Search Results (3,537)

Accurate uncertainty estimation in unlabeled data represents a fundamental challenge in active learning. Traditional deep active learning approaches suffer from a critical limitation: uncertainty-based selection strategies tend to concentrate excessively around noisy decision boundaries, while diversity-based methods may miss samples that are crucial for decision-making. This over-reliance on confidence metrics when employing deep neural networks as backbone architectures often results in suboptimal data selection. We introduce Distance-Measured Data Mixing (DM2), a novel framework that estimates sample uncertainty through distance-weighted data mixing to capture inter-sample relationships and the underlying data manifold structure. This approach enables informative sample selection across the entire data distribution while maintaining focus on near-boundary regions without overfitting to the most ambiguous instances. To address noise and instability issues inherent in boundary regions, we propose a boundary-aware feature fusion mechanism integrated with fast gradient adversarial training. This technique generates adversarial counterparts of selected near-boundary samples and trains them jointly with the original instances, thereby enhancing model robustness and generalization capabilities under complex or imbalanced data conditions. Comprehensive experiments across diverse tasks, model architectures, and data modalities demonstrate that our approach consistently surpasses strong uncertainty-based and diversity-based baselines while significantly reducing the number of labeled samples required for effective learning. Full article

As a key indicator for measuring urban green visibility, the Green View Index (GVI) reflects actual visible greenery from a human perspective, playing a vital role in assessing urban greening levels and optimizing green space layouts. Existing studies predominantly rely on single-source remote sensing image analysis or traditional statistical regression methods such as Ordinary Least Squares and Geographically Weighted Regression. These approaches struggle to capture spatial variations in human-perceived greenery at the street level and fail to identify the non-stationary effects of different drivers within localized areas. This study focuses on the Luolong District in the central urban area of Luoyang City, China. Utilizing Baidu Street View imagery and semantic segmentation technology, an automated GVI extraction model was developed to reveal its spatial differentiation characteristics. Spearman correlation analysis and Multiscale Geographically Weighted Regression were employed to identify the dominant drivers of GVI across four dimensions: landscape pattern, vegetation cover, built environment, and accessibility. Field surveys were conducted to validate the findings. The Multiscale Geographically Weighted Regression method allows different variables to have distinct spatial scales of influence in parameter estimation. This approach overcomes the limitations of traditional models in revealing spatial non-stationarity, thereby more accurately characterizing the spatial response mechanism of the Global Vulnerability Index (GVI). Results indicate the following: (1) The study area’s average GVI is 15.24%, reflecting a low overall level with significant spatial variation, exhibiting a “polar core” distribution pattern. (2) Fractal dimension, normalized vegetation index (NDVI), enclosure index, road density, population density, and green space accessibility positively influence GVI, while connectivity index, Euclidean nearest neighbor distance, building density, residential density, and water body accessibility negatively affect it. Among these, NDVI and enclosure index are the most critical factors. (3) Spatial influence scales vary significantly across factors. Euclidean nearest neighbor distance, building density, population density, green space accessibility, and water body accessibility exert global effects on GVI, while fractal dimension, connectivity index, normalized vegetation index, enclosure index, road density, and residential density demonstrate regional dependence. Field survey results confirm that the analytical conclusions align closely with actual greening conditions and socioeconomic characteristics. This study provides data support and decision-making references for green space planning and human habitat optimization in Luoyang City while also offering methodological insights for evaluating urban street green view index and researching ecological spatial equity. Full article

Shanxi Province hosts a rich diversity of traditional villages. From a river basin perspective, adherence to natural laws and the removal of administrative barriers are essential for reshaping the conservation paradigm. Using spatial analysis and multiscale geographically weighted regression, this study revealed the spatial patterns of 619 traditional villages and how environmental, socioeconomic, and historical–cultural factors shape the spatial heterogeneity. Villages clustered within the Yellow River Basin and the Haihe River Basin, forming an agglomeration belt and three high-density cores. Distance to rivers was a key factor in village siting, with 70.8% located within 3 km of the nearest river. Village density exhibited a U-shaped relationship with distance to roads, and an inverted U-shaped relationship with distance to county-level administrative centers. The interaction between intangible cultural heritage density and average annual precipitation showed the strongest explanatory power, with positive local regression coefficients exceeding 95% and 72%, respectively. Traditional villages constitute an evolving human–environment system in which water resources underpin spatial patterns and intangible cultural heritage sustains endogenous cultural vitality. These findings provide a theoretical framework for graded conservation and resource coordination at the river basin scale. Full article

