Search Results (16,788)

The soundscape serves as a critical determinant of the quality of urban wetland parks. This study employs a mixed-methods approach to comprehensively evaluate wetland soundscapes. First, field investigations combining sound level measurements and questionnaire surveys were conducted in Aixi Lake Wetland Park to analyze the spatiotemporal characteristics of the soundscape. Second, laboratory-based physiological tracking (using wearable sensors) and cognitive tests (Sustained Attention to Response Task, SART) were utilized to experimentally quantify the restorative benefits of typical soundscapes. The findings reveal that: (1) sound level indicators and sound harmonious degree in urban wetland parks exhibit significant spatiotemporal characteristics and distributional variations; (2) a marked competitive effect among biological, geophysical, and human activity sounds is observed in their spatial distribution; sound harmonious degree demonstrates significant spatial autocorrelation in both global and local models; (3) different sound sources possess varying restorative potentials, with bird song showing the highest restorative effect; the SHDs of biological and geophony, along with L_Aeq, are key factors affecting PRSS; (4) a positive correlation exists between L_Aeq and the PRSS up to 56.4 dB, beyond which PRSS declines with increasing L_Aeq; (5) at the physiological level, short-term exposure to urban wetland park soundscapes can rapidly alleviate stress, with the most pronounced restorative effects occurring within the first 60 s; and (6) in terms of attention, soundscape stimulation reduces SART response times and improves response speed, while bird song from treetops and musical sounds further decrease response errors. Full article

Shared GPU clusters often execute multiple distributed training jobs concurrently under fluctuating contention. We reinterpret this setting as a two-scale control problem, where the micro scale captures intra-job learning dynamics and the macro scale captures inter-job resource arbitration. We propose an entropy-guided hierarchical framework that links these two scales through a unified uncertainty signal computed from training logits. Unlike existing uncertainty-aware methods that typically use uncertainty for only a single level of decision making, our approach uses the same entropy-based signal to jointly support both intra-job adaptation and inter-job scheduling within a hierarchical control loop. At the micro level, each worker estimates predictive uncertainty via normalized entropy and converts it into stable weights that drive epoch-level controls for uncertainty-aware data sharding, fixed-budget batch-size reallocation, and learning-rate modulation, while remaining compatible with standard synchronous data-parallel training. At the macro level, the same signal is aggregated into a job utility score that guides admission, ordering, and GPU quota assignment under contention. In large-scale workload-driven simulation, our method reduces average job completion time (JCT) by 23.7% and shortens cluster makespan by 15.7% relative to a strong learning-unaware baseline, demonstrating that uncertainty-aligned scheduling can improve cluster-level efficiency while preserving training correctness. We further validate scalability using a calibrated simulator up to 1024 nodes. Full article

The rapid electrification of transportation and proliferation of distributed energy resources (DERs) are transforming distribution grids into highly dynamic, data-intensive, and cyber-physical systems. While reinforcement learning (RL), multi-agent coordination, and edge computing offer powerful tools for adaptive control, their deployment in safety-critical utility environments raises concerns regarding stability, certification compatibility, cyber-resilience, and regulatory acceptance. This paper presents an architecture-centric framework for edge-intelligent and cyber-resilient coordination of electric vehicles (EVs) and DERs that reconciles adaptive learning with deterministic safety guarantees. The proposed hierarchical edge–cloud architecture integrates multi-agent system (MAS) coordination, constraint-invariant reinforcement learning, and embedded cybersecurity mechanisms within a structured control hierarchy. Learning-enabled edge agents operate exclusively within standards-compliant safety envelopes enforced through supervisory constraint projection, control barrier functions, and Lyapunov-consistent stability safeguards. Protection-critical functions remain deterministic and isolated from adaptive layers, preserving compatibility with IEEE 1547 and existing utility protection schemes. The framework further incorporates anomaly triggered policy freezing, fail-safe fallback modes, and communication-aware resilience mechanisms to prevent unsafe transient behavior in non-stationary, distributed environments. Unlike simulation-only learning approaches, the architecture embeds progressive validation through software-in-the-loop (SIL), hardware-in-the-loop (HIL), and power hardware-in-the-loop (PHIL) testing to empirically verify transient stability, constraint compliance, and cyber-resilience under realistic timing and disturbance conditions. Beyond technical performance, the paper situates edge intelligence within standards evolution, governance structures, workforce transformation, techno-economic assessment, and equitable deployment pathways. By framing adaptive control as a bounded, auditable augmentation layer rather than a disruptive replacement for certified infrastructure, the proposed architecture provides a pragmatic roadmap for evolutionary modernization of distribution systems. Full article

