Search Results (382)

Road markings play a vital role in traffic safety and flow, yet their effectiveness relies entirely on drivers’ comprehension. This study seeks to assess the comprehension levels of a sample of drivers from Al-Najaf city (Iraq) and examine the extent to which their personal characteristics could influence these levels. While conventional linear statistical methods have provided foundational measures of driver comprehension, this study extends current research by utilizing a multilayer perceptron (MLP) artificial neural network (ANN) framework to capture complex, non-linear relationships between driver attributes and road marking comprehension in Al-Najaf, Iraq. Direct interviews were conducted with 402 drivers using a structured questionnaire to collect data on their personal attributes, driving behavior, and knowledge of 14 road markings. A set of correlational and group-comparison statistical analyses was initially performed before conducting the backpropagation-based ANN analysis; a supplemental sensitivity analysis for the best combination of activation functions and training/testing split ratios was performed. The analyses revealed an overall comprehension level of 72%. Crucially, the optimized ANN model revealed non-linear predictor importance hierarchies, demonstrating that drivers’ marking recognition and educational attainment are the primary determinants of conceptual comprehension, outweighing raw driving experience. These findings indicate that years of driving do not ensure adequate knowledge of road markings, revealing important limitations in current licensing standards. Consequently, this research offers an empirical foundation for transport authorities to transition from static licensing exams to continuous, adaptive driver education schemes targeting high-risk demographic groups. Full article

Emerging industrial agglomeration drives green transformation through knowledge spillovers and economies of scale. However, its effects exhibit pronounced nonlinearity and heterogeneity, shaped by spatial externalities and development stages. This paper investigates 19 Chinese urban agglomerations over the period 2014 to 2023. Kernel density estimation based on enterprise-level Point-of-Interest (POI) data is used to characterize spatial agglomeration patterns across eight emerging sectors. A two-stage dynamic network super-efficiency SBM model decomposes Green Transition Efficiency (GTE) into resource utilization and pollution control sub-stages. An XGBoost-SHAP-GEO analytical framework, combined with partial dependence analysis, then identifies nonlinear driving mechanisms. The main findings are as follows: First, emerging industrial agglomeration intensifies and polarizes toward the eastern coast, whereas GTE displays a “high-west, low-east” pattern. This produces a significant spatial mismatch, rooted in the near-saturation of environmental carrying capacity in eastern regions, where congestion effects exceed knowledge spillover dividends. Second, geographic characteristics constitute the primary factor shaping GTE and operate through nonlinear interactions with industrial agglomeration and R&D investment. Notably, their moderation direction is reversible, suggesting that geographic endowments should be understood as “conditional assets” rather than fixed advantages. Third, nonlinear patterns across sectors are highly heterogeneous. The bio-industry is the only sector to achieve a J-shaped positive breakthrough. Information technology and new materials exhibit persistent inhibition, while related services display an extremely narrow threshold window with the deepest negative reversal. Thus, “moderate agglomeration” is a multidimensional concept that shifts dynamically with industry type and regional endowment. Fourth, driving mechanisms display stage-dependent evolution. The incubation stage relies on natural endowments and basic industrial pull, with the green bottleneck residing in resource utilization efficiency. The growth stage faces multiple tensions from coexisting positive and negative effects. The optimization stage shifts toward R&D innovation and industrial greening, marking a qualitative transformation from MAR externalities to Jacobs externalities. In addition, the non-significant linear coefficient in the 2SLS instrumental variable test is consistent with the inverted U-shaped nonlinear finding, further validating the necessity of a nonlinear analytical framework. These findings provide differentiated governance evidence for balancing industrial agglomeration with green sustainable development. Full article

In recent years, aquaculture, a critical pillar of global food security, has faced challenges including germplasm resource degradation, frequent disease outbreaks, and insufficient environmental adaptability. To address these issues, omics technologies—centered on genomics, transcriptomics, and related fields—are driving the transformation of aquaculture toward precision and intelligence by establishing a three-dimensional “genotype–phenotype–environment” knowledge network. In this review, we outline foundational applications of omics in aquaculture and highlight the characteristics and current applications of distinct omics approaches. Furthermore, we summarize three core application domains of omics in aquaculture: (1) genetic breeding via genomic selection (GS), gene editing, and multi-omics molecular dissection of economic traits; (2) disease prevention through host–pathogen interaction studies and microbial engineering; and (3) environmental assessment through multi-omics characterization of biological and ecological responses. Additionally, we identify challenges in applying omics technologies to aquatic breeding and database development. Overall, this review aims to provide aquaculture practitioners with actionable insights by enhancing understanding of omics-driven innovations in sustainable aquaculture advancement. Full article

