Search Results (992)

As cyberattacks continue to rise in frequency and sophistication, extracting actionable Cyber Threat Intelligence (CTI) from diverse online sources has become critical for proactive threat detection and defense. However, accurately identifying complex entities from lengthy and heterogeneous threat reports remains challenging due to long-range dependencies and domain-specific terminology. To address this, we propose XLNet-CRF, a hybrid framework that combines permutation-based language modeling with structured prediction using Conditional Random Fields (CRF) to enhance Named Entity Recognition (NER) in cybersecurity contexts. XLNet-CRF directly addresses key challenges in CTI-NER by modeling bidirectional dependencies and capturing non-contiguous semantic patterns more effectively than traditional approaches. Comprehensive evaluations on two benchmark cybersecurity corpora validate the efficacy of our approach. On the CTI-Reports dataset, XLNet-CRF achieves a precision of 97.41% and an F1-score of 97.43%; on MalwareTextDB, it attains a precision of 85.33% and an F1-score of 88.65%—significantly surpassing strong BERT-based baselines in both accuracy and robustness. Full article

Background/Objectives: In routine practice, colonoscopy findings are saved as unstructured free text, limiting secondary use. Accurate named-entity recognition (NER) is essential to unlock these descriptions for quality monitoring, personalized medicine and research. We compared named-entity recognition (NER) models trained on real, synthetic, and mixed data to determine whether privacy preserving synthetic reports can boost clinical information extraction. Methods: Three Spark NLP biLSTM CRF models were trained on (i) 100 manually annotated Romanian colonoscopy reports (ModelR), (ii) 100 prompt-generated synthetic reports (ModelS), and (iii) a 1:1 mix (ModelM). Performance was tested on 40 unseen reports (20 real, 20 synthetic) for seven entities. Micro-averaged precision, recall, and F1-score values were computed; McNemar tests with Bonferroni correction assessed pairwise differences. Results: ModelM outperformed single-source models (precision 0.95, recall 0.93, F1 0.94) and was significantly superior to ModelR (F1 0.70) and ModelS (F1 0.64; p < 0.001 for both). ModelR maintained high accuracy on real text (F1 = 0.90), but its accuracy fell when tested on synthetic data (0.47); the reverse was observed for ModelS (F1 = 0.99 synthetic, 0.33 real). McNemar χ² statistics (64.6 for ModelM vs. ModelR; 147.0 for ModelM vs. ModelS) greatly exceeded the Bonferroni-adjusted significance threshold (α = 0.0167), confirming that the observed performance gains were unlikely to be due to chance. Conclusions: Synthetic colonoscopy descriptions are a valuable complement, but not a substitute for real annotations, while AI is helping human experts, not replacing them. Training on a balanced mix of real and synthetic data can help to obtain robust, generalizable NER models able to structure free-text colonoscopy reports, supporting large-scale, privacy-preserving colorectal cancer surveillance and personalized follow-up. Full article

Background/Objectives: Cardiorespiratory fitness (CRF) is of great interest in children and adolescents. Due to the limited availability of cardiopulmonary exercise testing, simple and reliable alternatives are needed. A stair climbing test (SCT) for the assessment of CRF developed at the Department of Pediatric Cardiology of the LMU University Hospital in Munich showed a strong correlation with VO₂max. The aim of this study is to prove its feasibility in a non-clinical setting and to analyse its results in a larger study population. Methods: During the “Hand aufs Herz” study, a comprehensive cardiovascular examination was carried out on 922 pupils and siblings (13.2 ± 7.8 years) at a high school in Bavaria. The SCT was performed to evaluate CRF: participants had to run up and down a total of four floors (14.8 m) as quickly as possible without skipping steps or holding on to the banister. Absolute time has been normalized over the standard height of 12 m to allow comparisons with different settings. An SCT Index was calculated to adjust results to the different weights of participants and the exact height of the staircase. Results: The SCT proved to be easily feasible and safe in non-clinical contexts. Out of 922 participants, 13 (1.4%) were not able to perform the test, and 3 (0.3%) had to interrupt it following fatigue or stumbling. A total of 827 participants aged from 9 to 17 years (13.1 ± 2.1 years, 45.8% girls) had a mean absolute SCT time of 53.4 ± 6.2 s and 43.3 ± 5.1 s when normalized over 12 m. Conclusions: The SCT represents a simple, cost- and time-saving test that allows a rapid and solid assessment of cardiorespiratory fitness in children and adolescents. We could demonstrate that it is safe and feasible in non-clinical contexts. Its short duration and universal applicability are valuable advantages that could facilitate the establishment of a repetitive cardiovascular screening in the pediatric population, particularly in outpatient departments or settings with low-resource systems. Full article

