduce a Vision Transformer (ViT) using a temporal sliding window and 3D patch tokenization to capture spatio-temporal dependencies in variable-length VFSS via attention. Training/evaluation used 1154 VFSS sequences from 107 individuals (548 abnormal, 606 normal) with 5-fold cross-validation and comparisons to VGG-16, ResNet-50, EfficientNet-V1/V2, and MobileNet. Results: The ViT achieved 84.37 ± 1.15% accuracy, 90.81 ± 2.11% sensitivity, 79.49 ± 1.66% specificity, 82.94 ± 2.76% precision, 85.68 ± 1.54% F1-score, and AUC 0.878 (5-fold). It outperformed all CNN baselines across metrics; paired t-tests confirmed significant gains (p < 0.05). Conclusions: The pure ViT’s attention-based spatio-temporal modeling yields robust VFSS classification and is well-suited for screening workflows requiring timely abnormality detection, providing a foundation for clinically deployable VFSS analysis. Full article

Recommender systems are widely deployed across digital platforms, yet their opacity raises concerns about auditability, fairness, and user trust. To address the gap between predictive accuracy and model interpretability, this study proposes a glass-box architecture for trustworthy recommendation, designed to reconcile predictive performance with interpretability. The framework integrates interpretable tree ensemble model (Random Forest, XGBoost), an NLP sub-model for tag sentiment, prioritising transparency from feature engineering through to explanation. Additionally, a Reality Check mechanism enforces strict temporal separation and removes already-popular items, compelling the model to forecast latent growth signals rather than mimic popularity thresholds. Evaluated on the MovieLens dataset, the glass-box architectures demonstrated superior discrimination capabilities, with the Random Forest and XGBoost models achieving ROC-AUC scores of 0.92 and 0.91, respectively. These tree ensembles notably outperformed the standard Logistic Regression (0.89) and the neural baseline (MLP model with 0.86). Beyond accuracy, the design implements governance through a multi-layered Governance Stack: (i) attribution and traceability via exact TreeSHAP values, (ii) stability verification using ICE plots and sensitivity analysis across policy configurations, and (iii) fairness audits detecting genre and temporal bias. Dynamic threshold optimisation further improves recall for emerging items under severe class imbalance. Cross-domain validation on Amazon Electronics test dataset confirmed architectural generalisability (AUC = 0.89), demonstrating robustness in sparse, high-friction environments. These findings challenge the perceived trade-off between accuracy and interpretability, offering a practical blueprint for Safe-by-Design recommender systems that embed fairness, accountability, and auditability as intrinsic properties rather than post hoc add-ons. Full article

Smart cities increasingly rely on intelligent technologies to improve urban infrastructure, sustainability, and quality of life. Traffic flow prediction is essential for the optimization of the transportation system, reducing congestion and improving mobility. However, real-world traffic data are often noisy, limited in size, and lack sufficient features to capture the flow dynamics and temporal dependencies, making accurate prediction a significant challenge. Previous studies have shown that recurrent neural network (RNN) variants, such as LSTM and GRU, are well-suited for time series forecasting tasks, but their performance is highly sensitive to hyperparameter settings. This study proposes a hybrid approach that integrates GRU with a metaheuristic optimization algorithm to address this challenge. After effective preprocessing steps and a sliding time window are applied to structure the data, particle swarm optimization (PSO) is utilized to optimize the hyperparameters of the GRU. The model’s performance is evaluated using RMSE, MAE, and $R^2$, and compared against several baseline approaches, including LSTM, CNN-LSTM, and a manually configured GRU. According to the experimental findings, the GRU model that was manually adjusted performed the best overall. However, the PSO-GRU model demonstrated competitive results, confirming that metaheuristics offer a promising alternative when manual tuning is not feasible despite the higher computational costs. Full article