While artificial intelligence (AI) is expected to drive paradigm-shifting transformations, many initiatives result in merely incremental optimization. Anchored in strategic human capital theory, this study shifts the analytical focus from the scale of elite technical talent, conceptualized as AI stars, to the configuration of their knowledge structures to unpack this paradox. Using a dataset of 1270 medical AI patents from corporate R&D teams, we employed high-dimensional fixed-effects models to examine these dynamics. The results reveal that while the knowledge diversity of AI stars acts as a potent engine for breakthrough innovation, this generative capacity is attenuated by excessive domain familiarity. Specifically, direct domain familiarity (derived from internal experience) and indirect domain familiarity (absorbed through external collaborative networks) negatively moderate this relationship, a dynamic theorized to operate through internal cognitive entrenchment and external relational conformity, respectively. Extending the efficiency-driven consensus regarding bilingual expertise, these findings demonstrate that excessive domain embeddedness transforms from an informational bridge into a restrictive constraint during paradigm-shifting innovations, particularly within highly institutionalized environments. Full article

Semantic segmentation is a critical task in computer vision aiming to identify and classify individual pixels in an image, with numerous applications, for example, in autonomous driving and medical image analysis. However, semantic segmentation can be highly challenging, particularly due to the need for large amounts of annotated data. Annotating images is a time-consuming and costly process, often requiring expert knowledge and significant effort; moreover, saving the annotated images could dramatically increase the storage space. In this paper, we propose a novel approach for semantic segmentation, requiring only rough information about the proportions of individual semantic classes, hereafter referred to as semantic proportions (SPs), rather than the necessity of ground-truth segmentation maps. This greatly simplifies the data annotation process and thus will significantly reduce the annotation time, cost and storage space, opening up new possibilities for semantic segmentation tasks where obtaining the full ground-truth segmentation maps may not be feasible or practical. Our proposed method of utilising semantic proportions can (i) further be utilised as a booster in the presence of ground-truth segmentation maps to gain performance without extra data and model complexity, and (ii) also be seen as a parameter-free plug-and-play module, which can be attached to existing deep neural networks designed for semantic segmentation. Extensive experimental results demonstrate the good performance of our method compared to benchmark methods that rely on ground-truth segmentation maps. Utilising semantic proportions suggested in this work offers a promising direction for future semantic segmentation research. Full article

The development and maintenance of the nervous system depend on a tightly regulated intracellular transport network in which kinesin superfamily (KIF) motor proteins drive microtubule-based delivery of synaptic vesicle precursors, organelles, mRNAs, and signaling components along axons and dendrites. Disruption of this machinery underlies a clinically heterogeneous spectrum of neurodevelopmental disorders (NDDs), including intellectual disability, epilepsy, autism spectrum disorder, microcephaly, malformations of cortical development, spasticity, and axonal neuropathy. Here, we synthesize current knowledge on how kinesin dysfunction shapes neurodevelopment. We outline the physiological roles of kinesins in neuronal polarity, organelle and mitochondrial positioning, synaptogenesis, and progenitor division, and survey principal disease-associated genes, including KIF1A, KIF5A, KIF7, KIF11, KIF2A, KIF5C, and emerging members such as KIF14, KIF15, and KIF16B. We detail how distinct pathogenic mechanisms, such as loss of motility, impaired cargo coupling, motor hyperactivity, mitotic spindle defects, and disrupted ciliary signaling, converge on shared cellular endpoints, and how tubulin isotypes and posttranslational modifications further modulate motor output. In this review, we discuss translational implications, including variant-resolved diagnosis and precision strategies to restore transport, dampen pathological hyperactivity, or stabilize the microtubule track. Collectively, these advances reframe kinesinopathies as mechanistically stratified disorders of neuronal transport. Full article

