In-depth exploration of spermidines in Goji berry: the identification and semi-quantitation of spermidines by liquid chromatography-tandem mass spectrometry with feature-based molecular network and high performance liquid chromatography.

CNNs vs. transformers: A benchmark for multi-class marine debris identification.

Using neural networks to understand static and dynamic cues in facial expression recognition.

Brokers serve as key connectors linking academic researchers who might otherwise remain unconnected, to co-authorship networks. This study examines whether more complex interdisciplinary co-authorships yield greater scholarly impact than ties with authors from the same discipline, which are facilitated by cognitive similarity. It also tests the extent to which such collaborations influence the academic performance of female researchers. The study analyzed 594 authors and 271 papers in the field of social learning. Despite their complexity, regression analyses confirm the benefits of novel sources and combinations of interdisciplinary knowledge. The higher coordination and communication costs of interdisciplinary collaboration are offset by the potential of the new knowledge generated. Interdisciplinary collaboration is associated with higher performance among female scholars, contributing to increased recognition and reputation. This pattern suggests that collaboration strategies and institutional support should reduce coordination barriers in interdisciplinary co-authorship and facilitate women's access to high-impact brokerage opportunities.

In asphalt mixture research, deep learning networks have been widely applied to analyze slice images, serving as a foundation for numerous further valuable research. However, the effectiveness of these networks is hindered by the high cost of acquiring large-scale datasets, which is a great challenge for slice image dataset construction. To address this challenge, we propose Slice Image Translation Generative Adversarial Network (SIT-GAN), a novel slice image translation network designed to generate realistic images across various mixture types. Leveraging an innovative hierarchical feature extraction module (HFEM), SIT-GAN effectively captures the complicated multiscale features. Both qualitative and quantitative evaluations, as well as validation experiments, demonstrate the superior generation quality of SIT-GAN. Moreover, we show that using SIT-GAN for data augmentation significantly enhances the performance of classification and segmentation networks, providing a robust foundation for intelligent analysis in pavement engineering.

Waterborne vs solvent-borne 1 K/2 K polyurethane coatings: Network formation and performance of Bayhydrol®-based systems

Multi-GPU quantum circuit simulation and the impact of network performance

MoFormer: A centrality-aware multi-task graph transformer with multi-gate mixture-of-experts for link-level network performance modeling

Recent advances in surgical robotic systems, high-speed communication networks, and information processing technologies have made the clinical implementation of remote surgery increasingly feasible. Although pilot clinical applications have been initiated worldwide, the safe, ethical, and sustainable adoption of remote surgery requires comprehensive guidance that addresses not only technical considerations, but also clinical practice, legal responsibility, and organizational frameworks. In response to these needs, the Japan Surgical Society has developed the second edition of the Clinical Practice Guidelines for Telesurgery through a multidisciplinary, consensus-based process involving multiple surgical societies. This updated edition builds on validation and verification studies conducted since the publication of the first edition and places particular emphasis on practical implementation in real-world clinical settings, including telesurgical support and telementoring. The guidelines provide expanded, implementation-oriented recommendations covering surgeon and support staff qualifications, institutional requirements, communication network performance and cybersecurity standards, registry-based governance, and structured approaches to remote surgical mentoring. In addition, legal and ethical considerations are strengthened through the inclusion of representative informed consent documents and contractual frameworks. To enhance international applicability, content that is broadly relevant across jurisdictions is presented separately from elements specific to the Japanese regulatory environment. These guidelines aim to support the responsible global dissemination of telesurgery by promoting safety, transparency, and clinical effectiveness.

Efficient cell association remains a fundamental challenge in fifth-generation (5G) vehicle-to-everything (V2X) systems due to rapid topology changes, heterogeneous deployments, and stringent latency requirements. Conventional learning-based approaches often rely on shallow representations or independent optimization strategies, limiting their adaptability in dense and highly dynamic environments. To address these issues, this study introduces a multi-level graph representation framework that models interactions between vehicles and base stations across hierarchical spatial structures. The proposed approach integrates contextual node embedding with attention-driven graph learning to capture mobility patterns, signal characteristics, and network load dependencies. Additionally, a training-stage optimization mechanism is incorporated to refine attention parameters, improving convergence behavior without increasing inference complexity. The framework is evaluated using a real-world vehicular mobility dataset, demonstrating consistent improvements in association stability, handover reliability, and overall network performance compared with existing deep learning and graph-based methods. Experimental results show gains in accuracy (94.17%) and F1-score (93.93%), indicating enhanced decision robustness under dynamic conditions. Although validation is conducted on an urban dataset, the proposed architecture provides a scalable foundation for adaptive cell selection in next-generation intelligent transportation systems.

