搜索结果：intelligence

Technology has offered considerable means to create, enhance, and upkeep biosecurity during the last quarter-century. Mostly this refers to routine health promotion means, more or less modified for the much more complex environment of Biosecurity, which includes perpetrated biothreats and thus involves hostile intelligence set to nullify response. This narrative review describes the complexity of the field and presents to bioscience-cognitive audience an update of the current state-of-the-art approaches, their usefulness in Biosecurity applications and their suitability for the diverse current format. The research involved plain internet (Google) search engine and Pubmed searches with matching terms from May 2025 to January 2026. Regarding diagnostics, the Molecular Diagnostics applicable in Resource-Limited Settings, the serodiagnostics, now available in solid formats as Rapid Diagnostic Tests, improved and more powerful Microscopy and cultures updated to Culturomics are the focus. Support functions, from bioinformatics to remote and cybernetic applications, such as lab and diagnostic automations, increase responsiveness and flexibility. Intervention, the second major arm of Biosecurity is also evolving despite the scarcity of new antibiotics. Different classes of Bioamenities combined with advanced and controllable delivery methods emerge for routine medicine and may be adopted, adapted, and modified for Biosecurity use.

This study aimed to develop and evaluate a deep learning-based surgical navigation system capable of recognizing the ureter, uterine artery, and bladder-uterine dissection plane during minimally invasive gynecologic surgery. An artificial intelligence (AI) model was developed at the University of Tokyo Hospital using videos of prior surgeries. Surgical videos of 27 laparoscopic or robot-assisted total hysterectomies were used to create training and validation datasets, with an additional set of cases serving as an independent test set. Key frames were manually annotated to train segmentation models for the ureter and uterine artery. A separate model visualized loose connective tissue fibers (LCTF) to aid in recognizing the bladder-uterine peritoneal dissection plane. Quantitative performance was assessed using standard segmentation metrics, and a qualitative evaluation was conducted by nine gynecologic surgeons using predefined scoring criteria. The segmentation models achieved moderate quantitative performance, with Dice similarity coefficients of approximately 0.51 for the ureter and 0.45 for the uterine artery. In contrast, qualitative evaluation demonstrated favorable clinical interpretability. The mean recognition scores assigned by nine expert surgeons were 4.12 for the ureter and 3.45 for the uterine artery on a five-point scale, indicating that most structures were recognized clearly with only minor misrecognition. For bladder dissection, visualization of connective tissue fibers enabled identification of the correct dissection plane in the majority of evaluated frames; more than 70-80% of connective tissue was recognizable in most frames, and substantial misrecognition was uncommon. This study demonstrates that a deep learning-based system can recognize three key elements of a total hysterectomy: the ureter, the uterine artery, and the bladder-uterine dissection plane. Despite modest quantitative metrics, qualitative assessments indicated strong clinical utility. These findings establish a foundation for an integrated AI-assisted surgical navigation platform to enhance the safety and standardization of minimally invasive gynecologic surgery.

The reliable deployment of artificial intelligence systems in medical imaging requires high diagnostic performance, robustness and interpretability. In this study, we developed and evaluated two automated frameworks for binary classification of shoulder radiographs (XRs) using deep learning (DL) and hybrid DL-machine learning (ML) approaches. A convolutional neural network (CNN) based on a fine-tuned VGG19 architecture was trained end-to-end on a large, balanced dataset of 4,268 shoulder XRs. In parallel, hybrid models were constructed by extracting deep feature representations from the trained network and combining them with traditional ML classifiers. Model performance was evaluated on independent internal (n = 480) and external (n = 308) validation sets. Both approaches achieved high discriminative performance. Paired comparison of Receiver Operating Characteristic (ROC) curves using the DeLong test revealed no statistically significant differences between the end-to-end CNN and the hybrid CNN-ML pipeline for either internal validation (AUC 0.956 vs. 0.961) or external generalization (AUC 0.940 vs. 0.942). Model interpretability was assessed using Grad-CAM and SHAP values. Our results suggest that while both frameworks are robust, the end-to-end DL approach offers a more streamlined workflow and more direct visual explainability via saliency maps. These findings support the potential of AI-based tools for shoulder XR analysis; however, prospective real-world validation, assessment under routine prevalence conditions, and direct comparison with human readers are still needed before clinical integration can be established.