Environmental NGOs (ENGOs) play an important role in promoting climate governance as Southeast Asia is highly vulnerable to climate change and has limited governance capacity. Despite the existing research, however, the factors that affect ENGO participation have been identified, but not the hierarchies between these factors, any structural dependency, or strategic implications, which could help to understand the conditions that enable effective participation. To fill this gap, this study proposes a framework of 16 important factors with a structured literature review and expert consultation and builds a matrix of factor relations on the basis of assessments made by 11 experts of various regional experiences. These different factors are used to reveal the hierarchical structure of these factors and their driving–dependence relationships by means of Interpretative Structural Modeling (ISM) and the Matrice d’Impacts Croisés-Multiplication Appliquée à un Classement (MICMAC) methods. The results show that the foundation of the structural conditions, such as institutional stability, resource predictability, organizational governance, and capacity allocation, is the most important factor in ENGO participation, and the strategic actions of social trust, network embedding, advocacy and accountability only work when these foundation conditions are met. Based on these results, four successive strategic pathways are suggested to enhance ENGO involvement in climate governance: institutional embedding, capability assetization, network embedding, and accountability and initiative upgrading. Theoretically, the study contributes to the field of climate governance research as a demonstration of a hierarchical process of structural preconditions for ENGO participation, not just strategic actions. In the practical sense, the evidence generated contributes to the formulation of policy and design of governance and participation strategies at ASEAN and at the national level by NGOs and other environmental groups. The results are based on expert knowledge used to build the ISM–MICMAC matrix and represent only a structural reflection of the relationships between the factors and not a causal relationship. Full article

SASH1 (SAM [sterile alpha motif] and SH3 [SRC-homology-3] domain-containing protein 1) is a multidomain scaffold implicated in pigmentation, innate immunity, receptor signaling, cytoskeletal dynamics, vascular biology, and tumor suppression. Although genetic and expression studies link SASH1 dysfunction to diverse diseases, a unifying mechanistic framework has remained elusive. Here, we synthesize current knowledge of SASH1 structure, interaction networks, and biological functions across cell types and disease contexts. SASH1 contains an intrinsically disordered SPIDER (SLy Proteins Associated Disordered Region), an SH3 domain, two SAM domains, and multiple linear motifs; together, these elements mediate interactions with EphA8 (ephrin type-A receptor 8), β-arrestin 1, TRAF6 (TNF receptor-associated factor 6), CRKL (CRK-like proto-oncogene), IQGAP1 (IQ-motif-containing GTPase-activating protein 1), cortactin, and TNKS2 (tankyrase-2). We propose that SASH1 functions as a context-dependent multi-docking scaffold that organizes signaling architecture. Its modular domains, intrinsically disordered regions, and dual SAM domains enable flexible, multivalent interactions with partners that can be grouped into three functional modules: receptor regulation, intracellular signaling, and cytoskeletal organization. Notably, many SASH1 partners are themselves scaffold or adaptor proteins, allowing integration into pre-existing networks in a hierarchical ‘scaffold-of-scaffolds’ manner. Through selective partner recruitment, SASH1 links cell-surface receptor inputs to downstream signaling pathways and cytoskeletal remodeling. This model provides a mechanistic framework for how SASH1 drives diverse, cell-type-specific outputs across physiology and disease, while revealing broader principles by which multidomain scaffolds encode cellular behavior. Full article

In this paper, we present LightBAL, an ultra-lightweight deep learning framework for real-time active cell balancing and onboard balancing control in electric vehicle (EV) battery management systems (BMSs). Although active cell balancing can improve battery utilisation and performance, applying deep learning-based balancing control strategies remains prohibitive in typical embeddable BMS platforms because of the computational complexity and inference latency of deep models. In response to this issue, we propose an AI-physics-informed controller that forecasts the voltage difference of a single cell, the SoC variation, and the optimal balancing current based on proportional feedback closed-loop (FCLL) control. The introduced framework exploits wavelet-based adaptive denoising, multi-scale hierarchical feature learning using a cooperative Principal Component Analysis (PCA) and autoencoder feature extraction technique, and a lightweight One-Dimensional Convolutional Neural Network (Conv1D) coupled with Bidirectional Long Short-Term Memory (BiLSTM) (Conv1D-BiLSTM). The implemented lightweight network is further trained by model compression methodologies such as knowledge distillation and 8-bit quantisation-aware training, aiming for efficient deployment on edge devices. Experimental validation on the multivariate battery time-series dataset demonstrates that LightBAL achieves an F1-score of 96.64%, a balancing efficiency of 94.30%, and a Mean Absolute Error (MAE) of 0.0379, outperforming methods based on conventional ANN, LSTM, and CNN. LightBAL without compression takes only 1.26 s to conclude on a PC workstation; the inference latency of the embedded light model is as low as 28.7 ms. In addition, hardware-in-the-loop (HIL) validation on the Raspberry Pi 4 platform indicates that the framework can fulfil real-time inference requirements under normal operating conditions, taking 28.7 ms per balancing process. Simulation shows that the proposed approach significantly decreases cumulative balancing energy loss by 12.4% across several driving cycle conditions. Full article

