Patient: Female, 29-year-oldFinal Diagnosis: SilicosisSymptoms: DeformityClinical Procedure: —Specialty: Plastic SurgeryObjective: Unknown etiologyBackgroundGluteal augmentation has gained significant popularity worldwide, with a parallel increase in complications related to unregulated procedures. The use of permanent fillers by non-medical personnel poses serious risks and can complicate future surgical interventions.Case ReportWe report the case of a 29-year-old woman seeking revision of a prior gluteal augmentation performed abroad with subcutaneous silicone implants. The patient desired greater upper-pole projection and improved contour. During liposuction of the lower back, an unexpected gel-like material was encountered throughout the subcutaneous tissue of the lower back, hips, and infra-gluteal fold, consistent with previously injected permanent filler. The procedure was immediately adapted: fat grafting was abandoned due to the risk of infection and poor graft viability, and thorough manual extraction and irrigation were performed. Implant exchange was completed successfully, with new biconvex silicone implants placed in intramuscular pockets. The postoperative course was uneventful, and at 3-month follow-up, the patient demonstrated improved contour and was satisfied with the outcome.ConclusionsThis case illustrates the need for accurate preoperative evaluation and the ability to adapt intraoperatively in response to unsafe prior procedures. It also highlights the growing concern over illegal filler use and reinforces the importance of regulation, patient education, and adherence to evidence-based surgical practices to ensure safety and satisfactory outcomes.

Automated strain-to-peptide conversion: a high-throughput proteome analysis platform empowering rational design of microbial cell factories.

exBAClock: A comprehensive database of published clocks for age quantification and age-related diseases.

The process of developing and updating an evidence gap map (EGM) is based on the principles of systematic reviews and requires extensive time and financial resources. Artificial intelligence (AI) tools, like prioritisation screening (PS), integrated into programmes such as EPPI-Reviewer (ER) and Copilot 365, can potentially mimic human performance in systematic review processes. ER is a subscription-based web application employed by systematic review groups, while Copilot 365, integrated into Microsoft 365, offers real-time assistance. Although ER shows promise in speeding up screening, the optimal threshold for accuracy remains unclear. Additionally, there is no evidence on the effectiveness of any version of Copilot in systematic review and EGM processes. Assess the accuracy and efficiency of Copilot 365 and PS integrated into ER at different stages of an EGM update, comparing it to human performance. We will conduct both manual and automated screening of references, full-text screening, data extraction, and critical appraisal. Two reviewers will independently screen studies for inclusion, extract data, and appraise included studies, resolving conflicts through discussion. We will assess the accuracy and efficiency of Copilot 365 and ER at different EGM update stages, comparing them to human performance. To evaluate the PS accuracy, we will use 20% and 40% manual screening thresholds, calculating the proportion of relevant references prioritised by PS and the total relevant citations missed. We will compare Copilot 365's full-text screening accuracy to reviewers' decisions and assess consistency using Cohen's Kappa. For automated data extraction and appraisal, we will manually inspect 20% of Copilot 365's outputs, comparing them to reviewers' results, measuring consistency with Cohen's Kappa, and evaluating time savings by comparing the time taken for manual extraction versus using Copilot 365. This study will offer insights into ER's accuracy in screening small samples of citations and potentially guide future applications in this context. Additionally, by evaluating Copilot 365, which shares similar features with other AI tools, we will gain a broader understanding of its applicability and limitations in evidence synthesis, making the results relevant to other AI applications in this field. Registered at Open Science Framework: https://doi.org/10.17605/OSF.IO/49BX8.

As semiconductor technology node advances, the number of parameters in the modern device compact model increases drastically. Manual extraction of these model parameters becomes not only tedious but also impossible, and the automatic method is strongly desired. Traditional black-box optimization suffers from poor scalability due to the curse of dimensionality, while deep learning–based methods typically require large amounts of training data. To address these challenges, we propose CDRPE: a combined deep learning and self-attention enhanced reinforcement learning framework for automatically extracting a large set of DCM parameters across multiple electrical characteristics. The framework leverages a pre-trained multilayer perceptron to initialize core parameters, incorporates device physics knowledge to guide the search, and employs a self-attention–enhanced RL agent for efficient exploration in high-dimensional parameter spaces. Experimental results on BSIM4, BSIMSOI, and BSIMCMG demonstrate that CDRPE can automatically extract 100 parameters with root-mean-square error below 5% relative to TCAD and silicon data. Compared with existing methods, the proposed framework achieves a 7.7X speed up. Moreover, the generated models show good convergence in both digital and analog circuit simulations, exhibiting the potential of this framework for future practical applications.