Due to global change, natural disasters such as floods have become more frequent in recent years. An effective impact assessment of highway networks before and during floods can help transportation departments prioritize resources and take necessary emergency measures. Although current works have assessed the flood impacts from different perspectives, none have comprehensively evaluated the integrated impacts that capture network structural, functional, and social features, limiting their reliability for decision-making and resilience planning in engineering management. To address this gap, we proposed a graph neural network (GNN)-based framework that incorporates two synthesized indicators—the disaster impact index and criticality score—to integrate structural, functional, and social features. These multidimensional features were inputs to the GNN model, enabling it to capture complex interdependencies and more accurately predict traffic flow and speed under disasters. The practicality of this framework was demonstrated in the case study of Harris County affected by floods caused by Hurricane Harvey. The results showed that Beltway 8, IH-10, and IH-45 were most vulnerable to potential impacts before the flood, while Beltway 8, US-59, and IH-10 were most impacted during the flood, highlighting the need for proactive preflood preparedness and prioritized postflood recovery for these critical roadways. The proposed framework captures complex interdependencies among multidimensional features and more accurately predicts traffic flow and speed. Consequently, it provides a more realistic prediction of the uncertainties in transportation network performance under disasters, offering a robust and practical tool for resilience planning and resource prioritization of other critical infrastructure systems in engineering management.

Atrial fibrillation (AF), the most prevalent cardiac arrhythmia, affects 2% to 4% of the global adult population and is associated with an increased risk of stroke. Early diagnosis of AF and atrial flutter (AFL) is crucial due to their association with stroke risk and the challenge posed by their often asymptomatic and episodic nature. Traditional electrocardiogram (ECG) interpretation requires substantial expert input and can be challenging, especially with poor-quality ECGs. This study aimed to evaluate the performance of a deep neural network (DNN) model in detecting AF/AFL from a large, heterogeneous set of long-term ambulatory ECG recordings, including clinical data collected over 6 months at a university hospital, and assess its effectiveness in a setting reflecting the diversity and complexity of real-world clinical data. The research combined public datasets totaling 10,248 patients, ECG data from our previous studies (648 patients), and authentic long-term ECG recordings from 4346 patients at Kuopio University Hospital for development of the DNN model. Its clinical accuracy and generalizability were assessed using a separate test dataset consisting of 1039 pseudonymized long-term ECG recordings from 1010 patients, all thoroughly reviewed and annotated by experts. The DNN model demonstrated high effectiveness, achieving 96.4% sensitivity and more than 99.99% specificity for time-level AF and AFL detection. At the recording level, it identified AF and AFL with 100% sensitivity and 98.77% specificity, producing false positives in only 1.2% (11/897) of recordings, of which 81.8% (9/11) had other non-AF/AFL arrhythmias. The model maintained high performance across diverse patient characteristics, including varying ages, comorbidities, coexisting arrhythmias, and poor-quality ECG recordings. The results demonstrate that the proposed DNN model may support automated screening for AF and AFL in long-term ambulatory ECG recordings and may reduce manual review workload in clinical practice.

ABSTRACT General aviation (GA) plays a crucial role in enhancing regional transport accessibility, especially in remote and underserved areas. With the increasing integration of GA airports into the existing air transport network, significant transformations are unfolding in both network topology and airport functionality. By detecting the emerging hub and revealing airport connection patterns, we investigate the regional benefits of integrating GA airports into the existing network. New network characteristics are observed: emerging hubs across regions assume diverse functional roles, with clear differences in intra- and extra-regional connectivity compared to core hubs. Meanwhile, the triangular route structures formed among hubs contribute to a more stable and efficient network, accompanied by increased route concentration and the emergence of new connecting flight patterns. These connectivity pattern changes strengthen regional geographic accessibility and optimize overall network performance. We suggest airlines develop regional markets by leveraging GA airport connection patterns. The government should provide priority strategies to support the development of GA airports.

The rapid advancement of wireless network technology and the Internet of Things (IoT) ecosystem in the industrial sector carries significant security risks, particularly regarding vulnerabilities within the management frame protocol. This study aims to evaluate the resilience of wireless network infrastructure against communication disruptions at the protocol level, specifically deauthentication attacks, at PT Jaring Solusi Persada (JSP) Tulungagung Branch. Applying a Research and Development (R&D) method with a Prototyping approach , empirical and real-time testing was conducted using a low-power ESP8266 Deauther Version 2 device as a penetration medium across 15 different access points (APs). Quality of Service (QoS) parameters, including throughput, latency, and packet loss, were measured using Wireshark software to analyze network performance degradation in detail. The experimental results reveal that deauthentication attacks inflict a highly destructive impact on vulnerable devices. Out of the 15 tested APs, 6 units (40%) consisting of consumer-grade devices (such as TP-Link, ZTE, and Totolink) were declared vulnerable because they did not support or activate the Management Frame Protection (MFP/IEEE 802.11w) feature. On these vulnerable devices, the attack successfully triggered complete client disconnections, a drastic 70% drop in throughput (from 567 kbps to 170 kbps), extreme latency spikes ranging from 100 to 300 ms, and a packet loss of 15%. Conversely, 9 AP units (60%), dominated by Ruijie Enterprise business-class devices, successfully withstood the attack and maintained stable network connections due to superior hardware specifications and better defense configurations. This study recommends upgrading network security standards to WPA3 and strictly enabling the Protected Management Frames (PMF) feature to ensure the company's operational continuity.

