Search Results (38,364)

The rapid advancement of machine learning and artificial intelligence (AI) has created new opportunities to enhance diagnostic accuracy in dermatology, particularly within primary care settings. Computer-aided diagnosis (CAD) systems have demonstrated potential to support General Practitioners (GPs) by enabling earlier and more consistent identification of skin diseases. This review critically examines the literature on explainable artificial intelligence (XAI) for skin disease classification, with a specific focus on the evolution of explainability frameworks and the methodological implications of dataset selection. A comprehensive review of studies published between 2020 and 2025 was conducted across multiple academic databases, encompassing research on skin lesion detection, classification, and monitoring. The analysis reveals that deep learning architectures, particularly those leveraging transfer learning with models such as EfficientNet, ResNet, and Xception, frequently report high classification accuracies—often exceeding 90% when evaluated on single benchmark datasets. However, studies employing multiple datasets consistently demonstrate more stable and generalisable performance, albeit with modest reductions in reported accuracy, highlighting a critical trade-off between performance optimisation and real-world robustness. The review further identifies a clear temporal progression in the adoption of XAI techniques. Early studies relied on a broader range of post hoc explainability while later work increasingly consolidated around Grad-CAM, SHAP, and related attribution techniques, followed by gradual diversification into more specialised frameworks such as TCAVs (Testing with Concept Activation Vectors) and Prototype-based Networks. Despite these advances, the lack of clinically grounded explanations, limited integration of ethical considerations, and reliance on non-clinical imagery continue to constrain clinical applicability which we have explored using a GRADE-style narrative. Notably, evidence suggests that CAD systems can improve GP diagnostic accuracy for conditions such as melanoma and seborrhoeic keratosis; however, sustained clinical adoption remains contingent on transparent, reliable, and context-aware explainability mechanisms. Full article

As hybrid work arrangements become more established in organisations, the need for effective design and implementation strategies has grown significantly. Evidence indicates that employee wellbeing and engagement in hybrid work environments are declining and this presents a critical challenge for human resource management (HRM) professionals. This presents HRM professionals with a critical imperative of improving wellbeing, while maintaining engagement and productivity at work. This aligns closely with the United Nations’ 17 Sustainable Development Goals, particularly those that promote wellbeing and decent work. Through a systematic synthesis of 78 studies, this research investigates the key determinants of employee engagement and wellbeing in hybrid work contexts. The conceptual framework for this study is grounded in existing theoretical perspectives from the Job Demands–Resources model, Saks Frameworks and wellbeing perspective presented by Guest. The analysis identifies five critical factors that influence engagement and wellbeing outcomes in hybrid work, accompanied by evidence-based propositions for practice. These recommendations encompass: establishing well-equipped workspaces with appropriate flexibility in both location and time; developing organisational culture and leadership through enhanced communication and collaboration mechanisms; strategically allocating jobs and tasks whilst fostering effective networks and collaboration tools and implementing targeted training interventions to mitigate technostress and burnout associated with digital workloads. We advocate for future research to develop comprehensive models, frameworks and wellbeing interventions to guide HRM professionals in addressing these challenges at both the local and global levels. Full article

The representativeness of point measurements in the atmospheric boundary layer is a fundamental challenge for interpreting observations and evaluating numerical models. In this study, we quantify the representativeness of near-surface wind measurements using a dense network of 13 meteorological towers from the Army Research Laboratory’s Meteorological Sensor Array. These towers are distributed over an approximately 3 × 3 km domain at the U.S. Department of Agriculture Jornada Experimental Range in southern New Mexico. The analyzed domain consists of relatively flat terrain within a broader region of more complex topography. Representativeness is assessed using pairwise differences between towers and deviations from the array mean. Spatial variability decreases with temporal averaging, with the largest reductions occurring between 1 and 10 min and diminishing improvements beyond 10–30 min. Wind measurements become progressively less similar with increasing separation distance, particularly at separations approaching 1 km. Representativeness errors are larger under unstable conditions due to enhanced turbulence and spatial variability, while stronger winds increase wind speed variability but enhance directional coherence. Deviations from domain-averaged conditions are comparable among towers, indicating that no single location is uniquely representative. These results quantify the extent to which temporal averaging, spatial separation, and atmospheric conditions influence representativeness, providing practical estimates of the associated spatial scales and residual errors. The results are useful for interpreting observations, evaluating models, and designing sampling strategies using fixed and mobile platforms, including Uncrewed Aircraft Systems. Full article

