Search Results (437)

As High-Performance Computing (HPC) architectures evolve towards the Exascale, storage infrastructure reliability has emerged as a critical operational challenge, with traditional reactive and static preventive maintenance strategies proving increasingly insufficient. This study addresses this gap by proposing a comprehensive methodological framework for the transition from predictive to predictive-prescriptive maintenance in large-scale storage environments. By integrating the CRISP-DM industry standard with a multi-layered eXplainable Artificial Intelligence (XAI) suite, we develop a system capable of isolating hardware degradation signals amidst massive volumes of routine telemetry. To validate our approach, we leveraged a publicly available disk failure dataset to evaluate multiple Machine Learning configurations, addressing the challenge of severe class imbalance through optimized oversampling and Gradient Boosting algorithms. The methodology employs global and local XAI techniques, including Permutation Feature Importance, SHAP, and surrogate decision trees, to translate probabilistic risk assessments into auditable hardware engineering rules. Our results demonstrate that this hybridization of robust predictive modeling with multi-layered explainability provides a transparent, evidence-based decision support system. Ultimately, we conclude that converting opaque risk predictions into technical justifications enables infrastructure managers to optimize hardware lifecycle management and minimize system downtime in mission-critical environments, establishing a viable pathway toward more resilient and auditable storage management. Full article

Bovine endometritis remains one of the most significant postpartum uterine disorders. It impairs uterine recovery, compromises fertility, and causes substantial economic losses in dairy production. Growing evidence suggests that the disease cannot be attributed solely to postpartum bacterial contamination; rather, it should be understood as a multifactorial failure to restore uterine homeostasis after calving. This review summarises the latest research findings on six interconnected aspects: the clinical significance of postpartum uterine disease; the diagnostic and biological differences between clinical and subclinical endometritis; the role of microbes in the uterus in health and disease; interactions between the host and uterine bacteria; the mechanisms of persistent inflammatory regulation; and current as well as emerging treatment strategies. Current evidence indicates that postpartum uterine disease is more strongly associated with dysbiosis, reduced microbial diversity, and disturbed microbial succession than with the presence of any single pathogen. Disease progression is driven by complex interplay among microbial ligands, epithelial and stromal immune responses, virulence-associated tissue injury, endocrine disruption, and impaired inflammatory resolution. Furthermore, persistent uterine inflammation is regulated by multilayered networks involving cytokines, prostaglandins, noncoding RNAs, extracellular vesicles, metabolic remodeling, and oxidative stress. Although conventional therapies remain relevant in certain clinical cases, microbiota-oriented approaches, particularly probiotic interventions, have emerged as promising adjunctive strategies for the prevention and control of the condition. Overall, bovine endometritis should be viewed as a disorder caused by disrupted interactions between the host, microbiota and inflammation. Future progress will depend on longitudinal, strain-resolved, and function-oriented studies to enable more precise and less antimicrobial-dependent interventions for postpartum uterine health. Full article

The exponential growth of cyber threats in modern digital infrastructure demands advanced detection systems that adapt to evolving attack patterns. Traditional cybersecurity approaches struggle with dynamic threats, requiring extensive labeled datasets and retraining for each new category. This paper presents a comprehensive transfer learning framework for cybersecurity threat detection, leveraging the CICIoMT dataset as a benchmark to enhance detection capabilities across heterogeneous cybersecurity environments. We propose a machine learning (ML)-enabled framework that employs systematic feature alignment, hybrid class balancing, and multi-algorithm evaluation using machine learning models, including Multi-Layer Perceptron (MLP), Support Vector Machine (SVM), Random Forest (RF), Gradient Boosting, and XGBoost. The proposed approach addresses the critical challenges of data scarcity and domain heterogeneity in cybersecurity by enhancing feature engineering with cybersecurity-specific features, statistical aggregations, and PCA embeddings. Extensive experimental evaluation across two target datasets (CICIoT and IoT-23) demonstrates both the exceptional successes and critical limitations of cross-domain transfer learning in cybersecurity. The framework achieved outstanding performance on domain-compatible datasets, with RF reaching 99.0% accuracy on CICIoT, Gradient Boosting achieving 98.9%, and XGBoost delivering 98.4%, demonstrating exceptional knowledge transfer from medical IoT to smart home IoT environments. However, transfer learning to IoT-23 was unsuccessful (50% accuracy, equivalent to random guessing), revealing that feature domain difference, where identical attack labels encode fundamentally different behavioral patterns, prevents effective knowledge transfer despite nominal class overlap. This research makes significant advances in adaptive cybersecurity systems by providing a rigorous evaluation of both the successes and limitations of transfer learning. This work demonstrates that ensemble methods (RF, XGBoost, and Gradient Boosting) achieve superior cross-domain performance compared with neural networks on compatible domains, while also revealing fundamental challenges when the source and target domains differ in their feature spaces. Full article