Urban pipeline infrastructure plays a vital role in ensuring the operational efficiency and service reliability of modern utility systems, especially in industrial regions. While previous studies have focused primarily on pipe failure prediction, limited research has addressed the forecasting of key pipeline performance indicators such as velocity, pressure, and head loss within the context of infrastructure asset management. This study investigates the performance of advanced machine learning (ML) models, PSO-ANN, Genetic CNN, Quantum SVR, Fuzzy Logic Tree, and Bayesian GPR, in predicting three critical output variables: velocity, head loss, and pressure. A dataset comprising 91 instances with geometric and hydraulic descriptors was employed, and descriptive statistics revealed significant variability in flow-dependent parameters. SHAP-based sensitivity analysis highlighted elevation (0.9287) as the dominant factor for pressure prediction, while flow rate (0.4574) and diameter (0.2273) strongly influenced head loss. For velocity, flow rate (0.1139) emerged as the most influential, though other parameters also contributed, justifying their inclusion in the modeling framework. The models were trained using data from the Gadhra Water Distribution Network (District Metered Area-03) in East Singhbhum, Jamshedpur, India. Model evaluation was conducted using R², RMSE, MAE, and MAPE. Results demonstrated a clear performance hierarchy, with Bayesian GPR and Fuzzy Logic Tree exhibiting superior accuracy and stability (R² ≥ 0.98, RMSE ≤ 0.06, MAPE ≤ 0.13), whereas PSO-ANN and Genetic CNN showed relatively weaker performance. The near-perfect R² observed for Fuzzy Logic Tree reflects the small dataset size and its high capacity, highlighting that generalization may be limited in larger or unseen datasets. The analysis of regressor plots, residual distributions, and normalized accuracy matrices further validated these findings. Overall, the study establishes Bayesian GPR and Fuzzy Logic Tree as robust predictive tools for hydraulic modeling while acknowledging dataset constraints that may affect generalization.

Deep learning (DL) techniques have been applied in lung cancer screening, assessing drug effectiveness, and enhancing prognosis prediction. Within this context, the combination of 18FDG PET/CT images with DL has demonstrated promising results, particularly in predicting programmed death ligand-1 (PD-L1) expression in lung cancer, improving overall prediction accuracy and offering a viable non-invasive complementary imaging biomarker to support clinical decision-making and patient stratification. An effective way to improve the performance of deep neural networks in most tasks is to increase the quantity of labeled data and the quality of labels. However, in medical imaging, high-quality annotations and large datasets are both challenging to obtain due to the need for expert knowledge and tedious procedures including regulatory obstacles. In this context, we propose a semi-supervised and unsupervised deep neural networks (USSLNet) using early fusion multi-modal PET/CT images within the context of predicting PD-L1 expression. By alternately running two tasks, label information is propagated to the unlabeled data, enabling the model to extract semantic information and mitigating the risk of overfitting to limited labeled data. Model performance was evaluated using the area under the receiver operating characteristic curves (AUCs) and 95% confidence intervals (CIs). Compared with current methods, our framework demonstrates improved robustness, reducing the impact of outliers and yielding superior performance in PD-L1 status classification. Moreover, the framework consistently outperformed current approaches when utilizing various types of unlabeled PET/CT images. These findings highlight the effectiveness of our approach in predicting PD-L1 expression through the use of limited in size and partially annotated multi-modal PET/CT datasets.

There are various aspects of life in which access to internet services is highly beneficial. The evolution of internet technology must keep pace with rising standards. The issue faced within the SMKS Pendidikan Karya computer network is sub-optimal bandwidth allocation, which results in data congestion and slows down connections, thereby impacting work efficiency. This study aims to optimise the utilisation of internet bandwidth in the SMKS Pendidikan Karya network to improve the quality of access and service distribution for users. The implementation method involves applying bandwidth management using the Hierarchical Token Bucket (HTB) method on a MikroTik router, which includes needs analysis, the design of a bandwidth distribution scheme, system configuration, and network testing. The analytical technique applied was comparative descriptive analysis, comparing network performance before and after implementation based on Quality of Service (QoS) parameters, including throughput, delay, and packet loss. The results of this study indicate that the implementation of the HTB method is capable of managing bandwidth distribution in a fairer and more structured manner in accordance with user priorities, improving connection stability, and reducing network congestion levels. Thus, the HTB method has proven effective in optimising internet bandwidth management on the SMKS Pendidikan Karya network