Synthetic Aperture Radar (SAR) has been widely applied for rice monitoring, especially in cloud-prone areas, due to its ability to penetrate clouds. However, only limited methods were developed to monitor separately irrigated rice and rainfed rice ecosystems. This study demonstrated the use of multi-temporal polarimetric dual-polarization (dual-pol) SAR (Sentinel-1B and ALOS PALSAR-2) data to monitor and discriminate the irrigated and favorable rainfed rice ecosystems in the province of Iloilo, Philippines. Key polarimetric parameters derived from H–A–α and model-based dual-pol decomposition were analyzed to characterize the rice phenology of both ecosystems. Segmented regression was performed to detect breakpoints corresponding to changes in rice phenology within each ecosystem and used to identify the parameters to use for classification. Based on the results, Sentinel-1B polarimetric parameters (entropy, anisotropy, and alpha) can capture the phenological dynamics, whereas ALOS2 polarimetric parameters were more sensitive to water conditions, as reflected in span and volume scattering. Furthermore, irrigated rice exhibited more stable and predictable scattering patterns than favorable rainfed rice. Using the Random Forest classifier, various combinations of backscatter and polarimetric parameters from Sentinel-1B and ALOS2 were tested to discriminate between the two ecosystems. The highest classification accuracy (81.81% overall accuracy; Kappa = 0.6345) was achieved using the combined backscatter (S1B VH, ALOS2 HH, and HV) and polarimetric parameters from both sensors. The results demonstrated that polarimetric parameters effectively capture phenological stages and associated scattering mechanisms, with the integration of Sentinel-1B and ALOS2 data improving the discrimination of irrigated and favorable rainfed rice systems. Full article

Alpine treeline ecotones (ATEs) are critical ecological boundaries that are highly sensitive to climate change, yet their long-term spatial dynamics remain understudied in mountainous regions. This study investigates four decades (1984–2023) of ATE elevational shift along the Eastern Slopes of the Canadian Rocky Mountains (ESCR) using the Alpine Treeline Ecotone Index (ATEI), developed by integrating NDVI gradients, elevation data, and logistic regression. Multi-temporal Landsat composites and Shuttle Radar Topography Mission (SRTM) data were processed in Google Earth Engine (GEE) to map ATE boundaries over nine composite intervals. Results show a 13.32% increase in ATE area (from 1494.17 km² to 1693.19 km²), indicating a general upslope expansion consistent with a warming climate and extended growing seasons. Although the Mann–Kendall test did not reveal a significant monotonic trend in area change (neither upward nor downward) (p-value > 0.05), notable spatial variability was observed (approximately 8 km²/year). North-facing aspects exhibited the greatest mean elevation gain (+40.21 m), and significant ecotonal changes occurred within the Bow and Athabasca watersheds (p < 0.05), which are equal to around 416 and 452 km², respectively. These findings highlight the complex, aspect- and watershed-dependent nature of alpine vegetation responses to climate forcing and demonstrate the utility of ATEI for monitoring vegetation biodiversity shifts in high-elevation ecosystems. Full article

Gully is the most visible sign of land degradation, but its effects on runoff and sediment dynamics during snowmelt conditions remain poorly understood. This study monitored a typical gully in the Mollisols region of Northeast China to investigate runoff and sediment transport at the Gully Head (GH) and Gully Tail (GT) during spring snowmelt. Results showed that gully significantly influenced snow distribution, with deeper snow accumulation than on slopes. Runoff at the GH lasted 9 days, while gully connectivity extended catchment runoff by 10 additional days. Runoff temporal variation at GH and GT was broadly consistent, with GH contributing 7.4% of the total runoff at GT. Peak runoff discharge and sediment concentration occurred on the sixth day after snowmelt onset, driven by snow cover and air temperature. Gully significantly increased the sediment concentration from the upslope runoff. Runoff responses to temperature varied by melt stage, with GT showing higher sensitivity, especially under high-runoff conditions. High sediment yield was linked not to snow depth, but to late-stage snowmelt and soil thawing, when erosion sensitivity peaked. Hysteresis analysis revealed dominant clockwise loops during this phase, contrasting with figure-eight and counterclockwise patterns in other stages. These findings highlight the importance of targeting erosion control during late snowmelt when runoff intensifies and soils thaw. Full article

Nanomedicine has now become a transformative platform that enhances the precision and efficacy of immunotherapy approaches and allows customizations like never before when it comes to cancer, as well as autoimmune conditions. Using platforms based on nanoscale, researchers have been able to manipulate immune responses operating across spatial and temporal scales to address key limitations of conventional immunotherapy associated with working with immune response such as immune evasion, systemic toxicity, and poor pharmacokinetics. Sophisticated nanoparticles (such as stimuli-sensitive ones, exosome-mimetic vesicle nanoparticles, and nanoparticles with CRISPR) allow directed immunomodulators, antigens, and gene-editing systems to reach one or more particular immune compartments. The innovations allow reprogramming of immune cells, immune tolerance rejuvenation, and expansion of antitumor immunity without significant off-target effects. Finding applications in integrating the artificial intelligence as well as multi-omics techniques, the process leads to personalization of the nano-immunotherapies based on patient-specific immuno-signatures. The chapter discusses the mechanistic rationale, therapeutic advancement, and the translational opportunities of nanotechnology-based immunotherapies that define them as part of a foundation of future generations of clinical approaches to precision immune modulation in oncology and autoimmune diseases. Full article