Safety remains a central challenge in autonomous driving: overly rigid safeguards can cause unnecessary stops and erode efficiency. Addressing this safety–efficiency trade-off requires specifying what behaviors to incentivize. In reinforcement learning, the reward provides that specification. Conventional reward surrogates—such as distance gaps and time-to-collision (TTC)—depend on instantaneous geometry and often miss unfolding multi-vehicle interactions, whereas sparse terminal rewards provide no intermediate guidance. Accordingly, we adapt Safety Potential (SP)—a short-horizon, time-weighted path-overlap forecast—into a dense reward-shaping term that provides a predictive risk-aware signal for anticipatory throttle/brake control. In the CARLA v0.9.14 roundabout environment, SP attains 94% success with 3% collisions; in percentage points, this is 16.00, 13.00, and 5.75 higher success and 18.75, 9.50, and 7.25 lower collisions than No-Safe, Distance, and TTC, respectively. Adding a lightweight reactive guard at inference further reduces collisions to below 1% without sacrificing success. These results indicate that injecting a predictive, overlap-based risk measure directly into the reward supplies temporally consistent safety cues and clarifies the trade-off between progress and risk in reinforcement-learning-based vehicle control. Full article

This study examines the role of private investment in promoting multidimensional poverty reduction in a sustainable manner in Vietnam by analyzing both spatial and temporal spillover effects. Provincial panel data for 2010–2024 are employed. To assess the spatial spillover effects, three econometric models are applied: SAR, SEM, and SDM. Diagnostic tests suggest that the SDM model is the most appropriate for the research data. Results based on the contiguity and inverse distance weight matrices show that private investment not only reduces poverty in recipient provinces but also generates benefits for neighboring areas, highlighting the need for coordinated planning of industrial zones and regional economic hubs. To analyze this relationship over both the short-term and long-term horizons, the study employs PMG and CCEP estimators, while the DCCEP model verifies robustness in a dynamic framework. The findings consistently confirm that private investment contributes to multidimensional poverty reduction. An additional result from the DCCEP model indicates that literacy and urbanization rate have significant positive effects on poverty reduction, while these relationships are not detected in other models. This finding carries important implications for building an enabling investment environment to attract and effectively utilize private capital to implement multidimensional poverty reduction strategies towards sustainability and aligned with sustainable development objectives. Full article

Rainfall stations in small and medium-sized river basins in China are sparsely distributed and unevenly spaced, resulting in insufficient spatial representativeness of precipitation data and posing challenges to the accuracy of flood forecasting. Spatial interpolation methods for rainfall data are a key tool for bridging the gap between discrete rainfall station data and continuous surface rainfall data; however, their applicability in flood forecasting for small and medium-sized river basins with sparse rainfall stations requires further investigation. Taking the Hezikou basin as the study area and focusing on the Liuxihe model, this study analyzes the distribution characteristics of the seven rainfall stations in the basin and the interpolation effectiveness of the original Thiessen Polygon Interpolation (THI) method in the model. It compares and discusses the applicability of the THI, the Inverse Distance Weighting (IDW) method, and the Trend Surface Interpolation (TSI) method in flood forecasting for this basin. Different rainfall station distribution scenarios (full coverage, upstream only, downstream only, single rainfall station) were set up to study the performance differences in each method under extremely sparse conditions. The results indicate that, under the sparse condition of only 0.0068 rainfall stations per square kilometer in the Hezikou basin, IDW interpolation yields the best flood forecasting results, with model Nash–Sutcliffe Efficiency (NSE) values all above 0.85, Kling–Gupta Efficiency (KGE) values exceeded 0.78, and the Peak Relative Error (PRE) was controlled within 0.09, significantly outperforming THI and TSI. Additionally, as rainfall station sparsity increased, IDW exhibited the smallest decline in performance, showing a weak negative correlation (p ≤ 0.05) between prediction performance and rainfall station sparsity, demonstrating stronger adaptability to sparse scenarios. When station information is extremely limited, IDW performs more stably than THI and TSI in terms of certainty coefficients (NSE, KGE) and flood peak error control. The Inverse Distance Weighting method (IDW) can provide reliable rainfall spatial interpolation results for flood forecasting in small and medium-sized basins with sparse rainfall stations. Full article

