Search Results (3,053)

Extended target tracking requires both accurate shape representation and efficient recursive estimation. In non-ellipsoidal extended target tracking, ellipsoidal random-matrix models are computationally efficient and suitable for Bayesian recursion, but they mainly describe the overall spatial dispersion of measurements and cannot represent local contour variations such as protrusions and concavities. In contrast, non-ellipsoidal contour models provide stronger shape representation but usually introduce higher computational complexity and stronger prior assumptions. To address this trade-off, this paper proposes an extension-difference-mapping-based Poisson multi-Bernoulli mixture filter, termed EDM-PMBM, for non-ellipsoidal extended target tracking. First, each local Bernoulli component carries a Fourier-based contour estimate and an ellipsoidal baseline propagated from the previous posterior. At the current scan, the predicted EDM function is used to map each candidate measurement subset into the EDM domain, where the EDM-induced GGIW likelihood is evaluated for PMBM data association. After the association is determined, the assigned measurement subset is used to update the posterior contour, the EDM ratio, and the EDM-domain state. The updated EDM information is then propagated to subsequent scans. In this way, shape differences are introduced into likelihood evaluation and data association without changing the basic recursive structure of the PMBM filter. Simulation results in two scenarios show that the proposed EDM-PMBM filter achieves lower GOSPA error than the compared filters and maintains more stable tracks in dense crossing situations. These results indicate that the proposed method improves the discrimination ability for non-ellipsoidal extended targets. Full article

In Underwater Acoustic Target Recognition (UATR), accurately extracting spectral lines from time–frequency spectra in complex ocean environments faces three critical challenges: low-frequency spectral confusion, line spectrum and noise mixture, and a computational efficiency vs. performance trade-off. To address these, we propose a deep learning-integrated framework based on application-oriented integration and adaptation of established techniques tailored to the underwater acoustic domain. The framework consists of the following: (1) the Line Spectrum Separation Network (LSS-Net), which integrates a Time–Frequency Joint LSTM and a Temporal Gated Cross-Attention (TGCA) module within an encoder–decoder architecture adapted for high-resolution underwater acoustic time–frequency spectra; (2) a physics-informed signal simulation approach that realistically models Doppler frequency drift and intensity fluctuations; and (3) a Peak-Tracking Line Extractor (PTLE) algorithm that leverages underwater acoustic-specific temporal constraints. The proposed framework achieves an MOTA of 0.89 on simulated data and 0.52 on real sea trial data, outperforming existing methods by 0.06-2.14 in MOTA and significantly suppressing high-resolution background noise. Full article

This paper examines stock market reactions to corporate blockchain adoption announcements and explores whether the visibility of such initiatives shapes investor response. While prior research documents strong valuation effects during early phases of technological hype, evidence from more mature stages of diffusion remains limited. Accordingly, this study provides exploratory evidence on investor behavior in a later-stage adoption context. We construct a hand-collected dataset of 51 announcements by publicly traded firms across multiple industries and employ a standard event-study methodology to estimate abnormal returns over short announcement windows, using both market-model and Fama–French factor specifications. Adoption visibility is conceptualized as a multidimensional construct capturing (i) the intensity of communication surrounding the initiative and (ii) whether the application is customer-facing or internally oriented. The results indicate that average abnormal returns around announcement dates are positive but economically modest and statistically insignificant. These findings suggest that blockchain adoption announcements no longer trigger uniform market repricing effects. Instead, investors appear to respond more selectively, potentially differentiating based on the perceived informational content and strategic relevance of the initiatives. Overall, the analysis offers exploratory evidence consistent with a shift in investor response as emerging technologies move beyond hype-driven phases toward more mature stages of diffusion. The results should be interpreted with appropriate caution and motivate further research using larger samples and complementary empirical approaches. Full article