Diagnosing tilapia diseases in complex aquaculture environments is severely hindered by noisy backgrounds and limited high-quality pathological data. To overcome these bottlenecks, this study presents a two-stage diagnostic framework integrating an enhanced Segment Anything Model (SAM) with an Adaptive Neuro-Fuzzy Inference System (ANFIS). In the first stage, SAM is augmented with a Convolutional Block Attention Module (CBAM) feature adapter and a Region Proposal Network (RPN)-based prompt encoder. This design enables the automated and precise extraction of irregular disease lesions by self-generating spatial prompts, thereby isolating water background noise. In the second stage, clinical color features extracted from the lesion masks are classified using ANFIS. To optimize performance on small-scale datasets, ANFIS parameters are trained via Particle Swarm Optimization (PSO) under a numerically stable One-vs-Rest (OvR) binary cross-entropy loss. Validated on the public dataset “Enhancing Disease Detection in Nile Tilapia”, our method delivers an average segmentation Dice coefficient of 86.2% and a classification accuracy of 93.5%. This hybrid approach demonstrates strong potential as a foundational baseline for the automated monitoring of aquaculture diseases. Full article

Obesity is increasingly recognized as a complex systemic disorder rather than a simple consequence of excess energy intake and fat accumulation. This review presents a systems biology framework that examines how obesity-driven disruption of inter-organ communication networks contributes to chronic disease susceptibility, with particular emphasis on colorectal cancer (CRC). Disrupted signaling among the brain, adipose tissue, liver, skeletal muscle, gut, and immune system generates maladaptive feedback loops that promote chronic metabolic inflammation (metaflammation), loss of physiological resilience, and progressive metabolic dysfunction. Within this framework, obesity is redefined as a network disease characterized by neuroendocrine dysregulation, adipose tissue remodeling, immune dysfunction, impaired organ crosstalk, and alterations in the gut microbiome. A central feature of this dysregulation is persistent low-grade inflammation driven by immune-metabolic reprogramming and sustained activation of inflammatory pathways. Obesity-associated metaflammation is further linked to accelerated biological aging through mechanisms involving cellular senescence, mitochondrial dysfunction, oxidative stress, and impaired metabolic resilience. These interconnected processes create a tumor-promoting environment by enhancing oncogenic signaling, disrupting intestinal barrier integrity, altering microbial and metabolic signaling, impairing immune surveillance, and promoting epithelial dysfunction, thereby increasing susceptibility to CRC. The review also examines how behavioral, circadian, environmental, and socioeconomic factors influence metabolic health and cancer risk. Finally, emerging translational opportunities, including biomarker-guided risk stratification, precision prevention, metabolic network restoration, and integrative lifestyle and pharmacological interventions, are discussed. Collectively, this review reframes obesity as a whole-body regulatory disorder and provides an integrated conceptual framework linking metabolism, inflammation, aging, and colorectal carcinogenesis to inform future prevention and therapeutic strategies. Full article

The role of externalities generated by enterprise digital transformation in advancing SDGs 8 and 9 has been largely overlooked in existing research. Taking Lighthouse Factory certification (LFC) as a quasi-natural experiment, this paper uses China’s county-level panel data from 2016 to 2023 and adopts the DID model to investigate the impact of leading enterprises’ digital transformation on regional digital entrepreneurship (RDE). The findings show that LFC promotes RDE by facilitating digital technology transfer, deepening digital technology cooperation, accelerating digital knowledge accumulation, and enhancing local digital industrial competitiveness. Moreover, this effect is more pronounced in regions with stricter environmental regulations and a stronger green transformation climate, yet is less constrained by local digital infrastructure. Interestingly, LFC exerts positive spillover effects on surrounding cities within 50–150 km and those beyond 250 km, whereas it exerts a significant siphon effect on cities within 50 km. Furthermore, LFC generates network spillovers among economically connected cities through regional digital technology transfer and cooperation networks. This paper provides empirical evidence for leveraging the demonstration effect of leading enterprises to promote the coordinated implementation of SDG 8, SDG 9, SDG 10, SDG 12 and SDG 13. Full article