Multi-exposure image fusion (MEF) aims to generate a well-exposed image by combining multiple photographs captured at different exposure levels. However, deep learning-based approaches are often highly dependent on the quality of the training data, which can lead to inconsistent color reproduction and loss of fine details. To address this issue, this study proposes a variance-driven hybrid MEF framework based on a U-Net architecture, which adaptively balances structural and chromatic information. In the proposed method, the variance of randomly cropped patches is used as a training weight, allowing the model to emphasize structurally informative regions and thereby preserve local details during the fusion process. Furthermore, a fusion strategy based on the geometric color distance, referred to as the Chroma scale, in the LAB color space is applied to preserve the original chroma characteristics of the input images and improve color fidelity. Visual gamma compensation is also employed to maintain perceptual luminance consistency and synthesize a natural fine image with balanced tone and smooth contrast transitions. Experiments conducted on 86 exposure pairs demonstrate that the proposed model achieves superior fusion quality compared with conventional and deep-learning-based methods, obtaining high JNBM (17.91) and HyperIQA (70.37) scores. Overall, the proposed variance-driven U-Net effectively mitigates dataset dependency and color distortion, providing a reliable and computationally efficient solution for robust MEF applications. Full article

Quantifying the regional source of long-lived ozone precursors (especially GHGs) transported to Hong Kong is hampered by sparse observational data and computational limitations. This study introduces an observation-driven analytical framework that integrates a tracer ratio (ethylbenzene/m,p-xylene), wind–source–distance correlations to constrain transport corridors, and inventory mapping to determine the province- and sector-specific contributions, operationalized by identifying transport periods from observations, classifying sources with diagnostic ratios into five emission categories, deriving seasonal weighting factors via frequency normalization, mapping high-resolution inventory classes to these categories to restructure sectoral inventories, and combining normalized provincial spatial weights with the restructured inventories to quantify cross-boundary CO₂ and CH₄ emissions by sector and region. High-resolution measurements were conducted at the Cape D’Aguilar Supersite (CDSS), which showed dominant wintertime regional transport with mean concentrations of 435.29 ± 7.64 ppm (CO₂) and 2083.45 ± 56.50 ppb (CH₄). Thirteen transport periods were quantitatively analyzed, and province–sector contributions were estimated. The dominant provincial contributors were Guangdong (20.66%), followed by Jiangxi (18.36%) and Zhejiang (11.15%). Motor vehicles (70%), fuel combustion (15%), and solvent use (10%) were the primary contributing sectors. The framework enables province- and sector-specific attribution under stated assumptions and provides a tool for measuring cross-boundary mitigation and developing air quality and climate strategies in monsoon-affected coastal cities. Full article