Cross-domain scene classification aims to mitigate the distribution discrepancy between domains through domain adaptation techniques. With the rapid advancement of Vision–Language Models (VLMs), utilizing them for cross-domain scene classification has emerged as a promising research direction. Current methods utilize domain-specific prompts to facilitate domain adaptation through the CLIP model. However, for remote sensing images, the considerable differences in visual features across domains pose significant challenges for learning domain-specific prompts, leading to suboptimal cross-domain performance. In addition, they cannot reduce the domain shift that exists between the source domain and the target domain. To address the above challenges, we propose a novel cross-domain scene classification method, DIPVR (Domain-Invariant Prompts and Visual Representations), which enhances model performance by learning domain-invariant features for both prompts and visual representations. Specifically, we propose learning domain-invariant prompts and introducing prior knowledge to guide the prompt-learning process. To learn domain-invariant visual representations, we propose a Visual Invariant Learning module that adaptively extracts the shared features between the source and target domains. Finally, visual features are matched with context features to align the domain distributions between the source and target domains. The experimental results on the cross-domain scene classification datasets demonstrate that our proposed method outperforms the baseline methods, achieving optimal cross-domain transfer performance. Full article

Robust vehicle detection in real-world traffic surveillance remains challenging due to degraded imagery caused by motion blur, adverse weather, and low illumination, which significantly increases detector sensitivity to hyperparameter configurations. This study proposes a “Frugal AI” distributed multi-GPU framework that optimizes hyperparameters via a stochastic simplex-based search coupled with five-fold cross-validation. Utilizing three low-cost NVIDIA GTX 1050 Ti GPUs, the framework performs parallel candidate exploration with an asynchronous model-level exchange mechanism to escape local optima without the overhead of gradient synchronization. Seven CNN backbones—VGG16, VGG19, GoogLeNet, MobileNetV2, ResNet18, ResNet50, and ResNet101—were evaluated within YOLOv2 and Faster R-CNN detectors. To address memory constraints (4 GB VRAM), YOLOv2 was selected for extensive benchmarking. Performance was measured using a harmonic precision–recall-based cost metric to strictly penalize imbalanced outcomes. Experimental results demonstrate that under identical wall-clock time budgets, the proposed framework achieves an average 1.38% reduction in aggregated cost across all models, with the highly sensitive VGG19 backbone showing a 4.00% improvement. Benchmarking against Bayesian optimization, genetic algorithms, and random search confirms that our method achieves superior optimization quality with statistical significance ($p < 0.05$). Under a rigorous IoU = 0.75 threshold, the optimized models consistently yielded F1-scores 0.8444 $\pm$ 0.0346. Ablation studies further validate that the collaborative model exchange is essential for accelerating convergence in rugged loss landscapes. This research offers a practical, scalable, and cost-efficient solution for deploying robust AI surveillance in resource-constrained smart city infrastructure. Full article

This paper presents a deterministic embedded control architecture for an industrial electroplating line. The validated system includes two autonomous trolleys, 18 station-aligned process positions, shared-track motion, and redundant grouped baths. The proposed controller addresses the limitations of rigid sequential automation by combining asynchronous finite-state trolley execution, runtime allocation of equivalent technological stations, dwell-time-preserving retrieval, distributed thermal supervision, and layered fail-safe protection within a single ATmega2560-based implementation. The core contribution is the integration of virtual process groups and temporal FIFO logic into a compact plant-side embedded controller. This enables adaptive bath selection and process-completion-based retrieval without reliance on a real-time operating system or a computationally heavy supervisory runtime. The architecture also incorporates predictive pre-start validation, runtime software arbitration, hardware-wired interlocks, binary-coded trolley positioning, and a distributed 1-Wire thermal measurement network. Validation was performed in a controller-centered hardware-in-the-loop representation of an 18-station zinc electroplating line. Over a 100-batch horizon, the proposed architecture reduced makespan from 1642 min to 1244 min, corresponding to a 24.2% throughput improvement. Average trolley idle time decreased from 18.4 min/batch to 4.1 min/batch. Grouped-bath utilization increased from 64% to 91%, while tracked bottleneck incidents decreased from 18 to 2. These results show that adaptive, resource-aware, and safety-layered electroplating control can be realized effectively on a compact embedded platform in an industry-representative HIL setting, while preserving dwell-time integrity and controller-level safety invariants. Full article