To satisfy the demands of 5G communication and reliable data connectivity for unmanned aerial vehicles (UAVs), a novel two-element dual-band MIMO antenna with an inherent self-decoupling property based on orthogonal linear polarization diversity is proposed. Distinct from conventional designs relying on extra decoupling components, the antenna realizes isolation enhancement via coupled currents between annular strips and S-shaped strips without additional decoupling structures, representing the core design novelty. Fabricated on a low-cost 1.6 mm thick FR4 substrate, the antenna features compact overall dimensions of 60 mm × 30 mm × 1.6 mm, covering the 2.40–2.73 GHz ISM band and 3.38–3.63 GHz 5G Sub-6 GHz band. Measured results demonstrate that the reflection coefficient remains below −10 dB across the entire operating bands, with port isolation exceeding 27 dB for the 2.4 GHz band and 20 dB for the 3.5 GHz 5G band. The measured realized gain is 0.7–1.5 dB in the lower band and 2.3–2.9 dB in the upper band. The radiation efficiency, which is obtained exclusively from ANSYS HFSS 2025 R1 simulation, is higher than 90% for the lower band and over 80% for the upper band. The calculated envelope correlation coefficient (ECC) is less than 0.15 throughout the working bandwidth, which effectively suppresses inter-channel electromagnetic interference and mitigates channel fading caused by varying UAV attitudes to improve system channel capacity. Further verifications via epoxy encapsulation and co-simulation on an eight-rotor UAV platform prove slight frequency drift after packaging and installation, whereas its bandwidth and isolation still meet practical engineering requirements. Benefiting from a compact layout and omnidirectional radiation performance, the proposed low-cost MIMO antenna is convenient for conformal integration into a UAV fuselage, improving the practicability of UAV-aided emergency communication, equipment inspection and 5G network coverage. Full article

Non-attributed graph sequence offers a powerful formalism for modeling the structural dynamics of complex systems—such as social networks, urban infrastructures, and document transmission pathways—where vertex interactions evolve over time without explicit attribute information. Mining association rules from such sequences to uncover recurring topological patterns have attracted growing interest. Yet two fundamental challenges remain: (1) how to effectively encode edge-level temporal dynamics in non-attributed settings, and (2) how to perform efficient and semantically meaningful temporal association rule mining under structural uncertainty. To address these within a systems-oriented framework, we propose two novel algorithms: the weighted temporal association rule mining algorithm and the fuzzy weighted temporal association rule mining algorithm. The first algorithm introduces time-dependent numerical weights to quantify the strength and persistence of vertex connectivity, integrating them into support and confidence measures to capture both the intensity and evolution of interactions. The second algorithm extends this by incorporating fuzzy set theory, modeling ambiguous or context-sensitive relationships (e.g., indistinct links or weakly correlated vertices) and generating fuzzy-weighted rules that enhance interpretability for real-world system analysis. Evaluated through five comprehensive experiments across diverse datasets and scales using standard metrics (support, confidence, rule count, running time), our methods produce more selective rule sets and achieve lower computational times compared to the classical Apriori algorithm. The proposed approaches thus establish a robust, data-driven foundation for analyzing temporal evolution and structural uncertainty in dynamic complex systems—providing a generalizable methodology applicable beyond domain-specific constraints. Full article

Against the backdrop of frequent global shocks and increasingly complex supply chain networks, supply chain disruption risk exposure has become a major challenge affecting firms’ operational stability and sustainable competitive advantage. Meanwhile, generative artificial intelligence is being increasingly embedded in business operations and has demonstrated strong application potential in information processing, risk identification, and decision support. Based on data from Chinese A-share listed firms from 2017 to 2024 and using text measures based on Management Discussion and Analysis (MD&A) disclosures of Generative AI application and supply chain disruption risk exposure, this study examines the relationship between Generative AI application and corporate supply chain disruption risk exposure, and further explores the channels through which this relationship may operate from the perspective of risk governance transformation. The results show that Generative AI application is significantly associated with lower corporate supply chain disruption risk exposure, and this relationship remains robust across a series of robustness checks and supplementary endogeneity analyses. Channel analyses suggest that this relationship may be related to firms’ risk governance transformation, mainly reflected in enhanced risk identification capability, improved resource allocation capability, and strengthened collaborative response capability. Heterogeneity analyses show that this association is more pronounced among firms facing higher environmental uncertainty, manufacturing firms, and firms located in cities with lower entrepreneurial vitality. This study provides text-based firm-level evidence for understanding the relationship between Generative AI application and supply chain risk governance, and offers managerial implications for firms seeking to promote scenario-based Generative AI application and enhance supply chain resilience and risk governance capability. Full article