This study comprehensively investigates the degradation performance of a Mn-doped Zn₂SnO₄ photocatalyst based on time-dependent UV-Vis absorption spectra. Before machine learning modelling, the effects of experimental parameters such as UV–Vis measurement wavelength, reaction time, and Mn doping ratio were statistically validated using One-Way Analysis of Variance (ANOVA) and Multiple Linear Regression (MLR) methods. To overcome the limitations of linear models in representing complex physical systems, an optimized Multi-Layer Perceptron (MLP) architecture was developed to capture the system’s nonlinear dynamics with high accuracy. To validate the model’s out-of-sample prediction capability and prevent data leakage potentially arising from spectral data correlation, the “Leave-One-Doping-Level-Out” (LODLO) cross-validation strategy was applied, during which performance metrics of R²$=0.8889$ and $MSE=0.00238$ were recorded. To make the neural network’s decision-making mechanism transparent, a dual-validation explainability framework comprising Shapley Additive Explanations (SHAP) and Permutation Feature Importance analyses was employed. By quantifying the relative contributions of the experimental parameters to the model predictions, this approach revealed that the UV–Vis measurement wavelength was the dominant predictive variable, followed by the Mn doping ratio and reaction time. This study presents a transparent methodology that offers both strong predictive capability and physically grounded data to shed light on the complex interactions in doped semiconductor photocatalysts. Full article

The corrosion of metal components leads to substantial economic losses and poses serious safety hazards. While organic coatings are regarded as an effective countermeasure, conventional epoxy resins (EPs) often exhibit high brittleness and insufficient corrosion resistance after curing. To overcome these limitations, this study proposes a novel modification strategy. A multilayer graphene-reinforced epoxy composite coating was fabricated via a layer-by-layer spraying process, employing uniformly dispersed modified aramid nanofibers (ANFs) and low-defect graphene as functional fillers. Polydopamine (PDA) was utilized to improve the dispersion of graphene oxide (GO), mitigate defect-associated permeation pathways, and enhance the interfacial bonding between the graphene layer and the epoxy matrix, thereby ensuring coating integrity. Tannic acid (TA) effectively improves the dispersion of ANF within the epoxy, preventing stress concentration. The corrosion resistance and mechanical properties of the composite coating were systematically evaluated. Results demonstrate that the coating achieves a low-frequency impedance of 1.98 × 10¹⁰ Ω·cm². With the incorporation of 0.05% TA-modified ANFs, the elongation at break increases to 68.79%, and the impact resistance is significantly enhanced, with the impact height reaching 50 cm. The composite coating preparation strategy in this work offers a novel approach for constructing multifunctional composite coatings, demonstrating broad application prospects. Full article

Structured Query Language injection (SQLi) remains a persistent threat to web applications due to the obfuscation, diversity, and evolving structure of malicious payloads, which limit the effectiveness of conventional rule and signature-based Web Application Firewalls (WAFs). Although prior studies have reported high detection performance using individual feature extraction methods or offline classification models, limited work has addressed deployment-oriented SQLi prevention through an integrated real-time inspection framework. This paper proposes a Machine Learning (ML)-based SQLi detection and prevention framework that combines hybrid feature representation, supervised dimensionality reduction, Genetic Algorithm (GA)-based hyperparameter optimization, and real-time WAF validation. Multiple public SQLi datasets were merged, cleaned, and deduplicated to improve exposure to diverse query patterns. SQL queries were encoded using Term Frequency–Inverse Document Frequency (TF-IDF), Word2Vec, and FastText features, which were fused and transformed through a Supervised Autoencoder into a compact discriminative representation. GA was then employed to optimize multiple classifiers, including Random Forest (RF), Support Vector Machine (SVM), Decision Tree (DT), and Multi-Layer Perceptron (MLP). The MLP achieved the best overall performance, with an accuracy of 0.998681. The optimized model was deployed within a lightweight Flask-based WAF for real-time Hypertext Transfer Protocol (HTTP) request inspection and malicious input blocking. SQLMap v1.8.4-based robustness testing and runtime analysis demonstrate that the proposed framework provides effective SQLi prevention with practical deployment efficiency beyond conventional offline benchmark evaluation. Full article