With the exponential growth of natural gas reserves and utilization scale in China, underground gas storage (UGS) facilities—critical infrastructure within the natural gas production-supply-storage-sales system—have entered a phase of rapid expansion. As the core component connecting subsurface reservoirs with surface systems, wellbore integrity directly influences operational safety and service lifespan of UGS facilities. However, current leakage detection and integrity analysis methodologies for gas storage wellbores remain deficient in effective real-time monitoring capabilities. Traditional methods, however, are constrained by limited spatial coverage and insufficient precision, rendering them inadequate for comprehensive, continuous safety monitoring requirements. To address this industry challenge, this study proposes a real-time wellbore integrity monitoring framework based on distributed fiber optic sensing technology, integrating distributed temperature sensing (DTS) and distributed acoustic sensing (DAS) devices into a synergistic monitoring system. The DTS component enables preliminary localization of potential leakage points through detection of minute temperature anomalies along the wellbore, while the DAS unit accurately identifies acoustic signatures caused by gas leakage within casings via monitoring of acoustic vibration signals propagating along the optical fiber. Through joint analysis of DTS and DAS data streams, real-time diagnosis of wellbore leakage events and integrity status can be achieved. Field trials demonstrated that this hybrid monitoring system achieved leakage localization accuracy within 1.0 m, effectively distinguishing normal operational signals from abnormal leakage characteristics. During actual monitoring operations, no indications of wellbore integrity compromise were detected; only minor noise and interference signals originating from surface construction activities were observed. Full article

This paper presents a comprehensive life-cycle analysis of distribution transformers, based on realized measurements of the increased power losses as a result of their long-term service under real-world conditions. The study is based on aggregated measured data from extensive fleets of oil-immersed distribution transformers characterized by diverse designs, manufacturing vintages, and service lives. The evolution of no-load losses and short-circuit losses is analyzed as a function of operational duration, structural characteristics, and the specific technologies employed for windings and magnetic core construction. Statistical models describing the variation in these losses are presented, highlighting the limitations of the static assumptions commonly utilized in power distribution network planning. On this basis, an approximation of the time evolution of the transformer’s total power and energy losses is proposed as appropriate for implementation in a life-cycle analysis model. Furthermore, the impacts of thermal loading and abnormal operating conditions—such as unbalanced loads, frequent short circuits, and repeated overheating of the transformer oil—are analyzed as drivers of accelerated transformer aging. These effects are integrated into a unified life-cycle framework, enabling the quantitative assessment of loss variations and their associated operational expenditures (OPEX). A numerical example is provided to evaluate the cost-effectiveness of “repair vs. replacement” scenarios, utilizing a discounted cash flow analysis that incorporates a carbon component. The findings establish a methodological foundation for a broader assessment of technical condition and energy performance, identifying the optimal intervention point for repair or replacement to support decision-making for Distribution System Operators (DSOs) amidst increasing requirements for efficiency and decarbonization. Full article

Extreme class imbalance is a persistent obstacle for machine learning-driven intrusion detection, as rare but high-impact cyberattacks occur far less frequently than benign traffic in training data. In many real-world cybersecurity datasets, this imbalance becomes extreme, with certain attack types containing a handful of samples, effectively placing the problem in a few-shot learning regime. This paper presents a controlled benchmarking study of Generative Adversarial Network (GAN) objectives for synthesizing minority-class cyberattack data. Using the UWF-ZeekData22 network traffic dataset, each MITRE ATT&CK tactic is framed as a separate binary detection task, and tactic-specific GANs are trained solely on minority samples to generate synthetic attack records. Four widely used GAN variants—Vanilla GAN, Conditional GAN (cGAN), Wasserstein GAN (WGAN), and Wasserstein GAN with Gradient Penalty (WGAN-GP)—are compared under unified training steps and fixed augmentation conditions. The utility of generated data is assessed by evaluating downstream detection performance using five traditional classifiers: Logistic Regression, Support Vector Machine, k-Nearest Neighbors, Decision Tree, and Random Forest. The results indicate that GAN augmentation generally strengthens minority-class detection across tactics and models, reducing false negatives and improving recall consistency, while not systematically harming majority-class performance. However, the effectiveness of each GAN objective varies significantly with data sparsity. Specifically, simpler adversarial objectives often outperform more complex architectures by preserving discriminative feature structure, while heavily regularized models may overly smooth minority-class distributions and reduce separability. Wasserstein-based objectives provide improved training stability, but additional regularization does not consistently translate to better detection performance. Overall, the results demonstrate that in extreme-imbalance settings, GAN effectiveness is governed more by data sparsity and structure preservation than by architectural complexity. These findings establish class-specific generative augmentation as a practical strategy for intrusion detection and provide empirical guidance for selecting appropriate GAN objectives for tabular cybersecurity data under highly imbalanced conditions. Full article

