Search Results (16,054)

Action chunking strategies in robot imitation learning struggle to dynamically balance between long-range motion efficiency and short-range operational precision due to their fixed planning horizon. This paper presents an Adaptive Action Chunking framework that enables robots to dynamically predict the optimal action chunk length based on real-time visual context. We design an end-to-end dual-branch network comprising a shared visual encoder, a parallel action prediction head, and a chunk-size prediction head. Experiments on two real-world bimanual robot manipulation tasks (transport-and-place and flip-and-handover) demonstrate that the method autonomously derives two distinct intelligent strategy patterns—phase-aware switching and sustained high-frequency adjustment—in response to task uncertainty. It significantly outperforms fixed-chunk baselines in both success rate and efficiency. Ablation studies confirm that the performance gain stems from the adaptive decision-making mechanism itself. Full article

Quantifying anthropogenic CO₂ increments is vital for assessing emission reductions. Using a seamless XCO₂ dataset over China reconstructed from OCO-2/3 satellite retrievals and machine learning, combined with EOF decomposition and LISA analysis, this study investigates XCO₂ anomalies and local anthropogenic increments (dXCO₂) at national and urban agglomeration scales. Nationally, XCO₂ anomalies exhibit a “southeast positive, northwest negative” spatial pattern aligning with human activities and a “winter high, summer low” seasonal cycle. EOF analysis reveals four dominant modes: anthropogenic–natural trade-offs, East Asian summer monsoon modulation, local emissions, and baseline context. At the regional scale, multi-year mean dXCO₂ (2015–2019) in Beijing–Tianjin–Hebei (BTH), Yangtze River Delta (YRD), and Pearl River Delta (PRD) are 3.46 ± 0.45 ppm, 1.30 ± 0.36 ppm, and 0.08 ± 0.14 ppm, respectively, showing higher values in northern heavy industrial zones. During the 2020–2022 pandemic, dXCO₂ decreased in BTH (2.28 ± 0.73 ppm) and YRD (1.16 ± 0.43 ppm) but increased in PRD (0.28 ± 0.27 ppm). Compared to pre-pandemic levels, lockdowns saw dXCO₂ decrease slightly in YRD while increasing in BTH and PRD, reflecting differential responses of regional industrial structures. This study demonstrates the potential of seamless XCO₂ data for monitoring anthropogenic enhancement signals, and the proposed LISA-based method offers new support for regionally differentiated emission reduction assessments. Full article

Social media platforms host abundant and timely descriptions of medication experiences that can complement traditional pharmacovigilance systems. Yet the linguistic informality of these data presents a major challenge for mapping adverse drug event (ADE) expressions to standardized medical terminology. In this study, we developed BERT-based language models to classify ADE mentions from social media into MedDRA System Organ Classes (SOCs). Using the SMM4H and CADEC corpora, as well as their combination, we performed 20 iterations of 20% holdout validation for 3-, 6-, 22-, and 25-SOC classification tasks with a selected fixed training configuration (learning rate, batch size, and training epochs) based on training-loss convergence. The models achieved accuracies ranging from 75% to 94%, demonstrating strong performance for SOC-level classification of noisy and informal ADE expressions under the evaluated settings. These results are based on a controlled mention-level evaluation using deduplicated adverse drug event strings and do not establish document-level or real-world deployment generalization. This work provides a systematic evaluation of BERT-based models for SOC-level classification of ADEs and demonstrates consistent performance within the evaluated datasets and label granularities. While direct comparison with prior studies is limited by differences in datasets and evaluation protocols, the results demonstrate that transformer-based models can effectively classify ADEs into SOCs. These findings support the use of transformer-based normalization for SOC-level aggregation of user-reported adverse events and their integration into large-scale social media pharmacovigilance pipelines as a downstream component under controlled conditions. Full article