Financial information plays a critical role in decision-making for stakeholders, including investors, regulators, and corporate managers. However, financial data is susceptible to deliberate manipulation, where some firms may distort disclosures to mislead stakeholders and potentially engage in fraudulent activities. With the rapid expansion of capital markets and advancements in information technology, financial fraud has grown increasingly sophisticated and concealed. As a result, conventional detection methods often struggle to identify emerging fraud patterns, rendering fraud prevention increasingly complex and less effective. In this paper, we propose a novel multi-layer architecture model that integrates business, internal control, and strategic features. Our framework leverages multi-layer neural networks for effective feature extraction and concatenates their outputs for classification. Furthermore, we develop this framework by incorporating explainable artificial intelligence (XAI) techniques to enhance interpretability. Empirical results show that the proposed framework provides competitive discriminatory ability and produces conservative, low-false-alarm fraud warnings under the full multi-layer feature setting while also offering interpretable insights for the diverse needs of stakeholders. This study contributes to the development of fraud detection tools that are both operationally useful and interpretable. Full article

The detection and degradation analysis of subsurface microcracks in mural paintings remain challenging due to their inhomogeneous multilayered structure and complex deterioration mechanisms. In this study, we propose a multimodal stepwise method for three-dimensional characterization and cross-scale degradation analysis by integrating digital holography (DH), infrared thermography (IRT), acoustic excitation (AE), and molecular dynamics (MD) simulations. In the first step, an adjustable field-of-view (FOV) digital holographic system is developed to capture subsurface deformation under acoustic excitation, enabling high-resolution planar characterization of subsurface microcracks. Infrared thermography is then employed to estimate crack depth through an inverse thermal model, achieving full three-dimensional reconstruction of crack geometry. Based on the reconstructed structures, MD simulations are conducted to investigate the evolution of stress, bond breaking, and crack propagation under varying temperature and humidity conditions, with particular emphasis on water molecule migration and chemically induced degradation. The results demonstrate that environmental factors promote stress concentration and material embrittlement at crack tips, leading to secondary microcrack formation and progressive deterioration. Experimental aging tests show strong agreement with simulation results, validating the proposed methodology. This work establishes a unified “characterization–simulation–validation” paradigm, providing new insights into the mechanisms of mural degradation and offering a robust framework for non-destructive evaluation and preventive conservation of multilayer cultural heritage materials. Full article

In the process of ascending repetitive extraction of multilayer coal seams, the evolution and expansion of excavation-triggered cracks dominate the structural instability of overlying rock layers and the deterioration of overall rock mass quality. In this study, a similar material model containing two coal seams (No. 2 upper coal seam and No. 4 underlying seam) was constructed. Through model excavation, the entire process of overburden movement was captured, and a box-counting program was developed for quantitative analysis to investigate the dynamic evolution law of overlying strata fractures under upward repeated mining disturbances. The results show that during the mining of the No. 4 coal seam, the fractal dimension of mining-induced fractures initially increases and then decreases, with the peak value occurring at an excavation distance of approximately 40 cm. For the No. 2 coal seam, the fractal dimension exhibits fluctuating evolution characteristics, with the global peak appearing at approximately 55 cm and a local peak at approximately 70 cm. The rate of change in fractal dimension during coal seam mining exhibits alternating fluctuations of increase and decrease. The initially mined No. 4 coal seam, subject to higher in situ stress, exhibits rapid fracture development with an earlier peak, while the No. 2 coal seam, influenced by pressure relief, presents progressive fracturing characteristics with a delayed peak. The fractal dimension can effectively characterize the evolution characteristics of the mining-induced fracture network in overlying strata and provide reference for strata control and disaster prevention. Full article