Background: Malnutrition and cancer-related fatigue (CRF) are prevalent in cancer patients, significantly impacting prognosis and quality of life. Oral nutritional supplements (ONSs) enriched with protein and ω-3 fatty acids may improve nutritional status and mitigate CRF. This study evaluates the effects of a high-protein, fish oil-enriched ONS (FOHP-ONS) on nutritional intake, body composition, fatigue, and quality of life in malnourished cancer patients. Methods: Cancer patients with malnutrition or inadequate food intake received 8 weeks of FOHP-ONS (2 cans/day, providing 4.2 g/day of ω-3 fatty acids). Dietary intake, body weight, handgrip strength, serum biochemical markers, nutritional status (PG-SGA), fatigue (BFI-T), and quality of life (EORTC QLQ-C30) were assessed at baseline, week 4, and week 8. Results: Of the 33 enrolled patients, 30 completed the study. Energy and protein intake significantly increased (p < 0.05), and body BMI and handgrip strength showed significant improvements (p < 0.05), while muscle mass did not change significantly. Nutritional status, assessed by PG-SGA, improved, with the proportion of severely malnourished patients (Stage C) decreasing from 46.7% to 13.3%, and moderately malnourished patients (Stage B) improving to well-nourished status (Stage A) from 10.0% to 30.0% (p < 0.001). Serum albumin levels increased significantly (p < 0.05), while fasting blood glucose significantly decreased (p < 0.05). Additionally, triglyceride levels significantly decreased (p < 0.05), while total cholesterol and LDL-C showed a downward trend. Cancer-related fatigue scores improved across all domains (p < 0.05), and quality of life significantly increased, particularly in physical and role functioning (p < 0.05). Conclusions: FOHP-ONS supplementation improved nutritional intake, body composition, and muscle strength while alleviating CRF and enhancing quality of life in malnourished cancer patients. These findings support its potential role in nutritional intervention for malnourished cancer patients. Full article

Nested named entity recognition (NER) is a task that identifies hierarchically structured entities, where one entity can contain other entities within its span. This study introduces a nested NER model for few-shot learning environments, addressing the difficulty of building extensive datasets for general named entities. We enhance the Biaffine nested NER model by modifying its output layer to incorporate label semantic information through a novel label description embedding (LDE) approach, improving performance with limited training data. Our method replaces the traditional biaffine classifier with a label attention mechanism that leverages comprehensive natural language descriptions of entity types, encoded using BERT to capture rich semantic relationships between labels and input spans. We conducted comprehensive experiments on four benchmark datasets: GENIA (nested NER), ACE 2004 (nested NER), ACE 2005 (nested NER), and CoNLL 2003 English (flat NER). Performance was evaluated across multiple few-shot scenarios (1-shot, 5-shot, 10-shot, and 20-shot) using F1-measure as the primary metric, with five different random seeds to ensure robust evaluation. We compared our approach against strong baselines including BERT-LSTM-CRF with nested tags, the original Biaffine model, and recent few-shot NER methods (FewNER, FIT, LPNER, SpanNER). Results demonstrate significant improvements across all few-shot scenarios. On GENIA, our LDE model achieves 45.07% F1 in five-shot learning compared to 30.74% for the baseline Biaffine model (46.4% relative improvement). On ACE 2005, we obtain 44.24% vs. 32.38% F1 in five-shot scenarios (36.6% relative improvement). The model shows consistent gains in 10-shot (57.19% vs. 49.50% on ACE 2005) and 20-shot settings (64.50% vs. 58.21% on ACE 2005). Ablation studies confirm that semantic information from label descriptions is the key factor enabling robust few-shot performance. Transfer learning experiments demonstrate the model’s ability to leverage knowledge from related domains. Our findings suggest that incorporating label semantic information can substantially enhance NER models in low-resource settings, opening new possibilities for applying NER in specialized domains or languages with limited annotated data. Full article