With the rapid growth of data center (DC) energy consumption and the large-scale integration of renewable energy, DCs increasingly face time-varying power upper-bound constraints jointly shaped by grid power supply capability, renewable energy fluctuations, and demand response mechanisms. Meanwhile, DC power consumption exhibits a typical information-load-driven characteristic. The computing tasks hosted by virtual machines affect server-side IT power consumption through resource utilization states such as CPU, memory, disk I/O, and network I/O, and are further coupled with non-IT auxiliary power consumption from cooling, power distribution, and networking equipment. In such cyber–physical operation scenarios, physical-layer sensing data and hypervisor-level virtualization monitoring data jointly provide the state basis for power estimation, power warning, and migration decisions. To address the mismatch between dynamic power upper bounds and time-varying information loads, this paper investigates the information load scheduling problem under constrained power loads and proposes a two-stage virtual machine (VM) migration optimization framework. In the VM selection stage, a Multi-Factor Balanced (MFB) algorithm is designed. By introducing a warning-line trend model based on the arctangent function, MFB comprehensively considers resource utilization, power load variation trends, and service level agreement (SLA) violation levels to dynamically identify candidate VMs for migration. In the VM placement stage, a Multi-Factor Equilibrium Ant Colony Optimization (MFEACO) algorithm incorporating a Random Roulette Wheel (RRW) selection mechanism is proposed. By constructing normalized multi-dimensional equilibrium factors, MFEACO coordinates the trade-off among energy consumption, load balancing, and SLA violations. Simulation experiments are conducted on an improved CloudSim platform using real-world cluster trace data from Google and Alibaba. The results show that, while satisfying dynamic power constraints, the proposed MFB–MFEACO framework achieves a favorable comprehensive trade-off among energy consumption control, SLA violation suppression, and migration reduction. Compared with traditional heuristic methods and a power-constrained genetic algorithm baseline, the proposed framework demonstrates better dynamic adaptability and scheduling stability. Full article

In response to the challenges in UAV-oriented ground small-object localization and detection, including the easy loss of tiny target features, insufficient scale adaptability, severe interference from complex backgrounds, as well as high missed and false detection rates and the inadequate localization accuracy of the conventional YOLOv11n model in such scenarios, this paper takes YOLOv11n as the basic framework and performs systematic optimization from three aspects, network structure, core modules, and feature enhancement, proposing a lightweight small-object-enhanced detection algorithm named YOLOSO for UAV applications. By introducing a P2 high-resolution feature branch with a stride of 4, a four-scale detection structure consisting of P2-P3-P4-P5 is constructed, which reduces the minimum detection stride from 8 to 4 and alleviates the loss of detailed feature information for ultra-tiny targets. A bidirectional “top-down + bottom-up” multi-scale feature fusion strategy is utilized to improve the complementation between deep semantic information and shallow detailed features, while the core modules C3k2SO and C2PSASO are optimized and redesigned, respectively; by adjusting the channel compression ratio (0.25 for shallow modules and 0.75 for deep modules in C3k2SO; 0.25 in C2PSASO), optimizing the convolution kernel configuration (combining 1 × 3 and 3 × 1 convolutions), increasing the number of attention heads (from 4 to 8), and introducing residual connections with a 1 × 1 convolutional branch, the refinement and focusing ability of small-object feature extraction are improved. Additionally, an Enhanced Dual-branch Convolutional Block Attention Module (ED-CBAM) is proposed to further suppress background interference. Experimental results on the VisDrone2019-DET dataset demonstrate that the proposed YOLOSO contains 3.56M parameters and maintains a lightweight structure, attaining P, R, and mAP50 values of 47.2%, 36.8%, and 37.3% in the test set, which are 4.5 percentage points, 4.8 percentage points, and 3.7 percentage points higher than those of the baseline YOLOv11n (42.7%, 32.0% and 33.6%), respectively. Meanwhile, the medium-to-large version YOLOSO-S (14.85M parameters, 45.3% mAP50) reduces the number of parameters by 53.6% compared with the same-scale Rtdetr-L (32.0M) while achieving significantly better performance (37.8% mAP50). Experiments on the DOTAv1 dataset further confirm the generalization of YOLOSO, achieving 62.2% precision and 27.3% mAP50, outperforming all compared YOLO models. Evaluated on the DOTA-v1 dataset, YOLOSO achieves a feasible FPS of 20.53. Although slightly slower than mainstream lightweight YOLO models, the substantial accuracy gains fully offset the minor inference speed loss, and such performance trade-off is acceptable for practical UAV deployment. Ablation experiments verify that structural optimization (2.8 percentage points mAP50 improvement, from 33.6% to 36.4%) and the proposed C2PSASO (0.7 percentage points mAP50 improvement to 34.3%) and C3k2SO (1.4 percentage points mAP50 improvement to 35.0%) modules all contribute positive performance gains with favorable complementarity. While retaining lightweight characteristics, the model effectively enhances the detection accuracy of small objects in unmanned aerial vehicle scenarios and can provide technical references for practical applications such as remote sensing monitoring and security patrolling. Full article