Background/Objectives: Loss of nuclear BAP1 (BRCA1-associated protein 1) expression is a well-established adverse prognostic marker in uveal melanoma (UM). However, data from Central and Southeastern European populations are limited. This descriptive study aimed to evaluate BAP1 immunohistochemical expression in a Croatian enucleation-based UM cohort, characterize its associations with clinicopathological parameters, and contextualize the findings within the published literature. Methods: Formalin-fixed, paraffin-embedded tumor tissue from 58 consecutive patients with primary choroidal and ciliary body melanoma treated with enucleation at University Hospital Centre Zagreb (2006–2016) was analyzed immunohistochemically for BAP1 nuclear expression. Associations with clinicopathological parameters were assessed using chi-square and Fisher’s exact tests. Survival analysis was performed using Kaplan–Meier estimation, log-rank tests, and Cox proportional hazards regression with a median follow-up of 11.2 years. Results: Loss of nuclear BAP1 expression was observed in 53/58 (91.4%) specimens, resulting in a severely imbalanced distribution (53 versus 5 patients) precluding meaningful comparative survival analysis. Five-year and 10-year overall survival rates were 72.4% and 51.7%, respectively, with a median overall survival of 14.5 years. BAP1 loss was associated with longer disease-free survival (log-rank p = 0.020); however, this finding likely reflects a statistical artifact attributable to the extremely small BAP1-retained group (n = 5) harboring concurrent adverse features and should not be interpreted biologically. The study was underpowered to draw prognostic inferences regarding BAP1 status. Exploratory survival analyses are presented for transparency but should not be interpreted inferentially. Conclusions: The exceptionally high prevalence of BAP1 loss reflects the selection bias inherent in enucleation-based cohorts, which are enriched for large, molecularly high-risk tumors. This study provides the first comprehensive BAP1 immunohistochemical data from Croatia, contributing to the growing evidence that enucleation cohorts represent a distinct, biologically high-risk subgroup in which BAP1 immunohistochemistry offers limited discriminatory value. The extended follow-up of 11.2 years confirms the prolonged natural history of UM. Future multi-center studies incorporating molecular validation and diverse treatment modalities are needed to establish the prognostic utility of BAP1 across the full spectrum of UM disease. Full article

Urban water distribution systems (WDSs) face increasing threats from accidental or intentional contaminant intrusion events. While contamination warning systems using water quality sensors enable early detection and rapid response to contamination events, traditional sensor placement approaches often rely on a single or limited performance metric, overlooking the multidimensional nature of system resilience. This study presents a multidimensional resilience-based framework for the optimal placement of water quality sensors in urban WDSs, integrating hydraulic and water quality simulations using the EPANET-MATLAB toolkit with a genetic algorithm (GA) optimization process. For Anytown Water Distribution Network, four distinct functionalities were formulated to capture different aspects of system performance during contamination events, and an integrated-multidimensional resilience metric was proposed as a collective measure. Results demonstrated that the optimal sensor configurations varied significantly depending on the selected functionality. However, the integrated multidimensional resilience-based approach yielded more balanced and effective sensor placements, simultaneously enhancing resilience levels for all individual functionalities. Furthermore, the findings indicated that adding more sensors beyond a certain number offers marginal improvements in system resilience, suggesting that sensor deployment should be guided by monitoring objectives (e.g., resilience) rather than simply increasing sensor numbers. The findings and discussion suggest practical insights for utilities to enhance water supply services with safe quality and system security against contamination threats in urban WDSs. Full article

Soybean (Glycine max [L.] Merr.) is a globally important crop with a rapidly expanding body of genomics literature driven by advances in sequencing and functional genomics. Thousands of studies reference soybean genes using standardized Glyma identifiers; however, systematic analyses of how these identifiers are distributed across chromosomes in the scientific literature remain limited. Here, we present a chromosome-resolved bibliometric analysis of soybean gene mentions using a reproducible rule-based text mining approach. PubMed abstracts published between December 2006 and December 2025 were mined for standardized Glyma gene identifiers using regular-expression-based entity extraction. A total of 377 PubMed records were retrieved, of which 340 abstracts (90.2%) contained at least one Glyma gene identifier. The median number of unique genes mentioned per abstract was 1, with a maximum of 14 genes reported in a single study. Our results reveal three major patterns. First, soybean genomics research remains predominantly gene-centric, with most abstracts referencing one or two genes. Second, apparent chromosome-level disparities exist in literature representation within the subset of studies using standardized Glyma identifiers, with chromosomes 3 and 16 exhibiting the highest frequencies of unique gene mentions. A Chi-square goodness-of-fit test confirmed that these differences deviate significantly from a uniform distribution (χ² = 123.71, p < 0.001), indicating non-random patterns of gene reporting. Third, a small subset of genes dominates the literature, while the majority of annotated genes are mentioned infrequently, reflecting a long-tailed distribution of research attention. This analysis captures reporting patterns in studies that explicitly use standardized Glyma identifiers and therefore represents a defined subset of the broader soybean genomics literature. Within this scope, the findings highlight uneven adoption of standardized gene nomenclature and chromosome-level differences in research emphasis. More broadly, this study demonstrates the utility of transparent, rule-based text mining approaches for large-scale bibliometric analyses in plant science and provides a scalable framework for comparative analyses across crop species. Full article