Patterns of social participation among older adults with chronic multimorbidity in the community: A qualitative study.

Abstract The low-frequency characteristics of infrasound signals enable them to play a critical role in the long-range detection of disasters such as earthquakes, landslides, and chemical explosions. However, the accurate classification of disaster types based on infrasound signals remains a significant challenge. Traditional methods rely on manual feature extraction, often failing to capture spatiotemporal patterns. While deep learning approaches, particularly convolutional neural networks (CNNs), show promise, they are limited by finite depth, inadequate temporal modeling capabilities, and network degradation. Additionally, a single CNN struggles to learn diverse feature representations and exhibits low fault tolerance to anomalies. To address these challenges, a machine learning-driven classification method is proposed featuring three synergistic components: a synergistic feature-deep learning architecture that integrates continuous wavelet-scale average coefficients-based time-scale transformation and deep feature learning to enhance discriminative capability; a triple-branch architecture that integrates multi-scale spatial convolutions and adaptive temporal regulation to model propagation dynamics; and a parallel confidence fusion module that dynamically weights branch outputs for final robust decisions. Experimental results show that the proposed method achieves 98.75% accuracy on a natural earthquake infrasound dataset and maintains a high accuracy on an open-source multi-class dataset, demonstrating strong robustness and generalization capability for natural disaster monitoring.

Cat breed classification is a significant challenge in the field of computer vision due to the high visual similarity between breeds (fine-grained classification) and pattern variations within a single breed. This study aims to compare the performance of two different machine learning approaches, namely Random Forest (RF) based on manual features and Convolutional Neural Network (CNN) based on automatic features. The research focuses on three cat breeds: Bombay, Siamese, and Persian. The research methodology uses a public dataset from Kaggle, divided in a ratio of 80:10:10. The RF pathway applies manual feature extraction through a combination of Histogram of Oriented Gradients (HOG) and Color Histogram. In contrast, the CNN pathway uses Transfer Learning techniques with the ResNet50V2 architecture. The test results show that CNN significantly outperforms RF with an accuracy of 93.33%, while RF only reaches 68.33%. The analysis shows that manual features in RF have difficulty capturing complex texture details in the Persian breed, while CNN is able to generalize well. It is concluded that the Deep Learning (CNN) approach is much more effective than traditional methods for animal breed classification.

The ability to continuously make inferences about another person's latent states from their behavior is integral to how people behave in social situations, yet is lacking from most artificial intelligence (AI) systems. The present study tests the capacity of cognitive models to assess latent motives in real time by evaluating different deep neural networks trained to infer a human player's intent during a continuous control task. These networks were trained by (a) directly using observable information or (b) selecting important features by estimating the parameters of a generative model of movement behavior inspired by approach-avoidance theory. Comparisons of classifier accuracy suggest that latent model parameters predict a participant's intent at a level exceeding human performance. Furthermore, classifier performance was best when model-based inferences were combined with summary statistics about behavior, yielding faster and more stable network training compared to networks that had no manual feature extraction. Equipping AI with cognitive models is a promising avenue for developing explainable, accurate, and trustworthy systems.

With the large-scale integration of power electronic converters, non-linear loads, and renewable energy generation, voltage and current waveform distortion in modern power systems has become increasingly severe, making harmonic resonance amplification and non-stationary distortion more prominent. Accurate and robust harmonic-level prediction and detection have become essential foundations for power quality monitoring and operational protection. However, traditional harmonic analysis methods remain highly dependent on pre-designed time–frequency transformations and manual feature extraction. They are sensitive to noise interference and operational variations, often exhibiting performance degradation under complex operating conditions. To address these challenges, a Unified Physics-Transformer-based harmonic detection scheme is proposed to accurately forecast harmonic levels in offshore wind farms (OWFs). This framework utilizes real-world wind speed data from Bozcaada, Turkey, to drive a high-fidelity electromagnetic transient simulation, constructing a benchmark dataset without reliance on generative data expansion. The proposed model features a Feature Tokenizer to project continuous physical quantities (e.g., wind speed, active power) into high-dimensional latent spaces and employs a Multi-Head Self-Attention mechanism to explicitly capture the complex, non-linear couplings between meteorological inputs and electrical states. Crucially, a Multi-Task Learning (MTL) strategy is implemented to simultaneously regress the Total Harmonic Distortion (THD) and the characteristic 5th Harmonic (H5), effectively leveraging shared representations to improve generalization. Comparative experiments with Random Forest, LSTM, and GRU systematically evaluate the predictive performance using metrics such as root mean square error (RMSE) and mean absolute percentage error (MAPE). Results demonstrate that the Physics-Transformer significantly outperforms baseline methods in prediction accuracy, robustness to operational variations, and the ability to capture transient resonance events. This study provides a data-efficient, high-precision approach for harmonic forecasting, offering valuable insights for future renewable grid integration and stability analysis.