Background: Multimodal breast ultrasound, including B-mode imaging, color Doppler flow imaging, and elastography, provides complementary information for lesion characterization. However, effectively integrating heterogeneous modalities remains challenging due to inconsistent feature distributions, limited cross-modal interaction, computational cost in existing methods, and sensitivity to noise and missing data. Methods: We presented an efficient Cross-Modal Interaction and Dynamic Fusion Network (CIDFNet) for multimodal breast ultrasound analysis. The framework integrates a multi-scale feature enhancement module to improve modality-specific representations, a cross-modal interaction module to enable early-stage feature exchange across modalities, and a dynamic fusion strategy to adaptively combine modality information based on feature reliability estimation. In addition, an invertible neural network is incorporated to reconstruct missing modality features during training. Results: Experiments on an internal dataset of 248 patients with 1532 images show that CIDFNet obtains an AUC of 85.69%, accuracy of 75.51%, recall of 50.00%, F1-score of 62.50%, and precision of 83.33%, while requiring 49.51 M parameters and 79.79 G FLOPs, respectively. Under a simplified Gaussian noise perturbation setting, performance degradation is observed. Conclusions: CIDFNet presents a framework for multimodal breast ultrasound analysis that reflects a trade-off between performance and computational efficiency. Full article

Large Language Models (LLMs) have emerged as transformative tools across various domains, exhibiting remarkable capabilities in natural language processing and generation. However, their reliance on static pre-training data limits their ability to access up-to-date and domain-specific information. The existing research often treats augmentation strategies in isolation, and limited efforts have been made to systematically compare them through the lens of information integrity. This review focuses specifically on Retrieval-Augmented Generation (RAG) and fine-tuning, identifying them as the two dominant paradigms for integrating external knowledge: RAG for retrieval-based context injection and fine-tuning for parametric knowledge adaptation. While existing surveys predominantly focus on performance metrics like accuracy or latency, this paper addresses the critical gap of data fidelity—the preservation of truthfulness, integrity, and fairness during augmentation. We systematically synthesize empirical findings from diverse methodologies to determine how each approach mitigates hallucinations and bias. By comparing the trade-offs between retrieval-based context injection and parametric knowledge adaptation, this survey provides unique value to readers by providing a structured taxonomy, a unified evaluation framework, and actionable insights to guide future research and practical deployment of robust, high-fidelity LLMs. Full article

The carbon emission allowance (CEA) market is a complex socio-technical and environmental-management system in which regulatory design, trading activity, liquidity conditions, and price volatility interact dynamically. Accurate systems-level tail-risk forecasting and early warning remain challenging because carbon-market losses are affected by nonlinear dependence, episodic liquidity stress, and time-varying volatility. This study proposes an Interpretable Quantile Temporal Network (IQTN) as a systems-oriented risk-monitoring framework for China’s national CEA market. By integrating a feature-gating mechanism, a causal temporal convolutional encoder, and a non-crossing quantile output layer, IQTN directly models the conditional tail distribution of future carbon-market losses. The framework produces multi-horizon Value-at-Risk (VaR) and Conditional Value-at-Risk (CVaR) forecasts for 1-day, 5-day, and 10-day horizons and converts predicted tail risk into operational early-warning signals. Compared with historical simulation, EWMA, GARCH-type models, machine-learning quantile models, and deep temporal benchmarks, IQTN achieved the lowest 95% VaR pinball loss across all horizons, with values of 0.1765, 0.3958, and 0.5732. VaR backtesting showed empirical exceedance rates of 5.23%, 6.04%, and 6.94%, closest to the nominal 5% level. Interpretability analysis identified rolling volatility, maximum loss, intraday range, trading value, and illiquidity as key risk drivers. The temporal importance results also show that recent observations dominated the risk forecasts, suggesting that the risk state of the CEA market is highly sensitive to short-term market information. This supports the use of a short-horizon temporal network as a systems-oriented tool for carbon-market tail-risk monitoring and early warning. Full article

Manufacturing equipment logs are an important source of information for quality monitoring, but building reliable defect prediction models from such logs is still difficult in practice. Defective samples are rare, and many process variables are missing because measurements are recorded only under certain sensing or process conditions. These properties make defect prediction difficult and limit the usefulness of accuracy-based evaluation. This paper evaluates defect prediction using the Bosch Production Line Performance dataset, with a supplementary validation experiment on the semiconductor manufacturing process (SECOM) dataset. Two feature configurations are compared: a baseline representation using imputed numerical variables and a missingness-aware representation that adds feature-wise missing indicators and a sample-level missing ratio. Logistic Regression, Random Forest, and LightGBM are evaluated using validation-based threshold selection. To examine the effect of imputation choice, zero, median, and KNN imputation are also compared in the SECOM experiment. In the Bosch experiment, explicitly representing missingness improves PR-AUC for all tested model configurations. The supplementary SECOM experiment shows a more mixed pattern, suggesting that the usefulness of missingness-aware features depends on the dataset, imputation strategy, and model family. The latency analysis further shows a practical trade-off: Random Forest with missingness-aware features gives the highest PR-AUC on Bosch but has the highest inference latency, while LightGBM provides a more balanced choice when prediction performance and response time are considered together. Full article