Artificial intelligence-driven image analysis has enabled significant advances in digital pathology. However, most approaches have focused on cell or organ structures. This manuscript presents a reproducible deep learning methodology for pixel-level analysis of amorphic patterns in haematoxylin and eosin-stained whole-slide histological images. This study analysed the pixel patterns in the extracellular matrix (ECM) part of connective tissue to identify differences in airway wall ECM compartments and their heterogeneity, which are microscopically similar and difficult to discern with the human eye. Through a targeted preprocessing pipeline, the deep learning model is guided to emphasise learning from pixel-level patterns in non-cellular tissue components while reducing the influence of cellular structures and artefacts. Combined with transfer learning, the model accurately distinguishes the characteristics of the airway submucosa and adventitia, achieving a test area under the curve of 0.84. Using visualisation techniques and statistical analysis, we demonstrate that random pixel imputation successfully reduces the effects of cellular structures on model learning. The framework is applied in a proof-of-principle study of lung tissue from patients with chronic obstructive pulmonary disease, illustrating how this quantitative approach can study population heterogeneity and inform novel research directions. Ultimately, this study provides an innovative and adaptable framework that unlocks the analytical potential of often-overlooked amorphic components in AI-empowered histopathology.

Chronic atrophic gastritis is a well-recognized precancerous condition, emphasizing the need for accurate endoscopic classification. However, existing endoscopic datasets lack comprehensive multi-dimensional annotations for systematic gastritis classification. We present the Endoscopic Gastritis Image Dataset (EGID), comprising 5,883 high-quality white-light endoscopic images from 229 patients retrospectively collected at Renji Hospital, Shanghai, using the Olympus CV-290 system between March and November 2024. Each patient's images are systematically annotated across four independent clinical dimensions: H. pylori infection status, presence of gastric atrophy, distribution of atrophy, and gastritis type. H. pylori infection status was determined based on rapid urease test results, whereas the three image-based endoscopic dimensions were independently assessed by two experienced endoscopists, with consensus adjudication by a senior gastroenterologist for discordant cases. The image-based annotation dimensions showed excellent inter-rater agreement (Cohen's Kappa: 0.947-0.970). EGID provides the first publicly available multi-label endoscopic gastritis dataset, enabling development of artificial intelligence (AI)-based classification systems and serving as an educational resource for clinical training in gastritis diagnosis and cancer risk stratification.

The spread of generative artificial intelligence and large language model technologies, such as ChatGPT, has sparked interest in their applicability and potential role in reproductive health counseling. This qualitative study explored the perspectives of professional counselors providing pregnancy termination counseling in Germany on the integration of ChatGPT into their work. Between November 2024 and January 2025, 20 semi-structured interviews were conducted with counselors working at state-accredited counseling centers, using a case vignette design to explore the potentials, challenges, meanings, needs for support, and requirements involved in case evaluation while sparring with ChatGPT-4o Mini.Thematic analysis revealed four main themes. Participants expressed persistent skepticism, curiosity, and encouraging first experiences regarding the reliability, contextual appropriateness, ethical alignment, and legal accuracy of ChatGPT-generated content. Counselors emphasized that interpersonal relatedness is a crucial marker of quality and meaning of counseling, encompassing empathy, subjective competence, and situational sensitivity. Reflections on professional roles revealed that ChatGPT was perceived as a primarily supportive tool for non-relational tasks. ChatGPT's potential was described as significantly constrained by specific needs, fantasized relief, and working conditions marked by structural limitations and individual barriers to digital innovation, such as the centers' equipment, digital readiness, and privacy policies.We discuss the findings in relation to AI and technology acceptance models and the theory of professional practice, contributing to a refinement of the concept of conceptualized skepticism toward a more nuanced understanding that may be specific to counseling contexts. The findings underscore positions that argue counseling encompasses more than its methods, relying on in-betweens and subjectivities.