Background: To provide more evidence in urolithiasis research, we have established the German Nationwide Register for RECurrent URolithiasis (RECUR) using local clinical data warehouses (CDWH). For RECUR and other registers relying on digitalized clinical data, it is crucial to ensure the data’s reliability for answering scientific questions. In this work, we aim to compare the results of different CDWH-based queries on urolithiasis cases next to manual case extraction from the primary source.Methods: Sources for data extraction included the Medical Center University of Freiburg (MCUF) hospital information system (HIS), MCUF performance data (a clinical data set with merged data from patients including data from various time points throughout their treatment), and MCUF reimbursement data. We extracted data on caseloads in urolithiasis algorithmically (performance and reimbursement data) and compared those to a reference group compiled of manually extracted data from the local HIS and algorithmically extracted data.Results: Algorithmic extraction based on performance data resulted in correct and complete case identification as compared to the reference group. The case numbers from manual extraction from HIS data and algorithmic extraction from reimbursement data differed by 14% and 12%, respectively. The reasons for deviations in HIS data included human errors and a lack of data availability from different wards. Deviations in reimbursement data arose primarily due to the merging of cases in the context of reimbursement mechanisms. As the CDWH at MCUF is part of the German Medical Informatics Initiative (MII), the results can be transferred to other medical centers with similar CDWH structure.Conclusions: The current study provides firm evidence of the importance of clearly defining a study’s target variable, e.g., urolithiasis cases, and a thorough understanding of the data sources and modes used to extract the target data. Our work clearly shows that, depending on various data sources, a case is not a case is not a case.

Background: High volumes of formalin-fixed paraffin-embedded (FFPE) samples are processed for Mycobacterium tuberculosis (MTB) real-time polymerase chain reaction (RT-PCR) testing, making manual DNA extraction prone to human error and cross-contamination. Automated nucleic acid extraction offers a more efficient alternative. Objective: To compare false positive rates, invalid results, DNA yield, and purity between automated and manual DNA extraction methods for MTB RT-PCR. Materials and Methods: One thousand three hundred eighteen FFPE samples were evaluated by a pathologist for tissue reactions and classified into histologic scores. Scores of 0, 1, 2, and 3 indicated no reaction, non-specific inflammation, non-necrotizing granuloma or caseous necrosis without granuloma, and necrotizing granuloma, respectively. Of these, 767 (58.19%) underwent manual extraction, and 551 (41.81%) underwent automated extraction. RT-PCR was performed to detect MTB, with false positives identified by reviewing PCR-positive samples that did not align with the histological scores. False positivity due to cross-contamination was confirmed if a repeat PCR test, performed on newly extracted DNA, yielded a negative result. DNA yield and purity were compared between the methods using a Mann-Whitney U test. Results: False positive rates were 1.69% for manual extraction and 0.91% for automated extraction, with invalid result rates of 2.09% and 3.27%, respectively. The manual method yielded higher median (IQR) DNA concentration and yield than the automated method at 334.60 (113.00 to 862.20) versus 120.80 (30.40 to 382.60) ng/μL and 10,038.00 (3,390.00 to 25,866.00) versus 6,040.00 (1,520.00 to 19,130.00) ng, respectively. DNA purity was also higher with the manual method with A260/A230: 2.22 (2.12 to 2.27) versus 2.02 (1.39 to 2.23), A260/A280: 1.94 (1.90 to 1.97) versus 1.90 (1.85 to 1.93). All differences were statistically significant (p<0.0001). Conclusion: Automated nucleic acid extraction reduced false positive rates by 0.78% but increased invalid result rates by 1.18%. It yielded lower DNA yield and purity compared to manual extraction. Despite these limitations, automation remains a practical option for high-throughput processing, offering substantial time and resource savings with manageable invalid result rates.