This paper investigates the quasi-synchronization (QS) problem for drive-response discrete-time delayed inertial neural networks (DTDINNS). Unlike existing studies that mainly rely on classical stability theorems, linear matrix inequality (LMI) methods, and matrix measure approaches (MMA), this work establishes three innovative quasi-synchronization criteria for DTDINNS by adopting the solution formula of second-order linear difference equations (SOLDES), infinite series summation techniques, and the solution formula of first-order linear difference equation group (Lemma 6). To the best of our knowledge, this study is the first attempt to introduce the general solution theory of second-order linear difference equations and infinite series summation methods to analyze the synchronization behavior of neural networks (NNS). The proposed framework offers a novel theoretical tool for the synchronization analysis of discrete-time delayed neural networks (DTDNNS), which bears important theoretical significance for relevant research fields. Full article

To address the two fundamental contradictions in distribution network digital simulation systems: “diversified simulation requirements versus standardized configuration” and “massive simulation outputs versus sparse decision-making information”, this paper builds upon existing work on Common Information Model (CIM)-based automatic simulation model generation and dynamic voltage regulation simulation and further constructs a user-oriented intelligent simulation service layer. This layer is collaboratively composed of an Orchestration_Agent (simulation orchestration agent) and an Analysis_Agent (result analysis agent), tasked with three responsibilities: based on multi-level simulation granularity (L0–L3) and a simulation template library, leveraging a large language model (LLM) to achieve natural language requirements parsing and automatic workflow orchestration; based on a simulation knowledge graph, implementing parameter recommendation, verification, and anomaly-adaptive recovery for process control; and based on a hybrid architecture of rule templates, statistical analysis, and causal graph models, achieving automatic result analysis, root cause reasoning, and structured report generation, with case feedback driving knowledge base iteration. Validation was conducted on data from a real 10 kV feeder with 91 distribution transformer areas over 30 consecutive days (2880 time cross-sections): comprehensive requirements-parsing accuracy of 96.3%, automatic parameter configuration coverage rate of 94.7%, anomaly identification recall/precision of 94.0%/96.9%, root cause reasoning accuracy of 92.1%, and the median end-to-end time per simulation shortened from approximately 36 min under the manual mode to 4.7 min. The results demonstrate that the proposed service layer provides a viable engineering technology pathway for the evolution of distribution network simulation from tool-oriented to service-oriented. Full article

Chronic lymphocytic leukemia (CLL) is characterised by the dependence of malignant cells on specialised tissue microenvironments within the bone marrow (BM) and secondary lymphoid organs (SLOs), which provide essential survival and proliferative signals. The CXCR4/CXCL12 and CXCR5/CXCL13 chemokine axes direct the trafficking of CLL cells into these anatomically distinct compartments, where stromal-derived survival signals protect them from both spontaneous and therapy-induced apoptosis. Although each chemokine axis has been extensively studied individually, no previous review has integrated both pathways into a unified mechanistic framework. This review proposes that the signal transducer and activator of transcription 3 (STAT3) function as a shared molecular hub that integrates niche-derived cytokine signals, including interleukin-6 (IL-6), IL-10, and IL-21, and may transcriptionally upregulate both CXCR4 and CXCR5, and reinforce tissue homing through a positive feedback loop. This review seeks to evaluate the expression, signalling, and clinical significance of each axis, their points of convergence and divergence and the therapeutic strategies that disrupt these parallel homing pathways. Complementing this framework, recent clinical evidence indicates that circulating CXCL13 serves as a robust prognostic biomarker in CLL, and that STAT3 inhibition may overcome bone marrow stromal-mediated cytoprotection. The CXCL12-CXCR4-STAT3-IL-10 immunosuppressive axis further drives T-cell exhaustion. Together, these pathways form an integrated oncogenic network that supports CLL cell survival, drives immune dysfunction, and promotes therapy resistance. Several important knowledge gaps remain. These include the lack of direct validation of the STAT3-CXCR5 transcriptional axis in primary CLL cells and uncertainty regarding whether CXCR4/CXCR5 dominance represents a stable transcriptional programme or a dynamic, microenvironment-driven process. Addressing these questions through single-cell transcriptomics, spatial transcriptomics, proteomics, and functional validation studies will be essential for developing rational combination therapies capable of simultaneously disrupting both homing axes. Full article