High-throughput technologies such as RNA-seq and single-cell transcriptomics generate increasingly large and high-dimensional gene expression datasets in which nonlinear dependence structures are common. Because classical methods primarily capture linear associations, they may fail to characterize many biologically relevant patterns of dependence. To address this limitation, diverse nonlinear dependence measures—including information-theoretic, rank-based, kernel-based, distance-based, copula-based, and clustering-based approaches—have been developed. However, the field remains fragmented, and comparative evaluations are often inconsistent. This review organizes nonlinear methods into major methodological families and critically compares their statistical behavior, strengths, limitations, and characteristic modes of failure. We emphasize that method selection depends on matching inferential objectives to estimator assumptions, analytical constraints, and characteristic failure modes. By identifying recurring trade-offs among flexibility, robustness, interpretability, and computational scalability, we provide scenario-based guidance for method selection in transcriptomics, network inference, and functional genomics. In doing so, we aim to align inferential objectives with analytical requirements, supporting principled and application-specific use of nonlinear dependence methods in modern omics research. Full article

Modern route planners such as Google Maps and Apple Maps serve millions of users worldwide, optimizing routes in large-scale road networks where fast responses are required for diverse cost metrics including travel time, fuel consumption, and toll costs. Classical algorithms like Dijkstra or A${}^{*}$ are too slow at this scale, and while index-based techniques achieve fast queries, they are often tied to fixed metrics, making them unsuitable for dynamic conditions or user-specific metrics. Customizable approaches address this limitation by separating metric-independent preprocessing and metric-dependent customization, but they remain limited by slower query performance. We recently introduced Customizable Tree Labeling (CTL) as a framework that combines tree labelings with shortcut graphs. The shortcut graph enables efficient customization to different cost metrics, while tree labeling, supported by path arrays, provides fast query answering. Although CTL enables optimizing routes with different cost metrics, it still faces challenges in storing and reconstructing path information efficiently, which hinders its scalability for answering millions of queries. In this article, we build on the CTL framework by developing several algorithmic variants that differ in the information retained within shortcut graphs and path arrays, offering a spectrum of trade-offs between memory usage and query performance. To further enhance scalability, we propose a batch processing strategy that shares path information across queries to eliminate redundant computation. We empirically evaluated the performance of our algorithms on 13 real-world road networks. The results show that they significantly outperform state-of-the-art methods, achieving speedups of up to factor 15 for route computation while maintaining practical memory requirements. Full article

Cross-border e-commerce (CBEC) has emerged as a critical mode of international trade; however, product uncertainty and transaction risk remain persistent barriers to purchase decisions. While digital platforms have developed various solutions, the role of offline experiential knowledge in shaping online purchase behavior remains underexplored. This study examines how tourism-derived information influences purchase intention in CBEC. Drawing on transaction cost theory and uncertainty reduction theory, we propose that tourism-derived information enhances product familiarity and perceived diagnosticity, which subsequently reduce product uncertainty and increase cross-border purchase intention, and further examine the moderating role of transaction uncertainty. A four-week survey in March 2026 collected data from 325 Chinese consumers who had visited Korea and encountered Korean cosmetics and beauty products; data were analyzed using PLS-SEM. Results show that tourism-derived information significantly enhances product familiarity and perceived diagnosticity while directly reducing product uncertainty; reduced product uncertainty, in turn, positively influences purchase intention. Transaction uncertainty strengthens the negative effect of product uncertainty on purchase intention. By reconceptualizing tourism experience as an experience-based informational resource in CBEC and providing a multidimensional perspective on consumer uncertainty, this study contributes to consumer behavior research in digital commerce and offers practical insights for CBEC platform operators and cross-border retailers. Full article