As multi-view datasets expand across diverse practical fields, feature selection (FS) has become an indispensable preparatory stage for machine learning models. Nevertheless, real-world multi-view data is often unlabeled and distributed among isolated clients, posing significant challenges to traditional centralized methods due to privacy concerns and communication constraints. Furthermore, existing centralized and federated approaches frequently suffer from entrapment in local optima and lack robust convergence guarantees. To address these issues, we propose Fed-MUFSHT, a federated framework for multi-view unsupervised FS (MUFS) that integrates tensor learning with a novel metaheuristic optimizer, Hierarchical-Cognitive Tianji’s Horse Racing Optimization (HC-THRO). Within the federated learning paradigm, Fed-MUFSHT follows a dual-stage local optimization process. Stage 1 applies HC-THRO, which integrates Hierarchical Competitive Learning and Adaptive Cognitive Mapping to simulate multi-level strategic competition and cognitive adaptation among individuals. This design enhances global exploration, adaptive learning, and fine-grained feature selection in high-dimensional spaces. Stage 2 employs a TL module based on canonical polyadic (CP) decomposition to perform missing-view imputation and refine latent representation learning. At the global level, a privacy-preserving aggregation strategy based on Normalized Mutual Information (NMI) and feature weights enables efficient model coordination without exposing raw data. Comparative experiments on several public benchmark datasets reveal that Fed-MUFSHT maintains clear advantages over strong competing methods, showing better optimization results together with more dependable convergence characteristics. The overall evidence suggests that the proposed approach is both robust and effective for distributed optimization tasks involving privacy protection. Full article

Electrotactile feedback is a compact approach for providing tactile cues in robotic teleoperation, but personalized calibration remains time-consuming because tactile perception varies across users. To address this problem, this study develops a subject-informed multi-layer finite element model of fingertip electric-field distribution coupled with a neural-response model and proposes a simulation-derived configuration-ranking method termed the Perceived Correctness Score (PCS). A gradient boosting regression model is then used to recommend among 36 candidate electrode diameter–spacing combinations. Validation was conducted using a custom-developed $3\times 2$ multi-channel fingertip electrotactile stimulation system in a shape/area recognition task involving six healthy subjects. The predicted PCS showed a moderate positive correlation with the measured mean recognition accuracy across configurations (Pearson $r=0.48$, $p<0.05$). The model achieved Top-1 exact matching for three of six subjects and Top-5 coverage for five of six subjects. Compared with conventional exhaustive psychophysical calibration, the proposed method reduced the average configuration time from 122.7 min to 16.0 min, corresponding to an efficiency improvement of 87.0%. These results show that model-guided ranking can substantially reduce the burden of individualized electrotactile configuration. Full article

This paper proposes a generalized high-order linear active disturbance rejection control (GHO-LADRC) method to suppress non-stationary disturbances in VICTS antenna direct-drive motors during high-dynamic scanning. First, a fourth-order generalized extended state observer is constructed, in which the derivative of the total disturbance is explicitly modeled as an extended state. This configuration enables real-time observation of the disturbance rate of change and suppresses the phase lag inherent in traditional ADRC during rapid disturbance variations through disturbance feedforward compensation. Secondly, drawing on singular perturbation theory and the motor’s dual-time-scale characteristics, this work precisely decouples and explicitly extracts the nonlinear friction and electromagnetic damping terms during the modeling stage. By integrating the extracted electromagnetic damping terms and the disturbance variation rate, an improved model-assisted control law is formulated, enabling active compensation for intense dynamic interference. Theoretical analysis and experimental results demonstrate that the proposed method significantly enhances disturbance rejection capability and satellite communication accuracy. As the first application of GHO-LADRC in the field of direct-drive VICTS antenna control, this work validates its effectiveness in improving system robustness within complex dynamic environments. Full article

Tube boilers are extensively employed in oil and gas refineries, as well as in petroleum, energy, and power generation industries, where they serve critical functions in local steam-generation units and combined-cycle gas turbine (CCGT) plants. However, these boilers are prone to defects arising from waterside corrosion (e.g., thinning of U-bend tubes), fireside corrosion, and material degradation caused by stress or creeping. Among these issues, wall thinning of tube bends is particularly severe, as it results in localized metal loss, reduced structural integrity, and an elevated risk of tube rupture or failure under high-temperature and high-pressure operating conditions. Such failures can significantly compromise boiler safety and efficiency, potentially leading to forced outages, costly unplanned repairs, or catastrophic damage if not detected in time. The current condition-monitoring policy for U-bends relies on scheduled preventive maintenance and unscheduled corrective interventions. In practice, this involves randomly checking approximately 10–20% of the tubes through spot scanning, partial scanning, or full scanning, with repairs typically carried out only after an undetected failure occurs. Such maintenance strategies generally require plant shutdowns, making the process time-consuming, labor-intensive, and ultimately not cost-effective. This paper reviews existing solutions, technologies, and research addressing the problem, and introduces femtosecond laser micromachined fiber optic sensors as a transformative approach for real-time monitoring of wall thickness reduction in U-bend boiler tubes, thereby opening pathways for further research. Full article

