Browse Articles

Advanced cloud intrusion detection framework using graph based features transformers and contrastive learning

Abstract This paper presents a modular and scalable intrusion detection framework that combines graph-based feature extraction, Transformer-based autoencoding, and contrastive learning to improve detection accuracy in cloud environments. Network flows are modeled as graphs to capture relational patterns among IP addresses and services, and a Graph Neural Network (GNN) is used to extract structured embeddings. These embeddings are refined through a Transformer-based autoencoder to preserve contextual information, while contrastive learning enforces clear class separation during classification. The system is evaluated on NSL-KDD and CIC-IDS2018 datasets under both binary and multi-class scenarios. Experimental results show an average accuracy of 99.97%, with high precision and recall across all attack types, including minority classes such as U2R and R2L. The model achieves low false-positive rates and demonstrates real-time inference performance with modest resource requirements. Key contributions include an interpretable pipeline using SHAP for feature attribution, a strategy for mitigating class imbalance, and validation across datasets with detailed security and generalizability analyses. These results support the practical applicability of the proposed approach in high-throughput, cloud-based network environments.

Posterolateral tibial plateau fracture with anterior cruciate ligament injury has biomechanical characteristics of anterolateral rotatory instability through finite element analysis

Reliability analysis of freight elevator systems based on FFMEC, FTA and BN

A deep neural network approach for optimizing charging behavior for electric vehicle ride-hailing fleet

Abstract The rapid advancement of Artificial Intelligence (AI) has led to a profound transformation in the transportation industry, particularly in driving the shift toward carbon neutrality and electrification. AI has proven to be a key enabler in formulating innovative strategies for optimizing electric vehicle (EV) fleets, thus advancing transportation services. While extensive research has been conducted on AI’s role in transportation innovation, there remains a significant gap in empirical studies focusing on optimizing the charging behavior of operational EV fleets, particularly within ride-hailing services. The rise of ride-hailing services has revolutionized the transportation landscape, and their transition to EV fleets presents a major opportunity. The integration of AI to optimize the operations of these EV ride-hailing fleets could substantially help achieve the dual objectives of reducing charging costs and simultaneously lowering carbon emissions. Therefore, this research develops a Neural Network (NN) trained with the Adaptive Moment Estimation (Adam) algorithm, based on 2.14 million charging events. The goal is to analyze current charging behaviors and evaluate the impact of key variables on costs and emissions, providing data-driven insights for potential improvements, thus addressing a critical research gap. The novelty of this study lies in its novel combination of deep learning algorithms with large-scale real-world charging data, proposing a new method for optimizing EV ride-hailing charging behavior, and providing practical solutions for promoting electric vehicle adoption and achieving low-carbon transportation.

Optimizing physical education strategies through circular intuitionistic Fuzzy Bonferroni based school policy formulation

Survival benefit of postoperative chemotherapy in stage III lung squamous cell carcinoma based on SEER database analysis through propensity score matching

Uptake of environmental halophilic archaea by human dendritic cells

Abstract Halophilic archaea are a group distinct from Bacteria and Eukarya, which belong to extremophiles living in highly saline environments. However, they can also exist in the human microbiome. Their impact on the human immune system is poorly known. In this study we examined the interaction of Halorhabdus rudnickae WSM-64T, isolated from the Barycz area of the Wieliczka Salt-Mine in Poland, and of Natrinema salaciae MDB25T from the brine of Lake Medee in Italy, with human monocyte-derived dendritic cells (Mo-DCs). We found that these halophilic archaea invade the cytoplasm and the nucleus of Mo-DCs, but, in contrast to intracellular bacterial pathogens, they do not cause cytotoxic effects on DCs, as no single- or double-stranded DNA breaks (SSB and DSB, respectively), nor chromatin aberrations were noted. Moreover, they did not induce cell cycle alterations, apoptosis or necrosis of DCs. Surprisingly, these halophiles were found to protect against genotoxic activities of Staphylococcus aureus enterotoxin B (SEB), as pre-incubation of the Mo-DCs with the halophilic archaea significantly reduced SEB-induced SSB and DSB, as well as cell cycle disturbance and apoptosis. Therefore, these halophilic archaea can be regarded as safe stimulators for the Mo-DCs to potentially be used as immunomodulators and protective agents for various disorders.

Variations in oral health outcomes and mycobiome composition among COVID-19 convalescents

Study on the influencing factors of land desertification degree in Xilingol Grassland

Abstract The study focuses on the Xilingol grassland in Inner Mongolia, collecting climate, surface, and human-related data for various banners and counties in Xilingol League. Initially, the correlation coefficient method was used to select the nine indicator factors most closely related to the degree of land desertification. Then, the Analytic Hierarchy Process (AHP) was utilized to rank the weights of these factors, constructing a system for evaluating the degree of land desertification, and an evaluation model was established using the comprehensive index method. This model introduced the land desertification evaluation index, dividing the degree of desertification into four levels, with highly and moderately desertified areas mainly distributed in the western part of Xilingol League, while moderately and mildly desertified areas mainly distributed in the eastern and southern regions. The study indicated that climatic factors are the primary factors affecting land desertification, followed by Humanistic and surface factors. Additionally, measures for preventing and controlling land desertification in different regions of the Xilingol Grassland were proposed, offering a scientific basis for formulating desertification prevention and control policies in the Xilingol League.