The rapid expansion of wireless data traffic is placing increasing strain on the energy consumption of current communication networks, intensifying the tension between performance and sustainability objectives. In interference-intensive multiple access scenarios such as power-domain non-orthogonal multiple access (NOMA), energy-efficient optimization is particularly challenging due to the strong coupling between power control and resource allocation decisions. In order to solve this issue, this paper introduces an AI-Enhanced Energy Optimization Framework (AEEOF), which uses deep spatio-temporal learning and reinforcement learning to provide adaptive and energy-aware network control. The proposed framework incorporates a Spatio-Temporal Graph Convolutional Network (ST-GCN) to learn spatial interference relationships and a Gated Recurrent Unit (GRU) to capture temporal traffic dynamics, embedding the resulting representations into a Multi-Agent Deep Deterministic Policy Gradient (MADDPG) controller to support sequential decision-making. Such a design provides the framework with the ability to dynamically distribute power and timing policies based on changes in network conditions. Extensive simulations in a realistic 5G-oriented environment of high interference levels prove the significant performance improvement. The suggested solution can save up to 15% of energy and make the energy use more efficient by about 40%, which is the number of bits delivered per joule. The overall system throughput goes up by 6.25%, the cell-edge user data rate goes up by up to 60%, the fairness goes up by 20%, and the chance of an outage goes down by 70%. A systematic ablation study with three architectural variants validates the individual contribution of each core component - the ST-GCN spatial module, the GRU temporal module, and the MADDPG reinforcement learning controller. Comparative evaluation against conventional orthogonal and non-AI-assisted baselines further supports the effectiveness of the proposed framework within the studied setting. These findings indicate that intelligent spatio-temporal learning is a promising approach for improving energy efficiency and network performance in interference-intensive wireless environments, as demonstrated within the studied 5G-oriented simulation setting.

Automated weed detection is essential for site-specific herbicide application, that can result into the reduced environmental footprint of conventional agriculture. However, for field deployment of automated weeding devices, occlusion remains a critical challenge that can weaken the precision of weed identification. Here, we compare the performance of Vision Transformers (ViT-B16 & PvTv2) and Convolutional Neural Networks (EfficientNet-B0 & ResNet-50) in accurate weed detection, using controlled synthetic occlusion levels (0%, 25%, and 50%). We found that ViT-B16 has superior occlusion resilience, with image testing accuracy increasing from 80% to 86% under 50% occlusion. In contrast, the testing accuracy of PvTv2, EfficientNet-B0 and ResNet-50 dropped from 45 to 76% under similar conditions. Multivariable regression confirmed architecture type as the dominant testing accuracy driver (p ≤ 0.001), with ViTs outperforming CNNs by an average of 14.56% points. These results suggest that occlusion resilience is not uniform across architectural variants but depends critically on attention-based design. Consequently, for real time deployable automatic weed detection systems, hybrid architectures that balance ViT global context with CNN computational efficiency represent a critical future direction. Such approaches can support precise herbicide application, reduce chemical inputs, and enable more sustainable crop protection through reliable AI-driven automation.

The present study aimed to construct a multidimensional performance profile of men's elite youth national teams from the 6 continental confederations (AFC, CAF, CONCACAF, CONMEBOL, OFC, and UEFA) by examining physical, technical-tactical, spatial, and passing network characteristics at the 2023 FIFA U-17 and U-20 World Cups. Data from 174 match observations and 943 player observations were analyzed using mixed-effects models. At U-17, confederation-related differences were observed across physical, spatial, and passing-related domains. AFC players covered more total distance than CONMEBOL (P = .006, effect size [ES] = 1.55), while CAF players reached higher top speeds than OFC (P = .030, ES = 0.87) and showed greater team length in the middle and defensive thirds (mean difference [MD] = 3.55 to 10.19m, P = .004-.040). UEFA showed stronger passing-related profiles than OFC, including higher pass and line-break completion (MD = 0.17 to 0.28, P ≤.005) and higher passing network density, global efficiency, and average clustering coefficient (MD = 0.08 to 0.16, P = .004-.036). At U-20, physical and most spatial differences were limited, whereas passing network differences persisted, with OFC teams showing lower density than all other confederations (MD = -0.17 to -0.11, P < .001-.018) and lower global efficiency than CONCACAF, UEFA, and CAF teams (MD = -0.09 to -0.08, P = .005-.012). Confederation-related profiles were more prominent at U-17 but became less evident in physical and spatial domains by U-20, whereas passing-related differences persisted and highlighted collective passing organization as an important feature of elite youth football development.