The natural antifungal peptide Histatin 5 (Hst 5) is a histidine-rich cationic peptide secreted by human salivary glands and a key component of oral innate immunity, but its moderate activity limits clinical use. Hst 5 enters Candida albicans via the membrane receptor Ssa1/2. Here, we integrated artificial intelligence-assisted and computer-aided drug design to rationally modified the sequence structure of Hst 5. Truncated derivatives of Hst5 were screened for antimicrobial potential using ESM2-AFPpred, and high-probability candidates were docked with Ssa1/2. The Hst 5-22 was identified, then redesigned based on alanine scanning to yield the optimized derivative Hst 5-22-RW. Compared with Hst 5, Hst 5-22-RW has a shorter sequence, stronger Ssa1/2 binding, and improved activity against C. albicans. It also shows superior activity against fluconazole-resistant strains. RT-qPCR and transmembrane tracking confirmed higher cellular transport efficiency in C. albicans. The CADD/AIDD-driven optimization successfully generated the highly active antifungal peptide Hst 5-22-RW, providing a novel strategy for rational modification of antimicrobial peptides.

To address the dilemma of homogeneous talent training and the efficiency bottleneck of human resource management in universities, this study proposes an innovative personalized training framework integrating artificial intelligence, big data, and deep learning. Based on the 18-dimensional full-cycle behavior dataset of 5,000 students and OULAD dataset, a multimodal heterogeneous data fusion pipeline is constructed. This study adopts Generative Adversarial Network (GAN) for data imputation and bias optimization, designs Hierarchical Attention Graph Neural Network (HA-GNN) to capture hierarchical correlations among features, and uses Long Short-Term Memory (LSTM) to model temporal behavior patterns. The experimental results demonstrate that, under 10 independent repeated runs with random seed variation, the Hierarchical Attention Graph Neural Network-Long Short-Term Memory (HA-GNN-LSTM) model achieves lower prediction error on the academic performance prediction task, with a Mean Absolute Error (MAE) of 4.2 ± 0.3. Compared with the Temporal Fusion Transformer (TFT) baseline model, MAE is reduced by 31.1%. Welch's two-tailed t-tests based on independent run results remain statistically significant after Holm-Bonferroni multiple comparison correction [Formula: see text]. The Normalized Discontinued Cumulative Gain at Top 5 (NDCG @ 5) index of personalized recommendation system reaches 0.90, which verifies the effectiveness of spatio-temporal feature modeling. At the management application level, the improvements in advisor allocation response time and resource idle rate are derived from simulation experiments based on historical data replay, rather than online deployment in real campus management systems. The simulation results demonstrate that, under established constraints and historical sample distributions, advisor allocation response time could be reduced by 60% and resource idle rate could be decreased by 63.4%. These findings indicate the framework's potential for optimizing educational resource allocation. However, its managerial benefits require further validation through subsequent real-world deployment and long-term follow-up studies.

Intracerebral hemorrhage (ICH) remains associated with high mortality and treatment variability. Current workflows rely on fragmented imaging interpretation and operator-dependent surgical planning. The objective was to develop and validate an agentic artificial intelligence (AI) framework integrating automated imaging analysis, guideline-based reasoning, and trajectory optimization for ICH treatment. Fifty consecutive computed tomography (CT) and computed tomography angiography (CTA) datasets from patients with spontaneous ICH were retrospectively analyzed. The system performed multi-class anatomical segmentation of skin, skull, brain, ventricles, and hematoma, followed by volumetric quantification and JavaScript Object Notation (JSON) based structured encoding of imaging biomarkers. A knowledge-based module incorporating international ICH guidelines generated risk stratification and treatment recommendations. When evacuation was indicated, an automated trajectory modeling module proposed a patient-specific minimally invasive surgical corridor. Overall agreement between AI-generated and expert treatment recommendations was 82% (41/50 cases), with substantial agreement beyond chance (Cohen's κ = 0.71). Discrepancies occurred primarily in borderline surgical indication scenarios. In evacuation candidates, the automated planner generated feasible trajectories in all 50 cases. Median angular deviation between AI-generated and expert-defined trajectories was 7.6°, interquartile range (IQR) 5.1-9.8°. AI-generated trajectories demonstrated equal or greater safety margins relative to expert planning in the majority of cases. End-to-end processing has a potential to substantially reduce simulated decision-support time compared with manual workflow. The proposed agentic AI framework enables structured, explainable, and workflow-integrated decision support for ICH management. This system may reduce operator variability and enhance precision in minimally invasive evacuation planning.