Comparative performance of viral landscape phylogeography approaches

The rapid evolution of RNA viruses implies that their evolutionary and ecological processes occur on the same time scale. Genome sequences of these pathogens therefore can contain information about the processes that govern their transmission and dispersal. Landscape phylogeographic approaches use phylogeographic reconstructions to investigate the impact of environmental factors and variables on the spatial spread of viruses. Here, we extend and improve existing approaches and develop three novel landscape phylogeographic methods that can test the impact of continuous environmental factors on the diffusion velocity of viral lineages. In order to evaluate the different methods, we also implemented two simulation frameworks to test and compare their statistical performance. The results enable us to formulate clear guidelines for the use of three complementary landscape phylogeographic approaches that have sufficient statistical power and low rates of false positives. Our open-source methods are available to the cientific community and can be used to investigate the drivers of viral spread, with potential benefits for understanding virus epidemiology and designing tailored intervention strategies.

A non-pyrophoric precursor for the low temperature deposition of metallic aluminium

Abstract The development of microelectronics prompts a search for precursors that can deposit conductive features. There is scarce research on Al as it is normally deposited using pyrophoric AlH 3 etherates/aminates. Ligands can impart increased stability while maintaining the ability to deposit target materials. Accordingly, we have engineered an aluminium complex that can undergo conversion to Al(0) at 100 °C. Our multi-step synthetic design features β-ketoiminate compounds, [Al(R-ketoiminate) 2 Cl] (R = Me, Et, i Pr, Ph and Mes, 1 - 5 ) as starting materials to obtain aluminium hydride complexes: the polymeric amidoalane Li[AlH 2 ( i Pr-Hacnac)AlH 3 ] n ( 6 ) and the imidoalane cluster [AlH 2 AlH 2 (N-Mes) 3 (AlH 2 ּ Li(Et 2 O) 2 ) 2 ] ( 8 ). Decomposition of 8 into aluminium metal is observed when heated under vacuum at 100 °C and is confirmed by XRD, TEM, XPS. Deposition of a highly conductive film of Al is achieved from 8 after three weeks under nitrogen at room temperature. This represents a route to metallic aluminium involving non-pyrophoric precursors at low temperatures.

Evolutionary conservation of the grape sex-determining region in angiosperms and emergence of dioecy in Vitaceae

Abstract In bunch grapes (Vitis spp.), flower sex is controlled by a ~ 200-kilobase sex-determining region (SDR) that contains genes involved in floral development. Here, we show that this region evolved from an ancient, highly conserved locus across angiosperms. Comparative genomic analysis of 56 plant genomes identifies homologous regions in all flowering plants but not in non-flowering lineages, suggesting a conserved role in floral function. Within the grape family (Vitaceae), we assemble and phase SDR haplotypes from six species, plus Leea coccinea as an outgroup, and find strong structural conservation, with size variation largely attributable to repetitive elements. Among the dioecious genera, Vitis and Muscadinia exhibit suppressed recombination in the SDR and share candidate sex-determining genes, whereas in Tetrastigma, the region appears to remain recombining, pointing to an alternative mechanism of sex determination. Altogether, our results suggest that dioecy emerged in grapes from a deeply conserved, collinear genomic region composed of multiple genes involved in floral development, morphology, and sexual fertility.

Variation and prognostic potential of the gut antibiotic resistome in the FINRISK 2002 cohort

Abstract The spread of antibiotic-resistant bacteria has severely reduced the efficacy of antibiotics and now contributes to 1 million deaths annually. The gut microbiome is a major reservoir of antibiotic resistance in humans, yet the extent to which gut antibiotic resistance gene load varies within human populations and the drivers that contribute most to this variation remain unclear. Here, we demonstrate, in a representative cohort of 7095 Finnish adults, that socio-demographic factors, lifestyle, and gut microbial community composition shape resistance gene selection and transmission processes. Resistance was linked not only to prior use of antibiotics, as anticipated, but also to frequent consumption of fresh vegetables and poultry, two food groups previously reported to contain antibiotic-resistant bacteria. Interestingly, resistance was not linked to the consumption of high-fat and high-sugar foods, but was consistently higher in females and urban high-income individuals, who currently have generally lower mortality rates. Nevertheless, during the 17-year follow-up, high resistance was associated with a 1.07-fold increase in mortality risk, comparable to elevated blood pressure, and with a heightened risk of sepsis. These findings highlight risks and socio-demographic dimensions of antibiotic resistance that are particularly relevant in the current context of global urbanization and middle-class growth.