Smart care applications impose increasingly stringent requirements on low-latency execution, privacy preservation, and continuous monitoring. These requirements are driving intelligent services from cloud-centric architectures toward edge-side deployment. When multiple care-related workloads are deployed on resource-constrained edge devices, performance bottlenecks arise not only from model inference itself, but also from process scheduling, inter-process communication, and resource coordination overhead. To address this issue, this paper presents a task scheduling and management platform for multi-workload smart elderly care on a single pure-edge CPU–TPU heterogeneous node. The platform adopts a shared-memory and event-driven synchronization mechanism together with fine-grained process partitioning, thereby establishing a data-sharing and runtime-coordination framework for concurrent multi-workload execution. To evaluate the effectiveness of the proposed platform, experiments were conducted under single-workload, multi-workload, multi-resolution, and long-term runtime settings. The results show that, compared with two baseline schemes, the proposed platform improves the average frame rate by 66.7% and 71.1%, reduces net memory usage by 96.3% and 45.3%, and lowers net power consumption by 46.8% and 37.7%, respectively, under the single-workload setting. Under 10 concurrent workload instances, the system still maintains a stable frame rate of 42.03 ± 0.73 fps, demonstrating strong concurrency scalability. Multi-resolution experiments further indicate that the performance degradation at higher resolutions is mainly constrained by the front-end data supply stage. A continuous 10-day runtime experiment additionally verifies the sustained operating capability and resource stability of the platform under pure-edge deployment. These results demonstrate that node-level shared-memory and event-driven coordination can effectively improve the execution efficiency, scalability, and stability of real-time multi-workload analytics on such pure-edge heterogeneous nodes, providing a useful basis for future extensions to multi-node edge environments and edge–cloud collaborative task scheduling. Full article

Accurately quantifying the driving mechanisms of land surface temperature (LST) is fundamental to developing climate-resilient urban strategies. However, traditional linear models often fail to capture the complex nonlinear interactions and spatial non-stationarity inherent in urban thermal environments, especially when hindered by multicollinearity among morphological indicators. This study proposes a multi-scale spatial explainability attribution framework by integrating an XGBoost machine learning model with SHAP (SHapley Additive Explanations) to decipher the thermal dynamics of Changchun, a representative cold-region city in China. Utilizing a 500 m grid-based dataset, we incorporated 3D urban morphology (BVD), land cover (NDVI, NDWI), and socioeconomic factors. The results indicate that the XGBoost model achieves superior predictive performance (R² = 0.694) compared to traditional OLS models. SHAP global attribution identified Building Volume Density (BVD) as the primary warming driver, as its three-dimensional volume creates “thermal traps” through radiation trapping and reduced ventilation. Notably, NDVI exhibits a significant nonlinear “cooling threshold effect” at 0.3, beyond which its mitigation efficiency stagnates or even reverses due to vegetation fragmentation and heat-induced physiological stress. Furthermore, spatial mapping reveals a distinct “sign reversal” in NDWI’s impact, reflecting the dualistic thermal regulation of water bodies across different urban–rural gradients. These findings suggest that urban thermal management strategies should shift from merely restricting 2D surface occupancy (e.g., Building Density) to a more sophisticated approach focused on precisely controlling 3D volume intensity (BVD). This study provides a “point-to-area” diagnostic tool supporting a transition to spatially targeted urban planning interventions. Full article