Lentiviral vectors (LVs) are indispensable tools in cell and gene therapy. Rising demand has created a global shortage of LVs, driving the development of novel packaging approaches. We report a novel vector packaging approach using autonomously replicating cytoplasmic RNAs (replicons) to express packaging proteins. Yellow fever virus (YFV) was used as a source of replicons encoding the HIV-1 Gag–Pol polyprotein together with reporter or selectable markers. YFV replicons were able to establish chronic infection in HEK293FT cells. Replicons expressing HIV-1 Gag–Pol containing the wild-type HIV-1 protease caused strong cytotoxicity, which prevented the selection of polyclonal cell pools harboring the replicon. In contrast, a replicon carrying the T26S mutation in the HIV-1 protease gene showed no measurable cytotoxic effects, enabling the generation of stable replicon-containing cell pools. The replicon cell pools were established using antibiotic selection and maintained Gag-Pol expression for at least ten passages under selection pressure. Using these first-generation replicon cell pools as packaging cells, LV production required only transient transfection of a transfer vector, a Tat/Rev plasmid, and an envelope plasmid. Yields reached ~10⁶ TU/mL prior to concentration and ~10⁹ TU from multilayer cell stacks, which fall within the range typically reported for conventional transient transfection systems under similar culture conditions. The resulting vectors efficiently transduced target cells, and no replication-competent lentivirus (RCL) was detected using a two-phase RCL assay with p24 ELISA detection. This demonstrator platform utilizing replicon cell pools represents a novel approach for LV packaging. Full article

Precision battery-health monitoring and rapid degradation detection are essential for improving the security, durability, and efficacy of electric vehicles (EVs). By incorporating short-term State-of-Health (SoH) forecasting, mid-term deterioration alarms, and degradation acceleration risk modeling into a temporally consistent machine learning architecture, this research suggests a hierarchical predictive-maintenance framework. The rolling-origin cross-validation approach is implemented to maintain the chronological order of the data and prevent any potential information leaks. The predictive core employs an ensemble learning approach that integrates Random Forest, Extremely Randomized Trees, and Histogram-Based Gradient Boosting. Validation-driven model blending and training only feature selection are implemented to improve generalizability. The one-hour SoH forecasting model for short-horizon monitoring exhibits exceptional accuracy in an assessment of health prediction, with an R² of 0.9254, an RMSE of 0.0033, and a MAPE of 0.32%. Early detection of anomalies and the provision of a seven-day degradation warning may be achieved by a proactive maintenance scheduling model with an area under the curve (AUC) of 0.7838 and a recall of 0.8205. In addition, the degradation acceleration risk module could identify rapid health decline with a robustness of 0.8796 and a precision–recall AUC of 0.7101 when operating under significant stress. Reliability in critical domains is demonstrated through validation using scenarios that simulate severe temperature and stress conditions. Achieving intelligent predictive maintenance of electric vehicle battery packs is now feasible due to the proposed multi-layer ensemble structure. Full article

This paper presents the Data Quality Assurance Model (DQAM), a formal model and multilayered architecture designed to guarantee data integrity and robustness in Reliable and Quality-Assured Information Systems. Recognising that inaccurate or corrupted sensor data can lead to system collapses and false alarms in critical services, the DQAM provides a standardised and systematic flow of actions to ensure data excellence for Artificial Intelligence (AI). The architecture is structured into three specialised layers (Acquisition, Processing, and AI Adequacy), implementing formal transformation functions that act as a rigorous filter against data degradation. A core contribution is the mapping of these functions to ISO/IEC 25012 and 5259-2 standards, providing a practical framework for reliable information management. It should be noted that quality dimensions regarding timeliness and data volume are outside the scope of this work, as they depend on external data issuers and end-service requirements. The model’s viability is validated through a real-world implementation on a university campus managing millions of data points, demonstrating its capability to optimise performance—achieving a speedup of up to 43%—and prevent service malfunctions. This work bridges the gap between raw IoT streams, and the high-integrity standards required by modern AI-driven applications. Full article