Accurate real-time prediction of labor productivity is crucial for the successful management of construction projects. However, it remains a significant challenge due to the dynamic and uncertain nature of construction environments. Existing models, while valuable for planning and post-analysis, often rely on historical data and static assumptions, rendering them inadequate for providing actionable, real-time insights during construction. This study addresses this gap by suggesting a novel hybrid AI-stochastic framework that integrates a Long Short-Term Memory (LSTM) network with Markov Chain modeling for dynamic productivity forecasting in repetitive construction activities. The LSTM component captures complex, long-term temporal dependencies in productivity data, while the Markov Chain models probabilistic state transitions (Low, Medium, High productivity) to account for inherent volatility and uncertainty. A key innovation is the use of a Bayesian-adjusted Transition Probability Matrix (TPM) to mitigate the “cold start” problem in new projects with limited initial data. The framework was rigorously validated across four distinct case studies, demonstrating robust performance with Mean Absolute Percentage Error (MAPE) values predominantly in the “Good” range (10–20%) for both the training and test datasets. A comprehensive sensitivity analysis further revealed the model’s stability under data perturbations, though performance varied with project characteristics. By enabling more efficient resource utilization and reducing project delays, the proposed framework contributes directly to sustainable construction practices. The model’s ability to provide accurate real-time predictions helps minimize material waste, reduce unnecessary labor costs, optimize equipment usage, and decrease the overall environmental impact of construction projects. Full article

In this study, we investigate a nonlinear Schrödinger equation relevant to the evolution of optical beams in weakly nonlocal media. Utilizing the modified F-expansion method, we construct a variety of novel soliton solutions, including dark, bright, and wave solitons. These solutions are illustrated through comprehensive graphical simulations, including 2D contour plots and 3D surface profiles, to highlight their structural dynamics and propagation behavior. The effects of the temporal parameter on soliton formation and evolution are thoroughly analyzed, demonstrating its role in modulating soliton shape and stability. To further explore the system’s dynamics, chaos and sensitivity theories are employed, revealing the presence of complex chaotic behavior under perturbations. The outcomes underscore the versatility and richness of the present model in describing nonlinear wave phenomena. This work contributes to the theoretical understanding of soliton dynamics in weakly nonlocal nonlinear optical systems and supports advancements in photonic technologies. This study reports a novel soliton structure for the weak nonlocal cubic–quantic NLSE and also details the comprehensive chaotic and sensitivity analysis that represents the unexplored dynamical behavior of the model. This study further demonstrates how the underlying nonlinear structures, along with the novel solitons and chaotic dynamics, reflect key symmetry properties of the weakly nonlocal cubic–quintic Schrödinger model. These results enhanced the theoretical framework of the nonlocal nonlinear optics and offer potential implications in photonic waveguides, pulse shape, and optical communication systems. Full article

The safe and efficient operation of nuclear power systems largely relies on sensor networks that continuously collect and transmit monitoring data. However, due to the high sensitivity of the nuclear power field and strict privacy restrictions, data among different nuclear entities are typically not directly shareable, which poses challenges to constructing a global digital twin with strong generalization capability. Moreover, most existing digital twin approaches tend to treat sensor data as static, overlooking critical temporal patterns that could enhance fault prediction performance. To address these issues, this paper proposes an Edge Temporal Digital Twin Network (ETDTN) for cloud–edge collaborative, sensor-driven fault detection in nuclear power systems. ETDTN introduces a continuous variable temporal representation to fully exploit temporal information from sensors, incorporates a global representation module to alleviate the non-IID characteristics among different subsystems, and integrates a temporal attention mechanism based on graph neural networks in the latent space to strengthen temporal feature learning. Extensive experiments on real nuclear power datasets from 17 independent units demonstrate that ETDTN achieves significantly better fault detection performance than existing methods under non-sharing data scenarios, obtaining the best results in both accuracy and F1 score. The findings indicate that ETDTN not only effectively preserves data privacy through federated parameter aggregation but also captures latent temporal patterns, providing a powerful tool for sensor-driven fault detection and predictive maintenance in nuclear power systems. Full article