The prevalence of brain tissue hypoxia (BTH; PbtO2&#x2009;<&#x2009;20&#xa0;mm Hg) in patients with spontaneous ICH is not well established. In this study, we aimed to quantify the prevalence of BTH and to assess determinants of brain tissue normoxia (BTN; PbtO2&#x2009;&#x2265;&#x2009;20&#xa0;mm Hg) and BTH resolution. This retrospective cohort study included 58 patients with ICH admitted to a neurological intensive care unit (ICU) between 2010 and 2020 with multimodal invasive neuromonitoring. BTN was sought by avoiding low cerebral perfusion pressure (CPP) and low blood hemoglobin levels, and by maintaining normocapnia, normoxemia, normothermia, and metabolic homeostasis. Hourly PbtO2, CPP, and temperature data were matched with intermittent variables (blood gases, hemoglobin, glucose, sodium, and microdialysis) over 10&#xa0;days. Regression analyses were performed&#xa0;using generalized estimating equations to account for repeated measurements. Patients were 61 [interquartile range (IQR), 55-69] years old and presented with an ICH score of 2 (1-3). Of the patients, 52 (90%) underwent surgical evacuation via hemicraniectomy and/or craniotomy, while 6 (10%) received invasive neuromonitoring only. The median initial ICH volume was 40.2 (IQR 29.5-55.8) mL. Surgical evacuation achieved a median reduction of 86.6% (IQR 69.0-94.2), leaving a median residual volume of 5.5 (IQR 3.0-14.1) mL. The overall prevalence of BTH was 31%. In multivariable analysis, the following factors led to the highest percentage of BTN: CPP 80-89&#xa0;mm Hg [odds ratio (OR) 1.88, 95% confidence interval (CI) 1.32-2.68, p&#x2009;<&#x2009;0.001; reference:&#x2009;<&#x2009;60&#xa0;mm Hg], partial pressure of oxygen (PaO2) 90-99&#xa0;mm Hg (OR 1.64, 95% CI 1.15-2.14, p&#x2009;=&#x2009;0.001; reference:&#x2009;<&#x2009;80&#xa0;mm Hg), core body temperature 36.0-37.4&#xa0;&#xb0;C (OR 2.10, 95% CI 1.34-3.28, p&#x2009;=&#x2009;0.001; reference:&#x2009;<&#x2009;36.0&#xa0;&#xb0;C), PaO2/fraction of inspired oxygen (FIO2) 100-199 (OR 3.52, 95% CI 1.60-7.75, p&#x2009;=&#x2009;0.002; reference:&#x2009;<&#x2009;100) in a model corrected for probe position. Only CPP [74.6 (66.8-82.7) vs. 72.5 (64.9-80.2) mm Hg, p&#x2009;<&#x2009;0.001)] was significantly higher after BTH resolution as compared with the time when PbtO2 was lowest during 229 BTH episodes. BTH was observed during 31% of the monitored time in patients with a large hematoma volume despite the use of a PbtO2-targeted therapy. These findings generate the hypothesis that physiological determinants, such as CPP, are significantly associated with the achievement of BTN in regions near to the ICH.

Predicting foundation pit deformation is a significant challenge for foundation pit engineering. The inaccuracy of deformation prediction is increased by the intricacy of subterranean space and the variety of construction conditions. Thus, this study develops a deformation prediction model that combines the attention mechanism and bidirectional long short term memory network (BiLSTM) in order to increase the accuracy of deep foundation pit deformation prediction. Meanwhile, to enhance the generalization ability of the model, this study introduces combined regularization in the loss function and adds a Dropout mechanism in the network structure. This study takes a deep foundation pit excavation project in Guangzhou as an example. Experiments shows that the model proposed in the study can complete convergence in about 30 training rounds, and the training loss is maintained at the 0.03 level. Meanwhile, the maximum absolute error of the model in the prediction of verification data is 1.44&#xa0;mm, and the minimum error is 0.001&#xa0;mm. The mean absolute error of the model is 0.311&#xa0;mm, the root mean square error is 0.433&#xa0;mm, and the R2 is 0.906, which is better than the comparison model. The attention mechanism and BiLSTM model suggested in this study provides good generalization performance and high prediction accuracy in deep foundation pit deformation prediction, according to experimental results. Its potential for use in engineering safety management is promising.