Accurate identification of weak nodes is a prerequisite for online security assessment, preventive control, and resilience enhancement in modern power systems. However, conventional single-layer graph-learning models mainly emphasize local neighborhood aggregation and are insufficient for characterizing vulnerability propagation from equipment-level disturbance to regional congestion and system-level transfer constraints. This paper proposes a mechanism-aware hierarchical graph-learning framework for weak-node identification in complex interconnected power grids. We emphasize that attention, fusion, and gating operations are standard neural-network mechanisms and are not claimed as new generic deep-learning blocks. The contribution of this paper is the power-system-specific formulation: constructing an electrically meaningful local-supernode hierarchy, defining reproducible mechanism-based node and branch-vulnerability proxies, and interpreting weak-node rankings through node–line–corridor coupling evidence. In the validated implementation, a local graph convolutional encoder and a supernode/global graph convolutional encoder generate 32-dimensional local embeddings and 16-dimensional global embeddings, which are concatenated and decoded by a 48 → 24 → 1 multilayer perceptron to obtain node vulnerability scores. Experiments are conducted on reproducible IEEE benchmark data generated from pandapower standard systems, with representative comparisons on the IEEE 57-bus, 145-bus, and 300-bus systems and a detailed structural interpretation on the IEEE 145-bus case. The present results validate the ability of the implemented local–global hierarchical model to reproduce the proposed mechanism-based vulnerability proxy on representative small- and medium-scale benchmarks. Full article

BK and JC Polyomaviruses (BKPyV and JCPyV) are ubiquitous human pathogens capable of establishing lifelong, asymptomatic persistence in the majority of the global population. While decades of research have focused on their lytic replication cycles and the development of severe diseases, such as polyomavirus-associated nephropathy (PVAN) caused by BKPyV and progressive multifocal leukoencephalopathy (PML) caused by JCPyV, their primary evolutionary strategy is one of persistence rather than pathogenesis. This review shifts the perspective from a replication-centric framework towards an evolutionary persistence model, detailing the multi-layered host and viral determinants that maintain the homeostatic balance. At the cellular level, viral genomes are restricted by chromatinization into minichromosomes and host S-phase licensing. These constraints are reinforced by innate immune sensing and adaptive T-cell and antibody responses that curtail systemic dissemination while permitting periodic, low-level urinary shedding, which is essential for horizontal transmission. In addition to these host barriers, the viruses utilize intrinsic regulatory mechanisms to prevent excessive replication and immune detection, including the stable archetype non-coding control region (NCCR), viral microRNAs that downregulate early gene expression, and the small t antigen (STAg). Finally, we address unresolved questions regarding the full spectrum of cellular reservoirs, the molecular triggers of reactivation, and the ecological factors shaping their transmission routes. Understanding these maintenance mechanisms is crucial for refining clinical interventions and managing the rare, devastating transitions from silent persistence to lytic disease. Full article

The Yellow River Delta, a young alluvial plain in China, is experiencing severe land subsidence that threatens its ecological security and sustainable development. However, the driving mechanisms of this subsidence exhibit strong spatial heterogeneity, which traditional global models fail to capture. This study integrates high-precision subsidence measurements from Sentinel-1A imagery and SBAS-InSAR technology (2017–2023) with multi-source environmental factors (topography, geology, land use, precipitation) to propose a Spatially Adaptive Ensemble Learning Model with feature selection (SA-GSE). The model concatenates predictions from base learners (CatBoost, XGBoost, Random Forest) with spatial features (e.g., distance to salt pans, local topographic variance) to form meta-features, which are then input into a multilayer perceptron meta-learner. Through 5-fold spatial cross-validation, SA-GSE learns spatially dynamic base-model weights, implicitly adapting to regional variations in subsidence drivers. The model achieves an R² of 0.7810 and RMSE of 40.55 mm/yr on the test set, outperforming individual base models and ordinary stacking. Residual spatial autocorrelation is substantially reduced, with SA-GSE yielding the lowest Moran’s I (0.0334, p = 0.206) among all evaluated models, confirming effective capture of spatial heterogeneity. Driving force analysis reveals that distance to salt pans is the most important predictor (permutation importance: 0.4456), underscoring the dominant role of brine extraction-induced aquifer compaction. Lagged precipitation importance (0.3191) exceeds that of current precipitation (0.2453), indicating a recharge lag effect. SHAP interaction analysis uncovers a nonlinear “precipitation decoupling” mechanism in salt pan areas, where high precipitation paradoxically exacerbates subsidence. The resultant map of predicted subsidence rates highlights elevated rate zones in the northern salt pans and along the Guangli River. While the map does not represent a full risk assessment—as it does not include exposure or vulnerability—it provides a spatially explicit estimate of hazard likelihood. This ensemble framework yields novel perspectives on subsidence drivers in heterogeneous regions and can support land subsidence prevention and groundwater management planning. Full article