Background/Objectives: Since the World Health Organization (WHO) recommended HIV self-testing as an alternative to traditional facility-based testing in 2016, it has been increasingly adopted worldwide. This study aimed to evaluate the performance of the ConfiSign HIV Self-Test (GenBody Inc., Republic of Korea), a newly developed blood-based immunochromatographic assay for the qualitative detection of total antibodies (IgG and IgM) against HIV-1/HIV-2. Methods: The evaluation included four components: (1) retrospective analysis of 1400 archived serum samples (400 HIV-positive and 1000 HIV-negative samples), (2) prospective self-testing by 335 participants (112 HIV-positive participants and 223 individuals with an unknown HIV status, including healthy volunteers), (3) assessment using seroconversion panels and diverse HIV subtypes, and (4) analytical specificity testing for cross-reactivity and interference. The Elecsys HIV combi PT and Alinity I HIV Ag/Ab Combo assays were used as reference assays. Results: In retrospective testing, the ConfiSign HIV Self-Test achieved a positive percent agreement (PPA) of 100%, a negative percent agreement (NPA) of 99.2%, and a Cohen’s kappa value of 0.986, showing excellent agreement with the reference assays. In the prospective study, the test showed 100% sensitivity and specificity, with a low invalid result rate of 1.8%. All HIV-positive samples, including those with low signal-to-cutoff (S/Co) values in the Alinity I assay, were correctly identified. The test also reliably detected early seroconversion samples and accurately identified a broad range of HIV-1 subtypes (A, B, C, D, F, G, CRF01_AE, CRF02_AG, and group O) as well as HIV-2. No cross-reactivity or interference was observed with samples that were positive for hepatitis viruses, cytomegalovirus, Epstein–Barr virus, varicella zoster virus, influenza, HTLV-1, HTLV-2, or malaria. Conclusions: The ConfiSign HIV Self-Test demonstrated excellent sensitivity, specificity, and robustness across diverse clinical samples, supporting its reliability and practicality as a self-testing option for HIV-1/2 antibody detection. Full article

Objective: The aim of this study was to determine the relationship between the components of physical fitness (PF), creativity and fluid intelligence, as well as to determine which components of PF are predictors of the analysed cognitive potential. Material and Methods: A total of 584 Spanish schoolchildren (6−11 years old; age = 8.62 ± 1.77 years) took part in this study. Creativity was assessed using the Torrance Tests of Creative Thinking (TTCT) and fluid intelligence through TEA-1. Moreover, PF components were evaluated using a 25 m sprint, handgrip strength, standing long jump and 20 m SRT. Results: Boys exhibited a better PF performance than girls (p range from = < 0.001 to 0.05), as well as higher creativity score (p < 0.001), the fluid intelligence score and QI score (p < 0.05, respectively). Moreover, PF components (CRF, strength and speed) were positively associated with creativity (p range from = < 0.001 to 0.001) and fluid intelligence (p range from = < 0.001 to 0.015). Regression analysis showed that the creativity model explained between 31.4% and 36.6% of the variance (R² = 0.314−0.366, p < 0.001), while the fluid intelligence model accounted for 25.5% to 33.1% of the variance (R² = 0.255−0.331, p < 0.001 to 0.001). Conclusions: A positive relationship was found between creativity, fluid intelligence, and PF components. Children with higher PF levels scored better in creativity, with notable differences between boys and girls. These findings highlight the educational value of incorporating structured physical activity into school settings to support both cognitive and physical development. Full article