Modular Access to Aliphatic Polycarbonates with Tunable Properties and Dual Closed‐Loop Recyclability by Polycondensation–Depolymerization–Repolymerization Strategy

Abstract Great achievements have been made in CO 2 /epoxide copolymerization and dialkyl carbonate/diol polycondensation; however, efficient preparation of high‐molecular‐weight (>100 kDa) aliphatic polycarbonates with tunable properties and recyclability under mild conditions still remains as a great challenge. Herein, we presented a “polycondensation–depolymerization–repolymerization” strategy for structurally diverse aliphatic polycarbonates. This involved a step growth polycondensation of dialkyl carbonate and diol to low‐molecular‐weight (0.5−1.9 kDa) polycarbonates under atmosphere pressure, which are then utilized to produce cyclic carbonate monomers through catalytic depolymerization. The ring‐opening polymerization of cyclic carbonate led to high molecular weight (>100 kDa) polymers, which can be converted back to cyclic monomer via ring‐closing depolymerization or diol/dialkyl carbonate via alcoholysis, enabling chemical recycling of polycarbonates via dual closed loops. The thermal and mechanical properties of the polycarbonates can be widely adjusted by varying the substituent, and polycarbonate with four‐membered spiro‐cyclic substituent shows a recorded high melting temperature (217 °C) and mechanical strength within the reported polycarbonate family. A(hard)‐B(soft)‐A(hard) triblock thermoplastic elastomers with good mechanical performance and elastic recovery were also created by sequential polymerization. The “polycondensation−depolymerization−repolymerization” strategy provided a powerful toolbox for developing high‐performance aliphatic polycarbonates.

A comprehensive evaluation of oversampling techniques for enhancing text classification performance

Abstract Class imbalance is a common and critical challenge in text classification tasks, where the underrepresentation of certain classes often impairs the ability of classifiers to learn minority class patterns effectively. According to the “garbage in, garbage out” principle, even high-performing models may fail when trained on skewed distributions. To address this issue, this study investigates the impact of oversampling techniques, specifically the Synthetic Minority Over-sampling Technique (SMOTE) and thirty of its variants, on two benchmark text classification datasets: TREC and Emotions. Each dataset was vectorized using the MiniLMv2 transformer model to obtain semantically rich representations, and classification was performed using six machine learning algorithms. The balanced and imbalanced scenarios were compared in terms of F1-Score and Balanced Accuracy. This work constitutes, to the best of our knowledge, the first large-scale, systematic benchmarking of SMOTE-based oversampling methods in the context of transformer-embedded text classification. Furthermore, statistical significance of the observed performance differences was validated using the Friedman test. The results provide practical insights into the selection of oversampling techniques tailored to dataset characteristics and classifier sensitivity, supporting more robust and fair learning in imbalanced natural language processing tasks.

Bayesian generalized poisson regression analysis of number of death attributed to household air pollution and associate factors in East Africa from 2010 to 2019 and projection up to 2030

Human clinical trial of plasmapheresis effects on biomarkers of aging (efficacy and safety trial)

Thyroid function test profile in patients who have thyroid nodule and its correlation with ultrasound TIRAD scoring system and cytological results

DNMT3B aggravated renal fibrosis in diabetic kidney disease via activating Wnt/β-catenin signaling pathway

Abstract The incidence of diabetic kidney disease (DKD) has increased rapidly worldwide in recent decades, and DKD is the leading cause of chronic kidney disease. The Wnt/β-catenin pathway is widely recognized as a critical contributor to DKD. However, how this pathway is activated in DKD is still unknown. Recent studies have revealed that epigenetic mechanisms play key roles in DKD. DNA methylation is an epigenetic mechanism that is essential for regulating gene transcription. Here, we demonstrated that reducing the expression of DNMT3B, a DNA methyltransferase, markedly decreased extracellular matrix (ECM) deposition and diabetic renal fibrosis (DRF). Furthermore, we found that DNMT3B activated the Wnt/β-catenin pathway by suppressing SFRP5 expression in HG-induced renal tubular epithelial cells. Mechanistically, we observed that DNMT3B increased the promoter methylation levels of sfrp5, which contributed to a decrease in SFRP5 protein expression. Additionally, Pharmacological disruption of DNA methylation (via 5-Aza) and genetic knockdown of DNMT3B suppressed the Wnt/β-catenin pathway, leading to the attenuation of ECM deposition and DRF. Thus, our study provides a novel understanding of the epigenetic regulation of DKD pathogenesis and a new therapeutic strategy for DKD by disrupting the Wnt/β-catenin pathway.

Attitude of patients with liver cancer toward clinical drug trials: empirical qualitative research