Complete sub-hourly rainfall datasets are critical for accurate flood modeling, real-time forecasting, and understanding of short-duration rainfall extremes. However, these datasets often contain missing values due to sensor or transmission failures. Recovering missing values (or filling these data gaps) at high temporal resolution is challenging due to the imbalance between rain and no-rain periods. In this study, we developed and tested two approaches for the imputation of missing 10-min rainfall data by means of machine learning (Multilayer Perceptron and Random Forest) and interpolation methods (Inverse Distance Weighting and Ordinary Kriging). The (a) direct approach operates on raw data to directly feed the imputation models, while the (b) two-step approach first classifies time steps as rain or no-rain with a Random Forest classifier and subsequently applies an imputation model to predicted rainfall depth instances classified as rain. Each approach was tested under three spatial scenarios: using all nearby stations, using stations within the same cluster, and using the three most highly correlated stations. An additional test involved the comparison of the results obtained using data from the imputed time interval only and data from a time window containing several time intervals before and after the imputed time interval. The methods were evaluated with reference to two different environments, mountainous and coastal, in Campania region (Southern Italy), under data-scarce conditions where rainfall depth is the only available variable. With reference to the application of the two-step approach, the Random Forest classifier shows a good performance both in the mountainous and in the coastal area, with an average weighted F1 score of 0.961 and 0.957, and an average Accuracy of 0.928 and 0.946, respectively. The highest performance in the regression step is obtained by the Random Forest in the mountainous area with an R² of 0.541 and an RMSE of 0.109 mm, considering a spatial configuration including all stations. The comparison with the direct approach results shows that the two-step approach consistently improves accuracy across all scenarios, highlighting the benefits gained from breaking the data imputation process in stages where different physical conditions (in this case, rain and no-rain) are separately managed. Another important finding is that the use of time windows containing data lagged with respect to the imputed time interval allows capturing the atmospheric dynamics by connecting rainfall instances at different time levels and distant stations. Finally, the study confirms that machine learning models outperform spatial interpolation methods, thanks to their ability to manage data with complicated internal structure. Full article

Adverse effects of transportation arise from the buildup of various stressors, which collectively compromise animal welfare. This study aimed to assess short-term behavioral responses, physiological stress, and meat quality as indicators of welfare in Japanese Black cattle on arrival at the slaughter facility. A total of 154 animals from different production farms were observed. Generalized linear mixed models were used, with fixed effects including animal type, weight, season, source, loading size, distance, transport experience, and their interaction with time periods. Significant post-transport behaviors and elevated cortisol concentration were observed, particularly in heifers, lighter animals, those transported in summer, from multiple farms, at high loading sizes, or without prior transport experience. Steers, heavier animals, and the same farm groups yielded higher carcass weights, while cattle transported under low loading size had improved marbling scores and a higher probability of achieving A5-grade carcasses. These findings suggest that management practices should focus on animals most susceptible to transport stress and strategies such as mitigating heat stress, transporting animals from the same production farm, and reducing loading sizes should be implemented to improve welfare and meat quality upon arrival. Full article

In forest resource plot surveys, tree relative positioning is a crucial task with profound silvicultural and ecological significance. However, traditional methods such as compasses and total stations suffer from low efficiency, high costs, or poor environmental adaptability, while single-sensor technologies (e.g., UWB or IMU) struggle to balance accuracy and stability in complex forest environments. To address these challenges, this study designed a multi-sensor fusion-based tree positioning device. By integrating the high-precision ranging capability of Ultra-Wideband (UWB) with the dynamic motion perception advantages of an Inertial Measurement Unit (IMU), a dynamic weight fusion algorithm was proposed, effectively mitigating UWB static errors and IMU cumulative errors. Experimental results demonstrate that the device achieves system biases of −1.54 cm (X-axis) and 1.27 cm (Y-axis), with root mean square errors (RMSE) of 21.34 cm and 23.93 cm, respectively, across eight test plots. The average linear distance error was 26.23 cm. Furthermore, in single-operator mode, the average measurement time per tree was only 20.89 s, approximately three times faster than traditional tape measurements. This study confirms that the proposed device offers high positioning accuracy and practical utility in complex forest environments, providing efficient and reliable technical support for forest resource surveys. Full article

Reducing noise in electric multiple-unit (EMU) gearboxes demands prediction tools that are both rapid and reliable. Gear sound pressure levels vary sharply with micrometre-scale changes such as tooth repair, inclination, or profile relief, yet traditional estimates depend on hours-long CAE simulations. We present a data-driven hybrid surrogate that combines k-means clustering and inverse distance weighting (CLS-IDW) within the ODYSSEE A-Eye platform to map geometry modifications directly to broadband noise. Trained on the open 200-case Romax dataset, the model returns predictions within milliseconds and reproduces unseen operating points, with R² = 0.75 and a mean absolute error of 2.33 dB, matching solver repeatability. Sensitivity analysis identifies a −7° tooth inclination coupled with a 10 µm repair depth as the most effective combination, lowering noise by 3–5 dB. Eliminating costly CAE loops, the surrogate supports acoustics-aware optimisation at the concept stage, compressing development cycles and enhancing passenger comfort while maintaining transparency for regulatory review. Full article