This study focuses on the synthesis of glucovanillin mediated by UGT109A1 and its mechanism against Type 2 Diabetes Mellitus (T2DM). Recombinant UGT109A1 successfully synthesized glucovanillin from vanillin using UDP-Glc as the sugar donor. Through network pharmacology, 140 potential targets were identified. Seven key targets were further screened using LASSO and SVM-RFE algorithms. Among these, SLC5A1 and ADK showed strong diagnostic potential, with AUC values ranging from 0.85 to 0.89. Immune infiltration analysis linked these core targets to M2 macrophages. Single-cell transcriptomics revealed that ADK is widely expressed but enriched in B cells, while TLR9 is confined to plasmacytoid dendritic cells (pDCs). Cell-to-cell communication analysis identified a pDC-to-B cell signaling axis. In vitro assays demonstrated that glucovanillin exhibits concentration-dependent inhibitory activity against α-glucosidase with moderate potency, with an IC₅₀ of 413.84 ± 12.80 μM. Molecular docking, 200 ns molecular dynamics simulations (MD), and MM/PBSA calculations showed that glucovanillin binds more strongly to α-glucosidase (−7.4 kcal/mol) than vanillin (−5.4 kcal/mol). Therefore, the glycosylation mediated by UGT109A1 enhanced the bioactivity and targeting specificity of vanillin. In summary, glucovanillin exerts anti-T2DM effects through a dual mechanism involving α-glucosidase inhibition and regulation of key targets, making it a promising lead compound for T2DM treatment. Full article

Deep learning has achieved remarkable progress in electrocardiogram (ECG) analysis, but its heavy dependence on labeled data greatly increases annotation cost. This work proposes wall-aligned vector embedding (WAVE), a self-supervised learning framework that effectively extracts prior knowledge from unlabeled ECGs to reduce reliance on labels. WAVE fully leverages the diversity, synergy, and lead correlation of multi-lead ECGs by explicitly incorporating the correspondence between ECG leads and cardiac walls. Specifically, a multi-branch network captures lead-wise diversity; wall-wise synergy is modeled by concatenating leads from the same wall and projecting them via shared projection; and a dual alignment task is designed to learn correlations both within and across cardiac walls. Experimental results demonstrate that WAVE consistently surpasses all baselines under various evaluation settings, and maintains strong performance even when only a small fraction of labeled ECGs is available. Furthermore, components such as dual alignment, shared projection, wall-based concatenation, and mean target embedding are empirically verified to significantly enhance pretraining quality. In summary, WAVE learns highly informative ECG representations from unlabeled data, enabling low-cost and label-efficient ECG analysis for real-world cardiovascular diagnostics. Full article

The health system is now more complex, uncertain, interdependent, and dynamically interconnected than ever, making traditional planning decisions based on static, reductionist models increasingly impracticable. Systems analytics approaches such as system dynamics, agent-based modeling, and network analysis are often deployed in isolation and fail to capture cross-level interactions and emergent system behavior. This study proposes an integrated multi-layer systems analytics framework that combines these analytical paradigms within a unified architecture to support adaptive health systems planning under uncertainty. The proposed framework introduces an Adaptive Policy Intelligence Layer (APIL), which enables continuous feedback-driven policy adaptation through dynamic monitoring, scenario evaluation, and real-time adjustment mechanisms. The model is evaluated under multiple simulation scenarios, including baseline conditions, demand shocks, resource constraints, and digital transformation environments. The findings provide strong quantitative and analytical evidence of improved system performance and resilience. More specifically, the digital transformation scenario achieved the lowest mean system pressure (0.128) and the highest resilience index (0.887), while the demand shock scenario produced the highest peak system pressure (0.306). The results demonstrate enhanced system resilience, more efficient resource deployment, and superior policy responsiveness compared with traditional single-method approaches. The originality of this study lies in integrating multi-level systems analytics with adaptive policy intelligence into a unified, feedback-driven decision-support framework for resilient health systems governance. The study contributes to systems analytics literature by advancing a synergistic and adaptive modeling paradigm capable of supporting policymakers in navigating complex and unstable healthcare environments. Full article