The present study aimed to examine the temporal dynamics of tuberculosis incidence, mortality, and TB–HIV coinfection in the First Health Region of Alagoas from 2001 to 2022, with particular attention to sex-specific differences. The analysis revealed pronounced divergences between men and women. The male series exhibited significant positive autocorrelation and high interannual variability, indicating strong temporal dependence and heightened sensitivity to external disruptions such as the COVID-19 pandemic. The female series displayed a more regular pattern without autocorrelation. Although both sexes showed declining incidence, only the reduction among women reached statistical significance; the male trend remained unstable and inconclusive. Disease burden was consistently higher among men, who accounted for most cases and maintained incidence levels above elimination targets. TB–HIV coinfection increased in both sexes, with a sharper rise among men and a statistically significant upward trend among women, accompanied by a structural shift in 2010. Additional change points in 2014 and 2018 are likely to reflect alterations in surveillance or broader public health events. The weak performance of linear models underscores the role of persistent social determinants and inequities in healthcare access. Overall, the findings demonstrate that tuberculosis remains a major public health concern and that differentiated strategies by sex are essential for effective prevention and care. Full article

Glycation-induced modifications of extracellular matrix (ECM) proteins, including collagen, are increasingly recognized as critical modulators of cellular behavior, particularly in pathophysiological contexts such as aging and diabetes. While their impact on general cell adhesion has been explored, the specific consequences for mesenchymal stem cell (MSC) mechanotransduction remain poorly defined. In this study, we investigated the temporal and mechanistic aspects of adhesion and mechanosensitive signaling in adipose-derived MSCs (ADMSCs) cultured on native versus glycated collagen substrates. Our findings identify two temporally distinct adhesion mechanisms: an initial pathway mediated by the receptor for advanced glycation end-products (RAGE), which is activated within the first 30 min following substrate engagement, and a later-stage adhesion process predominantly governed by integrins. Immunofluorescence analysis demonstrated maximal nuclear localization of YAP/TAZ transcriptional regulators during the initial adhesion phase, coinciding with RAGE engagement. This nuclear enrichment was progressively attenuated as integrin-mediated focal adhesions matured, suggesting a dynamic shift in receptor usage and mechanotransductive signaling. Interestingly, glycated collagen substrates accelerated early cell attachment but impaired focal adhesion maturation, suggesting a disruption in integrin engagement. Endogenous collagen synthesis was consistently detected at all examined time points (30 min, 2 h, and 5 h), suggesting a constitutive biosynthetic activity that remains sensitive to the glycation state of the substrate. Atomic force microscopy (AFM) demonstrated that glycation disrupts collagen fibrillogenesis: while native collagen forms a well-organized network of long, interconnected fibrils, GL-1 substrates (glycated for 1 day) displayed sparse and disordered fibrillary structures, whereas GL-5 substrates (5-day glycation) exhibited partial restoration of fibrillar organization. These matrix alterations were closely associated with changes in adhesion kinetics and mechanotransduction profiles. Taken together, our findings demonstrate that collagen glycation modulates both adhesion dynamics and mechanosensitive signaling of MSCs through a dual-receptor mechanism. These insights have significant implications for the design of regenerative therapies targeting aged or metabolically compromised tissues, where ECM glycation is prevalent. Full article

In the context of global warming, hydroclimatic conditions in the monsoon marginal zone are governed by two primary drivers: the East Asian monsoon and the westerly winds. As a sensitive indicator of climatic change, this region experiences disproportionately amplified adverse effects of climate change are markedly amplified, positioning it as a focal area for climatological research. However, the limited temporal coverage of instrumental records poses significant challenges for understanding historical hydroclimatic variability and its underlying mechanisms. To address this limitation, tree-ring width indices derived from 73 cores of Styphnolobium japonicum ((L.) Schott (1830)) are hereby employed to reconstruct summer maximum temperatures over a 433-year period in the central monsoon fringe zone—specifically, the northwestern Yan Mountains. Results confirm a strong correlation between the tree-ring width index of Styphnolobium japonicum and local summer maximum temperatures (r = 0.770, p < 0.01). Compared to the 19th century, the frequency of temperature fluctuations has increased substantially, with four abrupt regime shifts identified in the reconstructed series (1707, 1817, 1878, and 1994). Spectral analysis reveals cyclical patterns at interannual (2–7 years), decadal (10–30 years), and multidecadal (50 years) timescales. These oscillations align closely with known climate modes, including the EI Niño–Southern Oscillation (ENSO), the Pacific Decadal Oscillation (PDO), and the Atlantic Multidecadal Oscillation (AMO). Among them, the AMO presents particularly strong coherence with the reconstructed temperature variability. These outcomes improve insights into long-term temperature dynamics in the region and highlight the value of dendroclimatic proxies in reconstructing past climate conditions. Full article