Background/Objectives: This study aims to develop a deep learning (DL) model integrating ultrasound images and multidimensional clinical information to improve the diagnostic accuracy of breast non-mass lesions (NMLs). Methods: A total of 794 multicenter retrospective cases of NMLs were selected, stratified, and randomly divided into a training set (635 cases) and validation set (159 cases) at an 8:2 ratio. Multidimensional clinical information (including age, reproductive history, menstrual history, medical history, and findings from palpating the lesions) was incorporated to develop a DL model integrating ultrasound images and clinical data. To evaluate the diagnostic performance of the DL model, the area under the curve (AUC), accuracy, specificity, and sensitivity were employed. Results: The diagnostic model for NMLs integrating ultrasound images and multidimensional clinical information achieved an AUC of 0.8520 (95% CI: 0.7898–0.9068), F1 score of 0.7563, accuracy of 0.8176, sensitivity of 0.7031, and specificity of 0.8947. Its performance was superior to that of the model using only ultrasound images (AUC 0.8520 vs. 0.7571). SHAP analysis evaluating the reasons for the improved performance revealed that palpation with indistinct margins, abnormal axillary nodes, and older age were the three features with the highest contribution to predicting malignant risk. Conclusions: The DL model integrating ultrasound images and multidimensional clinical information demonstrated promising diagnostic performance in differentiating benign and malignant breast NMLs, suggesting the complementary value of multidimensional clinical information in the differential diagnosis of NMLs, though the reported AUC of 0.8520 is a preliminary internal estimate that awaits external validation. Full article

The impact of PM_2.5 pollution on the health inequality of vulnerable groups is a core issue in environmental justice research. However, existing studies in China mostly focus on severely polluted areas in northern China. They lack comparative cases in economically developed eastern regions. They also rarely consider changes in the impact of air pollution on residents’ health amid rapid urbanization. Based on multi-source data, this study employed spatial visualization, spatial autocorrelation analysis and spatial regression models. It investigated the impact of PM_2.5 pollution on the health inequality of vulnerable elderly groups in 92 districts and counties of Jiangsu Province from 2010 to 2020. The results show that: first, the regional pattern of health inequality between PM_2.5 pollution and vulnerable elderly groups in Jiangsu has continuously evolved, with a “lower in the south and higher in the north” pollution pattern and high overlap between high-pollution areas and high elderly health risk areas in northern Jiangsu. Second, the spatial coupling between PM_2.5 and elderly health risks has gradually strengthened, showing significant positive spatial agglomeration in 2020, confirming obvious spatial agglomeration characteristics of air pollution’s health impact. Third, the adverse health impact of PM_2.5 on vulnerable elderly groups became significant in 2020, exhibiting cumulative and lagged characteristics; urbanization and regional coordinated development have played a positive role in alleviating regional health inequality, while a lagging energy structure further exacerbates the health vulnerability of the elderly. This study fills the gap of insufficient research on economically developed eastern regions and provides targeted empirical references for urban refined governance and precise prevention and control of environmental health inequality. Full article

Accurate estimation of residential building energy consumption and associated CO₂ emissions is essential for refined urban carbon management. This study develops a hybrid framework that integrates physics-based simulation and machine learning to estimate residential building energy use and energy-related CO₂ emissions in Shenzhen in 2020. Representative building archetypes were first simulated and then used to train machine-learning models for large-scale applications. Building-level energy estimates were further combined with a bottom-up inventory to generate high-spatiotemporal-resolution maps of residential CO₂ emissions. The results show that: (1) the selected model achieved good accuracy and temporal robustness, with strong agreement between estimated and reference energy use at daily, monthly, and annual scales; (2) residential energy use was primarily driven by meteorological conditions, especially daily mean temperature and the duration of high-temperature conditions, and exhibited clear weekly and seasonal patterns, with higher values on weekends and in summer; (3) residential CO₂ emissions in Shenzhen reflected the combined effects of scale and intensity, with Longgang and Bao’an contributing the largest total emissions, Self-built residential buildings contributing the largest aggregate emissions, and Old residential buildings showing the highest average emissions per building; (4) emissions were highly concentrated in a small number of high-emission buildings, which were more frequently distributed along road-adjacent block perimeters. Overall, the proposed framework improves the fine-scale characterization of residential building CO₂ emissions and provides a useful basis for hotspot identification and targeted mitigation. Full article