The study examines the evolution, knowledge structure, and trends in snowmelt runoff prediction models. It identifies research hotspots, future directions, and offers a theoretical basis for accurate simulation and prediction. Utilizing CiteSpace software, 556 core Chinese and English publications from 2010 to 2025 were visually analyzed. Research on snowmelt runoff simulation shows: (1) Chinese publications are prominent in core journals like “Journal of Glaciology and Geocryology,” while English publications appear in high-impact journals like “Water Resources Research.” (2) Institutions like the University of Chinese Academy of Sciences, the Northwest Institute of Eco-Environment and Resources, and the University of California have formed a cross-regional research network. (3) International collaboration involves 42 countries, with a focus on China, the United States, and India. However, domestic institutional cooperation needs improvement. (4) Research trends in snowmelt runoff simulation have progressed from empirical statistics to remote sensing and model-driven physical mechanisms, and now to the integration of artificial intelligence with physical models. (5) The Chinese literature focuses on cold regions, while the English literature emphasizes intelligent modeling. This shift indicates a move towards “physical–intelligent” hybrid modeling. Future research should address challenges like model applicability in data-scarce areas, improving interpretability of complex models, quantifying uncertainties, and developing physically constrained deep learning models. Collaboration among institutions is crucial for enhancing water resource management and disaster warning systems in cold regions. Full article

High-dynamic amusement ride conditions involving impacts, rapid rotations, and abrupt posture changes introduce severe motion artifacts that degrade vital sign quality and destabilize physiological state recognition. This study aims to develop an engineering-ready closed-loop framework for robust passenger monitoring and intelligent diagnosis. A multimodal sensing and modeling pipeline was designed to jointly leverage physiological signals such as heart rate and SpO₂ and kinematic measurements, including acceleration, angular rate, velocity, and attitude. Inertial and PPG signals were preprocessed into supervised samples through wavelet multiresolution denoising and coordinate frame unification, while a strapdown inertial navigation system was used to propagate a 12-channel physical quantity sequence. To ensure interpretability and standards compliance, constraints from GB 8408-2018 were translated into executable threshold rules, enabling standards-driven auto-labeling and rule-based early warning. Building on this foundation, three learning modules were developed: a fusion model for high-dynamic heart rate estimation, a CNN–LSTM dynamic-threshold-enhanced network TAPNet for rapid kinematic anomaly screening, and an attention-augmented hybrid model HS-BANet integrating one-dimensional residual blocks, bidirectional LSTM, and multi-head attention for fine-grained arrhythmia classification. Experimental results demonstrated accurate and consistent heart rate estimation with RMSE of 1.18 bpm on HSSH-I and 1.24 bpm on the independent HSSH-II set, strong agreement with training and testing correlations of 0.9928 and 0.9865, and near-zero bias in Bland–Altman analysis. TAPNet achieved 96.9% validation accuracy and 98.2% test accuracy for kinematic anomaly recognition, maintaining robust generalization under class imbalance. HS-BANet enabled multi-class identification of PVC, PAC, VT, SVT, and AF, achieving an accuracy of 92.37%, an F1-score of 86.87%, a precision of 88.45%, a sensitivity of 88.14%, and a specificity of 89.42%. Overall, the proposed two-stage multimodal closed-loop—fast, interpretable early warning based on physical quantity thresholds followed by fine-grained diagnosis from physiological signals—supports stable feature extraction and reliable decision-making under strong motion artifacts and non-stationary dynamics, balancing responsiveness and diagnostic credibility, while showing potential for practical safety early warning and future deployment-oriented operational support in amusement ride scenarios. Full article

This paper addresses the consensus problem for networked robotic sensors characterized by general linear dynamics and strict communication bandwidth limitations. We propose a weighted sliding integration-based adaptive dynamic event-triggered control (WSI-ADETC) strategy. First, we design a bounded adaptive parameter using a nonlinear protocol to enhance sensitivity to changes in consensus error. To further alleviate the communication burden on the sensing network, we propose a weighted sliding integration-based event-triggering mechanism to reduce the number of triggers compared to traditional adaptive dynamic event-triggered control (ADETC) approaches. Using Lyapunov analysis, we establish sufficient conditions for asymptotic consensus and demonstrate that the proposed controller effectively eliminates Zeno behavior. Numerical simulations demonstrate that the proposed WSI-ADETC strategy significantly reduces communication frequency while maintaining satisfactory consensus performance. Compared with recent adaptive dynamic event-triggered methods, the proposed method reduces the total triggering number by more than 53%, providing a communication efficient solution for resource-constrained robotic sensing networks. Full article