To address the stringent demands of intelligent connected vehicles for high bandwidth, low latency, and highly reliable communication, this paper systematically summarizes the semantic communication technology of the Internet of Vehicles (IoV) based on information “meaning” transmission, covering basic theory, key technologies, application practice and challenge and trends. First, the paper expounds the knowledge driven and task oriented paradigm characteristics of semantic communication and its efficiency advantages in the IoV. Second, in terms of key technologies, semantic extraction achieves efficient feature compression through multimodal fusion and Generative Artificial Intelligence (GAI); semantic coding employs hierarchical codebooks and adaptive strategies to optimize transmission efficiency; semantic transmission leverages deep reinforcement learning for the joint scheduling of resources such as spectrum and power; and semantic decoding utilizes reconstruction networks and GAI to enhance resilience against impairments. Application practices demonstrate that semantic communication can significantly compress image data transmission volume for autonomous driving collaborative perception while maintaining high-fidelity reconstruction under adverse channel conditions. It significantly reduces the communication load and improves the system utility in vehicle-to-infrastructure coordination and in-vehicle service. Despite facing technical challenges such as semantic consistency, dynamic adaptability, and security trustworthiness, future semantic communication will evolve towards deep integration with distributed collaborative knowledge networks, lightweight real-time decision-making agents, and integrated “communication, sensing, and computing” architectures, positioning itself as a key enabling technology for empowering Sixth Generation mobile communication (6G) of intelligent vehicular networks. Full article

Breakthrough innovation—central to industrial competitiveness and the ongoing clean-energy transition—remains persistently constrained by information homogenization and weak cross-domain integration in single-layer innovation networks. Technology Innovation Composite Networks (TICNs) have therefore been advocated as dual-layer platforms coupling knowledge and collaboration networks, yet the cross-layer mechanism through which they generate breakthrough outputs has not been specified. This paper specifies and tests how knowledge-layer structural holes open access to heterogeneous information that must cross into the collaboration layer to be recombined into breakthroughs. Two distinct boundaries shape the outcome. Inventors’ finite cognitive processing capacity makes integration returns decay along an inverted U-shape; separately, excessive degree and closeness centrality drive the collaboration layer into homogenization and localization, narrowing the range of structural holes it can productively absorb and shifting the breakthrough peak toward lower structural-hole levels. Together, they delineate an optimal cross-layer integration zone. Using panel data on 10,681 patents, 948 inventors, and 5631 inventor-year observations from new energy (2004–2018), a fixed-effects negative binomial model confirms the inverted U-shape and the steepening, peak-shifting moderations of degree and closeness centrality; a Lind–Mehlum test places the turning point inside the observed data range, and negative binomial (robust SE), Poisson and zero-inflated Poisson specifications—together with a stricter top-1% breakthrough threshold—yield consistent results. The study moves multilayer network research from structural description toward mechanism-level identification and offers actionable network-design guidance. Full article

Solar radiation drives virtually every process in Earth’s climate system—from atmospheric circulation and the hydrological cycle to ecosystem carbon uptake and agricultural productivity. How this energy flux is changing under anthropogenic climate forcing and what the consequences might be have become central preoccupations of modern Earth system science. Yet despite a rapidly growing literature spanning atmospheric physics, ecology, remote sensing, and energy engineering, no study has attempted to map the global scientific output on solar radiation and climate change as a unified research domain. This study addresses this gap through a large-scale bibliometric analysis of 8473 publications retrieved from the Web of Science Core Collection (1991–2025). Using the Bibliometrix R package (v5.0.1) and VOSviewer (v1.6.20), the study examined publication growth, country and institutional productivity, journal performance, co-authorship structures, keyword networks, thematic evolution, and emerging research fronts. The literature has grown at an annual rate of 14.87%, with China and the USA accounting for nearly half of all output—though American research shows markedly higher citation impact. Bradford’s Law identified 27 core journals, which accounted for roughly one-third of total publications; the Journal of Geophysical Research–Atmospheres ranked first. Consistent with Lotka’s Law, a large majority of authors (78.9%) appear only once in the dataset, pointing to a broad but peripherally engaged scientific community. Keyword co-occurrence mapping revealed five thematic clusters: ecological and biosphere impacts; climate dynamics and variability; atmospheric processes and data-driven methods; solar geoengineering; and energy and renewable applications. The most rapidly rising topics after 2020—machine learning, CMIP6, solar geoengineering, and heatwaves—suggest that the field is shifting toward data-driven methods and active climate intervention debates. These findings offer a structured overview of where the field stands and the most urgent knowledge gaps. Full article