Accurate mapping of lunar mineral distributions is essential for understanding the Moon’s origin and evolution and for enabling future in situ resource utilization (ISRU). Yet mineralogical inversion from orbital hyperspectral observations remains challenging due to limited spatial resolution, complex photometric conditions, and sparse returned samples. We present PGU-Net, a Hapke physics-guided deep autoencoder for nonlinear blind unmixing of lunar hyperspectral data. The encoder adopts a dual-attention design to enhance discriminative spectral features. The decoder performs linear mixing in the SSA domain and then reconstructs reflectance through a lightweight nonlinear module, while physics-consistent losses encourage radiative-transfer plausibility. Experiments on a synthetic lunar regolith dataset demonstrate that PGU-Net achieves consistently lower endmember SAD and abundance aRMSE than representative baselines across multiple noise levels. Additional validations on the terrestrial AVIRIS Cuprite benchmark and on Moon Mineralogy Mapper (${\mathrm{M}}^3$) observations near the Chang’e-5 (CE-5) and Chang’e-6 (CE-6) landing regions yield physically plausible mineral distributions. The ${\mathrm{M}}^3$ maps are broadly consistent with Kaguya MI mineral products and returned-sample constraints, supporting the practicality of PGU-Net for lunar mineralogical mapping. Full article

Accurate multivariate long-term time series forecasting (LTSF) is critical for smart grid operations. However, non-stationary distribution shifts frequently induce compounding error accumulation in conventional architectures. This study proposes the Mamba-LSTM-Attention (MLA) framework, a distribution-aware architecture engineered for forecasting stability. The pipeline integrates Reversible Instance Normalization (RevIN) to neutralize statistical drift. To address computational bottlenecks, the architecture utilizes a linear-time Selective State Space Model (Mamba) to capture global trend dynamics, cascaded with a single-layer gated Long Short-Term Memory (LSTM) unit to model localized non-linear residuals. A terminal information bottleneck structurally bounds cross-step error propagation. Empirical results across standard ETT and Electricity benchmarks reveal a precision–stability trade-off. By prioritizing structural resilience, the MLA framework limits error accumulation on highly volatile datasets, yielding MSEs of 0.210 and 0.128 on ETTh2 and ETTm2 at the T = 96 horizon. This structural bottleneck inherently smooths high-frequency periodic patterns, yielding lower absolute accuracy on stationary benchmarks such as ETTh1 and ETTm1. Ultimately, the architecture establishes a computationally efficient, structurally stable baseline tailored for non-stationary anomaly tracking in smart grids. Full article

Polyethylene terephthalate is a prominent polymer known for its mechanical properties, chemical resistance, and recyclability, and it is widely utilized across various industries. Enhancing the properties of polyethylene terephthalate (PET) through nanocomposite technology, particularly with the inclusion of nanoscale fillers, has garnered significant attention. This study investigates synthetic layered double hydroxides (LDHs), specifically MgAl LDH modified with calcium dodecylbenzene sulphonate in n-butyl alcohol (CDS) organic surfactant, as an alternative to natural clays for PET nanocomposites. Additionally, modified LDH serves a dual role as both a catalyst and a dispersive agent, promoting effective exfoliation within the PET matrix. A polymerization process was employed to ensure proportional and effective dispersion of the nanofillers, addressing the critical challenge of achieving uniform distribution. The resulting nanocomposites demonstrated superior mechanical strength, thermal stability, and barrier properties compared to traditional intercalated counterparts. Moreover, synthetic LDHs present a more sustainable solution, reducing the environmental footprint associated with natural clay mining, which includes land degradation, water pollution, energy consumption, and biodiversity loss. This research provides a promising pathway for developing high-performance, environmentally friendly PET nanocomposites, with significant implications for various industrial applications, from packaging to automotive and electronics. The findings highlight the potential of synthetic LDHs to advance material science while aligning sustainable development goals. Full article