This paper presents a Multimodal Hidden Markov Model (MHMM) framework specifically designed for real-time human proficiency assessment, integrating physiological (Heart Rate Variability (HRV)), behavioral (Task Completion Time (TCT)), and subjective (NASA Task Load Index (NASA-TLX)) data streams to infer latent human proficiency states in industrial settings. Using published empirical data from the surgical training literature, a comprehensive simulation study was conducted, with the MHMM (Trained) achieving 92.5% classification accuracy, significantly outperforming unimodal Hidden Markov Model (HMM) variants 61–63.9% and demonstrating competitive performance with advanced models such as Long Short-Term Memory (LSTM) networks 90%, and Conditional Random Field (CRF) 88.5%. The framework exhibited robustness across stress-test scenarios, including sensor noise, missing data, and imbalanced class distributions. A key advantage of the MHMM over black-box approaches is its interpretability by providing quantifiable transition probabilities that reveal learning rates, forgetting patterns, and contextual influences on proficiency dynamics. The model successfully captures context-dependent effects, including task complexity and cumulative fatigue, through dynamic transition matrices. When demonstrated through simulation, this framework establishes a foundation for developing adaptive operator-AI collaboration systems in Industry 5.0 environments. The MHMM’s combination of high accuracy, robustness, and interpretability makes it a promising candidate for future empirical validation in real-world industrial, healthcare, and training applications in which it is critical to understand and support human proficiency development. Full article

This study evaluated the effects of one-time application of controlled-release fertilizer (CRF) on rice (Oryza sativa L.) grain yield, grain quality, and agronomic nitrogen use efficiency (ANUE, ANUE (kg/kg) = (Grain yield with N application − grain yield without N application)/N application amount) in coastal saline soils. A two-year field experiment (2023–2024) was conducted using two rice varieties (Nanjing 5718 and Yongyou 4953) under four nitrogen treatments: N0 (no nitrogen fertilization), N1 (270 kg·hm⁻², with a ratio of 5:1:2:2 at 1-day before transplanting, 7-day after transplanting, panicle initiation, and penultimate-leaf appearance stage, respectively), N2 (270 kg·hm⁻², one-time application at 1-day before transplanting as 50% CRF with 80-day release period + 50% urea), and N3 (270 kg·hm⁻², 50% one-time application of CRF with 120-day release period at the seedling stage + 50% urea at 1-day before transplanting). Compared with N1, the N3 treatment significantly increased grain yield by 10.2% to 12.9% and improved ANUE by 18.5% to 51.6%. It also improved processing quality (higher brown rice, milled rice, and head rice rates), appearance quality (reduced chalkiness degree and chalky rice percentage), and taste value (by 19.3% to 31.2%). These improvements were associated with lower amylose, protein, and soluble sugar contents and favorable changes in starch composition and pasting properties. While N2 slightly improved some quality traits, it significantly reduced yield and ANUE. Correlation analysis revealed that starch and protein composition, as well as pasting properties, were significantly associated with taste value and related attributes such as appearance, stickiness, balance degree, and hardness. Overall, one-time application of CRF with a 120-day release period at the seedling stage, combined with basal urea, offers an effective strategy to boost yield, quality, and ANUE in coastal saline rice systems. Full article

At a time of climate change, farmers face difficulties in providing food for a growing population. This results in the overuse of water and fertilisers. The aim of the research was to test the possibility of introducing waste sheep wool fibres into a hydrogel to obtain a stable material that could improve water retention and could serve as a fertiliser material matrix. Wool fibres and hydrogel were chosen because of their ability to store water and their degradability. An evaluation of the swelling degree of different alginate-based hydrogel matrices was performed to select the matrix. The stability and water bonding of hydrogels with different wool fibre content were analysed and evaluated by thermogravimetric analysis (TGA) and differential scanning calorimetry (DSC). The microstructure and the effect of fibres on the uniformity of the hydrogel were assessed using SEM and optical microscopy. The degree of water retention in the soil was also evaluated. The results showed that it is possible to incorporate wool fibres into the hydrogel matrix and the wool fibres make the composite porous, which allows water penetration into the material much more easily. This research has shown the possibility of using waste wool fibres as an active ingredient in sustainable fertiliser materials. Full article