To address the challenges of high-dimensional complexity and increasing heterogeneity in expert opinions, this study proposes a novel numerical aggregation model for multi-expert decision making based on triangular intuitionistic fuzzy numbers (TIFNs) and the Plant Growth Simulation Algorithm (PGSA). The proposed framework transforms experts’ fuzzy preference information into five-dimensional geometric vectors and employs the PGSA to perform global optimization, thereby yielding an optimized collective decision matrix. To comprehensively evaluate the aggregation performance, several quantitative indicators—such as weighted Hamming distance, correlation sum, information intuition energy, and weighted correlation coefficient—are introduced to assess the results from the perspectives of consensus, stability, and informational strength. Extensive numerical experiments and comparative analyses demonstrate that the proposed method significantly improves expert consensus reliability and aggregation robustness, achieving higher decision accuracy than conventional approaches. This framework provides a scalable and generalizable tool for high-dimensional fuzzy group decision making, offering promising potential for complex real-world applications. Full article

This study evaluated the decadal trends and spatial distribution of four irrigation suitability indices—Electrical Conductivity (EC), Sodium Adsorption Ratio (SAR), Magnesium Hazard (MH), and Kelley’s Ratio (KR)—using agricultural groundwater data collected from 157 monitoring sites across Korea between 2015 and 2024. Internationally recognized classification criteria were applied, long-term trends were analyzed using the Mann–Kendall test and Sen’s slope estimator, and spatial distributions for 2015, 2020, and 2024 were visualized using Inverse Distance Weighting (IDW). The results showed that EC and SAR remained at generally low absolute levels but exhibited statistically significant increasing trends with Sen’s slopes of +0.0038 and +0.0053/year, respectively, indicating the necessity of long-term salinization management. KR remained largely stable throughout the study period. In contrast, MH displayed a distinct pattern, with unsuitable levels concentrated in Jeju Island—approximately 15% of monitoring sites were classified as unsuitable for irrigation. This was interpreted as the combined effect of the basaltic aquifer’s geological and hydrological characteristics, seawater intrusion, and the relatively high mobility of Mg compared with Ca. This study uniquely integrates temporal trend tests with spatial mapping at a national scale and offers a mechanistic interpretation of MH vulnerability in Jeju’s volcanic aquifers. These findings emphasize the need for tailored regional management centered on groundwater abstraction control and continuous monitoring to ensure the sustainable use of agricultural groundwater. Full article

With the rapid expansion of data scale, compute-intensive tasks will become a core application of 6G networks. As Unmanned Aerial Vehicle (UAV) technology advances, UAVs can assist in task offloading for mobile edge computing by collaborating to overcome individual UAV limitations in battery life and computational capacity. Hence, in this paper, we propose a task offloading algorithm for multiple UAVs based on a temporal graph. We first formulate an optimization problem to minimize the total completion time of UAV swarm task offloading by classifying tasks and determining task priorities and subtask dependencies. To solve this problem, we introduce a temporal graph to simulate service nodes and task sequences in computing networks. It can reveal task execution priorities by calculating proximity indices, which indicate the ratio of physical distance to the sum of task weights, and determining timestamp offsets. In the following, to reduce unnecessary waiting and computation resource allocation risks, we transform the optimization problem into a directed acyclic graph connectivity problem, which identifies the fastest temporal paths for each UAV, forming a dedicated service network. Finally, we propose a two-stage matching algorithm that achieves optimal matching based on service node locations, statuses, task types, and offloading demands. Simulation results demonstrate that the algorithm performs exceptionally well, reducing task completion times and significantly outperforming other algorithms in terms of task utility. Full article