Hydrogen trapping at carbide/matrix interfaces is important for improving the resistance of steels to hydrogen embrittlement. In this work, the segregation behavior of hydrogen at the coherent α-Fe/V₄C₃ interface was investigated by first-principles calculations. Representative hydrogen sites were considered systematically, including interstitial sites in the near-interface region, interfacial sites, and carbon-vacancy sites in V₄C₃. All of the sites examined are energetically favorable for hydrogen trapping, but the carbon vacancy inside V₄C₃ exhibits the strongest trapping tendency. Charge density, Bader charge, and density-of-states analyses indicate that hydrogen at this site gains more electrons and forms stronger interactions with neighboring V atoms, leading to enhanced stability. The behavior of H₂ at the internal carbon vacancy was also evaluated. After structural relaxation, the H₂ molecule dissociated into two separate H atoms, indicating that hydrogen is more stably trapped in atomic rather than molecular form. These findings reveal the crucial role of carbon vacancies in regulating hydrogen trapping at the α-Fe/V₄C₃ interface and provide atomic-scale insight into the hydrogen trapping mechanism of vanadium carbide precipitates in steels. Full article

Severe heavy metal contamination affects the karst landscapes of Guangxi Zhuang Autonomous Region, China, which are highly polluted and complex. However, integrated assessments of heavy metal sources, distribution, ecological risks, and speciation in karst agricultural soils remain limited. Additionally, there is a gap regarding the interactions between cadmium (Cd), zinc (Zn), and selenium (Se) in natural rice fields. This study employed the pollution load index (PLI), ecological risk index (RI), and Positive Matrix Factorization (PMF) models to evaluate the sources and characteristics of heavy metal contamination in farmland soils. The results showed significant pollution in agricultural soils of Guangxi karst due to Cd, chromium (Cr), copper (Cu), and nickel (Ni). Among these, Cd poses the highest ecological risk. Heavy metal accumulation in the surface soil far exceeds that in deeper layers, and the main sources of Cd were contributed from soil parent material and agricultural activities. Speciation analysis revealed the high bioavailability of Cd, while Zn and Se existed in more stable forms. Despite elevated soil Cd levels, rice grains remained within the safety limits. Using transmission electron microscopy (TEM), Cd was primarily detected in the cell walls of rice stems and husks, which was attributed to Zn’s competitive uptake, reducing Cd absorption and Se forming complexes with Cd to enhance its fixation. Statistical correlations revealed positive associations between Cd in soil and rice. Cd also demonstrated a positive correlation with Se, but a negative correlation with Zn, suggesting a synergistic mechanism between Zn and Se that acts to mitigate the absorption of Cd. This study provides practical guidance for managing farmland soil heavy metal contamination and protecting agricultural soil resources in the karst areas. Full article

The river buffalo is an economically important livestock species supplying milk and meat. However, a multi-tissue transcriptomic atlas for the key dairy river buffalo breeds, Murrah and Nili-Ravi, has not yet been established, and the lack of stable reference genes has hindered in-depth studies of their biological functions and the molecular mechanisms underlying key economic traits such as lactation. We established a multi-tissue gene expression atlas across 20 tissues and identified 717 housekeeping genes (HKGs), and RPL37A and EEF2 were further shown to be stable candidate reference genes under the conditions tested. We found 8368 tissue-specific genes (TSGs), predominantly enriched in the reproductive system. Exploratory analysis of mammary tissue (dry-period vs. public lactating samples, confounded by batch effects) revealed mammary-enriched hub genes including LALBA; these findings are preliminary and require validation. Dynamic analysis across lactation stages (early, peak, mid-, late) identified candidate genes including SEC14L2 and ACSM3. Phenotypic data showed strong negative correlations between milk yield and protein/fat content, and a positive correlation with lactose content. However, causal or regulatory roles were not inferred due to lack of paired individual-level data. Cross-dataset comparisons are descriptive only, and are not key conclusions. In summary, this study lays the foundation for advancing research in lactation trait genetics and functional genomics in river buffalo, with novel reference genes and lactation stage-specific transcriptional dynamics as its main contributions. Full article