This paper studies the informational hierarchy of individual stocks in the Colombian Stock Exchange (BVC), with particular attention to the role of cross-listed securities. The paper addresses a gap in the literature on small emerging markets, where evidence on intra-market information and return transmission remains scarce, particularly in the presence of illiquidity, cross-listing, and external risk exposure. Using daily data for 2015–2024, we estimate a five-asset vector autoregression VAR (3) with exogenous global controls and compute generalized forecast error variance decompositions within the Diebold–Yilmaz connectedness framework, with residual-bootstrap inference and CBOE Volatility Index (VIX)-based regime analysis. The VIX regimes are used to distinguish low-, medium-, and high-global-risk environments because global risk appetite is a key channel through which external shocks affect emerging equity markets. Three results stand out. First, total connectedness is moderate in the full sample, at 25.2%, but rises sharply with global risk, from 17.5% in low-VIX periods to 28.4% in high-VIX periods. Second, Ecopetrol’s American Depositary Receipt listed on the New York Stock Exchange (EC, NYSE) emerges as the dominant net transmitter of return innovations, and its informational leadership becomes stronger as global uncertainty increases. Third, when the local Ecopetrol share is excluded, leadership shifts to Bancolombia’s ADR (CIB), suggesting that directional spillover leadership is associated not only with firm identity but also with the offshore trading venue. These findings document a regime-dependent and venue-driven informational hierarchy, consistent with ADR-listed securities acting as dominant transmitters of return innovations to the domestic Colombian equity system. For portfolio managers, the results imply that diversification across local Colombian equities may overstate the number of independent information sources, especially during high-risk periods, and that monitoring ADRs, global volatility, oil prices, and exchange-rate conditions may improve hedging and risk management. Full article

Introduction: The global ornamental fish trade is a rapidly expanding sector and a major pathway for the introduction of non-native species, particularly in freshwater ecosystems in developed countries. The introduction of non-native species can result in a range of ecological impacts, including predation, competition, hybridization, and disease transmission, often leading to ecosystem degradation and biotic homogenization. Therefore, it represents a clear ecological risk, especially serious in freshwater systems with a high endemism rate, such as the Iberian Peninsula. The occurrence of ornamental non-native species in the Iberian Peninsula has been common, yet little has been done to describe the overall ornamental fish trade as a first step to evaluate invasion risk. Objective: This study characterizes the import dynamics of ornamental freshwater fish in Portugal between 2020 and 2024 and evaluates its potential role as a pathway for species introductions. Methodology: Data were obtained from the Institute for Nature Conservation and Forests database, including information on species composition, quantities, sizes, prices, and countries of origin. A total of 431 records were analyzed, resulting in 27,689 validated entries of imported freshwater fish, which were taxonomically verified and filtered to retain only freshwater species. Results: A total of 666 species from 88 families were identified, with an average of 380 species imported annually, reflecting high taxonomic diversity. Import volumes increased from approximately 1.25 million individuals in 2020 to 1.75 million in 2024, while total import value nearly doubled from €300,000 to €600,000. Imports were predominantly from five Southeast Asian countries, particularly Indonesia and Vietnam, and largely supported by aquaculture production (88%). A stable core of highly traded species, including Carassius auratus, Poecilia reticulata, and Paracheirodon innesi, suggests a sustained and very high propagule pressure, while some species variability was observed on yearly basis, suggesting the importance of monitoring programs on actual imports. Conclusions: Overall, the ornamental fish trade represents a significant and growing pathway for biological invasions in Portugal. The combination of increasing trade volume, high species diversity, and persistent dominance of key taxa highlights the need for improved monitoring, regulatory frameworks, and public awareness to mitigate ecological risks. Full article

Modern connected vehicles rely on the controller area network (CAN) to disseminate safety-critical in-vehicle information, including sensor-related and vehicle-state signals such as engine revolutions per minute (RPM) and gear state, among electronic control units (ECUs). Because CANs lack built-in authentication and encryption, malicious message injection and spoofing can compromise the integrity and availability of vehicular sensing and control functions. Existing deep-learning-based intrusion-detection systems (IDSs) show a clear trade-off: supervised methods perform well on known attacks but rely on costly labels, whereas unsupervised methods can identify unseen attacks but often suffer from high false-positive rates. To address these limitations, this paper proposes a semi-supervised generative adversarial network (SGAN) framework for CAN bus intrusion detection that combines image-based CAN representation with adversarial learning. Consecutive CAN messages are converted into $64\times 9$ grayscale images, and the proposed framework is trained in three phases. First, the discriminator establishes an initial decision boundary using a small labeled subset. It then refines this boundary through distribution-level likelihood objectives and generated samples. Finally, the generator is trained to produce realistic samples capable of deceiving the discriminator. The proposed method was evaluated on the Hacking and Countermeasure Research Lab (HCRL) car-hacking dataset using leave-one-class-out experiments to simulate unknown attacks and achieved an average accuracy of 99.73% and an average F1-score of 99.63% on unknown attacks. Moreover, with only 0.21 M parameters and 3.25 M floating-point operations (FLOPs), the model is well suited for resource-constrained in-vehicle platforms. These results indicate that the proposed framework can serve as a practical cybersecurity component for protecting CAN-carried data in vehicular sensing applications. Full article