The construction of marine ranching is a crucial component of China’s Blue Granary strategy, yet the fragmented knowledge system in marine ranching equipment impedes intelligent management and operational efficiency. This study proposes the first knowledge graph (KG) framework tailored for marine ranching equipment, integrating hybrid ontology design, joint entity–relation extraction, and graph-based knowledge storage: (1) The limitations in existing KG are obtained through targeted questionnaires for diverse users and employees; (2) A domain ontology was constructed through a combination of the top-down and the bottom-up approach, defining seven key concepts and eight semantic relationships; (3) Semi-structured data from enterprises and standards, combined with unstructured data from the literature were systematically collected, cleaned via Scrapy and regular expression, and standardized into JSON format, forming a domain-specific corpus of 1456 annotated sentences; (4) A novel BERT-BiGRU-CRF model was developed, leveraging contextual embeddings from BERT, parameter-efficient sequence modeling via BiGRU (Bidirectional Gated Recurrent Unit), and label dependency optimization using CRF (Conditional Random Field). The TE + SE + R_i + BMESO tagging strategy was introduced to address multi-relation extraction challenges by linking theme entities to secondary entities; (5) The Neo4j-based KG encapsulated 2153 nodes and 3872 edges, enabling scalable visualization and dynamic updates. Experimental results demonstrated superior performance over BiLSTM-CRF and BERT-BiLSTM-CRF, achieving 86.58% precision, 77.82% recall, and 81.97% F1 score. This study not only proposes the first structured KG framework for marine ranching equipment but also offers a transferable methodology for vertical domain knowledge extraction. Full article

Background: Local audits of Helicobacter pylori (H. pylori) prescriptions and outcomes are necessary to assess guideline awareness among clinicians and treatment effectiveness. Aims: The aims were to investigate first-line prescriptions and effectiveness over a 10-year period in Ireland and evaluate the influence of the 2017 Irish consensus guidelines on these trends. Methods: Data were collected at e-CRF AEG-REDCap from the European Registry on H. pylori management (Hp-EuReg) and quality reviewed from 2013 to 2022. All treatment-naïve cases were assessed for effectiveness by modified intention-to-treat (mITT) analysis. Multivariate analysis was also performed. Results: Data from 1000 patients (mean age 50 ± 15 years; 54% female) were analyzed. Clarithromycin (C) and amoxicillin (A) triple therapy represented 88% of treatments, followed by sequential C, A, and metronidazole (M) therapy (4.3%) and triple C + M (2.7%). Bismuth quadruple therapy was prescribed in 1.7% of cases. Treatment durations of 14, 10, and 7 days accounted for 87%, 4.5%, and 8.5% of prescriptions, respectively. High-, standard-, and low-dose proton pump inhibitors (PPIs; 80 mg, 40 mg, and 20 mg omeprazole equivalent b.i.d.) were used in 86%, 0.9%, and 13% of cases, respectively. The overall eradication rate was 80%, while it was 81% for triple C + A. Good compliance and high-dose PPI were associated with higher overall mITT eradication rates (OR 4.5 and OR 1.9, respectively) and triple C + A eradication rates (OR 4.2 and OR 1.9, respectively). Overall eradication rates increased from 74% pre-2017 to 82% (p < 0.05) by the end of 2022. Similarly, the triple C + A eradication rates increased from 76% to 83% (p < 0.05). Conclusions: While first-line treatment effectiveness improved in clinical practice over time, cure rates remain below 90%. Alternative first-line strategies are required in Ireland. Full article