暂无摘要（点击查看详情）

Chronic ocular graft-versus-host disease (coGVHD) after allogeneic hematopoietic stem cell transplantation (allo-HSCT) may lead to irreversible ocular surface damage and even vision loss. Current management of coGVHD faces challenges, with frequent missed or misdiagnosed cases. This study aimed to leverage a multimodal large language model (MLLM) to develop an early warning and diagnostic system for coGVHD. A total of 666 post-allo-HSCT patients (early warning model) and 805 post-allo-HSCT patients (1574 eyes, diagnostic model) were enrolled for construction, internal validation, and external validation of the corresponding models. We proposed the GVHD-MLLM, a multitask multimodal network that fused latent representations from four modal sequences to provide high-precision, real-time predictions for two tasks. The GVHD-MLLM achieved high performance in internal testing, with AUROCs of 93.44% (95% CI: 91.85-95.03%) for early warning, 98.98% (95% CI: 98.59-99.36%) for diagnosis, and 98.24% (95% CI: 98.05-98.43%) for disease severity grading. In external validation, the early warning AUROC was 83.45%, while diagnostic AUROCs across three external sites were all above 96.0%. The disease severity of patients seeking medical treatment after using the early warning model was significantly lower. Junior ophthalmologists also improved diagnostic accuracy using the model as an auxiliary tool. The GVHD-MLLM can process rich multi-modal information collected in clinical practice, and is expected to become an effective tool for managing coGVHD.

Object detection in soccer videos plays an important role in intelligent broadcasting, tactical analysis, and player tracking. Frequent view switching driven by multi-camera coordination in soccer broadcasts introduces two distinct challenges for existing detection methods. First, in the temporal dimension, shot transitions cause abrupt visual changes between adjacent frames. Existing temporal aggregation methods forcibly fuse features from heterogeneous views, injecting cross-view noise that can degrade detection accuracy below single-frame baselines. Second, in the feature dimension, target appearance varies significantly across views: players in panoramic shots are low-resolution small objects characterized mainly by coarse contours and color, whereas the same players in close-up shots become high-resolution large objects with rich texture and fine-grained detail. A unified detection strategy struggles to handle both extremes simultaneously. We propose ViewAdapt-Det, a framework consisting of two complementary modules. The View-Aware Temporal Gating module (VATG) detects shot transitions and dynamically controls temporal aggregation intensity, suppressing historical frame contributions during view changes while preserving temporal enhancement during continuous sequences. The View-Conditioned Detection Modulation module (VCDM) infers the current view type and conditionally modulates detection features, allowing the detector to apply view-specific feature processing strategies that match the visual characteristics of each viewpoint. Experiments on SoccerNet-Tracking, SportsMOT, and our self-constructed BroadcastSwitch-Soccer dataset show that ViewAdapt-Det outperforms existing methods and demonstrates stronger robustness under abrupt shot transitions.

Acute inflammation, when unresolved, can lead to complications that impair tissue repair and therapeutic outcomes. In this study, we employed a model of lipopolysaccharide (LPS)-induced acute subcutaneous abdominal inflammation in mice to investigate the modulatory effects of elastic compression. LPS administration elicited a robust inflammatory response, characterized by increased leukocyte infiltration, edema, and upregulation of pro-inflammatory mediators. Elastic compression significantly attenuated this response, reducing leukocyte counts in subcutaneous lavage, histological inflammatory infiltrates, and the expression of key pro-inflammatory genes and proteins, including NF-&#x3ba;B, IL-1&#x3b2;, and TNF-&#x3b1;, at both 24 and 72 hours post-induction. Mechanistically, these effects may result from the compressive force altering microvascular dynamics and modulating macrophage polarization and mechanotransduction pathways, including TLR4 and integrin signaling. Additionally, compression preserved redox homeostasis, as indicated by stable oxidative stress markers and antioxidant responses. To our knowledge, this is the first study to demonstrate that elastic compression modulates inflammation at molecular, cellular, and tissue levels in an acute inflammation model. These findings support the therapeutic potential of elastic compression as a non-pharmacological strategy for managing acute inflammation, with possible applications in postoperative care, traumatic edema, and other soft tissue inflammatory conditions. Further translational and clinical studies are warranted to validate these outcomes and guide evidence-based application protocols. This journal requires that authors assign a level of evidence to each submission to which Evidence-Based Medicine rankings are applicable. This excludes Review Articles, Book Reviews, and manuscripts that concern Basic Science, Animal Studies, Cadaver Studies, and Experimental Studies. For a full description of these Evidence-Based Medicine ratings, please refer to the Table of Contents or the online Instructions to Authors www.springer.com/00266 .