Background and Objectives: Oxygen consumption (VO₂) recovery plays a critical role in reestablishing homeostasis within multiple physiological processes. This study aimed to assess the differences in the fitness profiles, fatigability, patterns of VO₂ recovery, and sleep quality among individuals with different body fat percentages. Thus, we evaluated the predictive effects of body fat percentage, CRF, fatigability, and sleep quality on VO₂ recovery patterns following exercise. Materials and Methods: Eighty healthy participants aged 18–52 years were included in this case-control study. The participants were divided into two groups based on body fat percentage: normal-fat (CON; n = 40) and high-fat (HFG; n = 40) groups. The PSQI questionnaire was used to assess sleep efficiency, and a 10 min walk test was performed to assess fatigability. Both groups underwent a symptom-limited treadmill exercise test to assess VO₂ using a modified bulk protocol, followed by 6 min of passive recovery. Results: The participants in the CON group had a higher mean VO₂ peak than those in the high-fat-percentage group (p = 0.0003). The half-time recovery (T1 and T2) demonstrated higher amounts of VO₂ in the CON group compared to the HFG group (p = 0.0007 and p = 0.0005), respectively. Those in the HFG reported greater performance fatigability (p = 0.01) and poorer sleep quality compared to the CON group (p < 0.001). The multiple linear regression model indicated that a higher recovery amount of VO₂ was associated with the fat percentage, VO₂ peak, and fatigability index and explained 72% of the variance (F = 39.58, p < 0.001). Conclusions: The findings of this study revealed that the participants with higher fat percentages exhibited increased performance fatigability and a reduced peak VO₂ and reported poor sleep quality compared to the normal group. CPF, body fat, and performance fatigability were associated with VO₂ recovery after high-intensity exercise. The interplay between body fat, fatigability, sleep quality, and VO₂ recovery highlights the need for a holistic approach to healthcare. Full article

Background/Objectives: Keratoconus (KC) is a corneal ectatic disease, characterized by the progressive thinning of the cornea, myopia, and astigmatism, which lead to a decrease in visual acuity. Corneal collagen crosslinking (CXL) is an efficient method of stopping the progression of the disease. The objective of this study is to investigate the endothelial and biomechanical properties of the cornea in keratoconus patients, before and after undergoing corneal collagen crosslinking. Methods: A total of 66 eyes were diagnosed with progressive keratoconus and were recommended epi-off corneal crosslinking. Before the procedure, they were investigated with corneal topography (for minimum, maximum, average keratometry, and corneal astigmatism), specular microscopy (for the following endothelial cell parameters: number, density, surface, variability, and hexagonality), and an ocular response analyzer (for the following biomechanical parameters: corneal hysteresis and resistance factor). All measurements were repeated 1 month and 6 months after the intervention. Results: Several parameters differ according to the Amsler–Krumeich stage of keratoconus: in more advanced stages, patients present higher endothelial cell variability, a lower number of endothelial cells in the paracentral region of the cornea, lower CCT and CRF, and higher keratometry and astigmatism. Endothelial cell variability and number correlate with average keratometry, and there are also strong correlations between topography and CH and CRF. After CXL, the paracentral number of endothelial cells decreased; cell variability and average cell surface increased. Conclusions: More advanced keratoconus cases present with altered corneal biomechanics and topographical parameters, the endothelial layer also being affected proportional to the stage of the disease and also slightly affected after corneal collagen crosslinking. Full article

In order to enhance the management and application of fault knowledge within intelligent production lines, thereby increasing the efficiency of fault diagnosis and ensuring the stable and reliable operation of these systems, we propose a fault diagnosis methodology that leverages knowledge graphs. First, we designed an ontology model for fault knowledge by integrating textual features from various components of the production line with expert insights. Second, we employed the ALBERT–BiLSTM–Attention–CRF model to achieve named entity and relationship recognition for faults in intelligent production lines. The introduction of the ALBERT model resulted in a 7.3% improvement in the F₁ score compared to the BiLSTM–CRF model. Additionally, incorporating the attention mechanism in relationship extraction led to a 7.37% increase in the F₁ score. Finally, we utilized the Neo4j graph database to facilitate the storage and visualization of fault knowledge, validating the effectiveness of our proposed method through a case study on fault diagnosis in CNC machining centers. The research findings indicate that this method excels in recognizing textual entities and relationships related to faults in intelligent production lines, effectively leveraging prior knowledge of faults across various components and elucidating their causes. This approach provides maintenance personnel with an intuitive tool for fault diagnosis and decision support, thereby enhancing diagnostic accuracy and efficiency. Full article