Low-velocity impact damage significantly compromises the structural integrity and residual strength of carbon fiber-reinforced polymer (CFRP) composites. This study proposes a hybrid approach integrating progressive damage finite element modeling with convolutional neural networks (CNNs) to accurately predict compressive strength after impact (CAI). A 3D continuum damage model was developed to characterize interlaminar damage, employing a bilinear traction-separation law combined with the Benzeggagh-Kenane (B-K) criterion for delamination simulation. Finite element results under varying impact energies demonstrated strong agreement with experimental data in terms of force-time and force-displacement responses. A dataset pairing delamination damage profiles with corresponding CAI values was constructed from simulations across different impact scenarios. A deep CNN architecture achieved an Root Mean Square Error (RMSE) of 2.5197 MPa in mapping damage images to residual strength, eliminating the need for manual feature extraction or material parameter dependency. This image-driven method enables high-fidelity strength prediction and shows promising potential for intelligent health monitoring of composite structures.

Continuous monitoring of the cutting tool insert’s condition is essential to enhance product quality and efficient machining process, by reducing the machine downtime. But the available tool condition monitoring approaches are often limited by coolant induced visibility loss in the cutting zone that reduces the feature reliability. This study proposes a transfer learning based deep learning method where the machining vibration signals are converted into visual representations and classified using ResNet 18, MobileNet V2, SqueezeNet, ShuffleNet, DenseNet 201, and EfficientNet B0 pretrained convolutional neural networks. This combination enables the model to learn deep wear profiles from vibration data without the manual feature extraction. Also, this method enhances signal strength, making it highly suitable for smart, scalable, and real world manufacturing environments. The effects of the proposed pretrained network hyperparameters, such as mini batch size, solver type, learning rate, and filter size, were studied and EfficientNet B0 was identified as the best performing network with a classification accuracy of 89.23% for tool condition monitoring tasks.

Accurate estimation of tool wear in machining processes is essential to ensure product quality and optimize maintenance strategies. This work presents a Machine Learning methodology for the PHM-AP 2025 Data Challenge. The objective of the challenge is the cutter flank wear prediction in a CNC mill-turn machine using accelerometer, acoustic emission, and controller data. The training data consists of six datasets with a limited number of labeled samples, resulting in a few-shot learning scenario. To address these constraints, a manual feature extraction method is proposed. Features are computed by aggregating data from the controller and sensors in the time and frequency domains across five-cut intervals. In this way, the wear behavior is captured, and the sensitivity to missing data is reduced. Then, an optimization process is performed to select the most relevant features based on correlation values. These 14 identified features are used to fit a Multilayer Perceptron through a leave-one-dataset-out cross-validation process. Results reveal variability between training sets, with pronounced errors in the 17-21 cutting interval in four datasets. However, in the evaluation stage, the model achieved a competitive performance: RMSE of 11.486, MAPE of 8.518, and R2 of 0.875, placing fourth in the challenge.

BackgroundDilated cardiomyopathy (DCM) represents the nonspecific conclusion to a diverse array of genetic and acquired myocardial insults and is the leading indication for heart transplantation worldwide. Free-text narratives in electronic health records (EHRs) capture real-world variation in patient trajectories and outcomes in detail, critical for improved disease characterisation. However, manual information extraction is resource-intensive and error-prone, highlighting the need for automated patient identification tools.PurposeWe aimed to develop a natural language processing (NLP) pipeline to identify patients with a non-ischaemic, non-valvular dilated cardiomyopathy phenotype from EHRs.MethodsA named entity recognition (NER) and contextualisation pipeline based on word2vec embeddings and bidirectional Long Short-Term Memory architecture was developed using MedCAT, an open-source NLP toolkit. Thirteen clinical concepts defining DCM and its exclusion were selected from the biomedical ontology SNOMED for entity recognition and linking. Two medical professionals annotated 1,200 documents with the selected concepts, adding context with labels of ‘Experiencer’ (Patient, Family, Other), ‘Presence’ (Present, Not Present, Hypothetical), and ‘Temporality’ (Past, Recent, Future) to create a ground-truth dataset for supervised training. NLP performance was evaluated by comparing model predictions to expert annotations.ResultsIn training, the NER model achieved high performance in extracting concepts and linking them to SNOMED terms, with a mean positive predictive value (PPV) of 0.94 and a mean sensitivity of 0.86 (Figure 1). Several outliers included sensitivity for myocardial infarction, aortic valve regurgitation, and aortic valve stenosis which are commonly expressed as abbreviations in text, potentially complicating detection by the model.For contextualisation, the ‘Experiencer’ model achieved a mean PPV of 0.91 and a mean sensitivity of 0.82, the ‘Presence’ model achieved a mean PPV of 0.61 and a mean sensitivity of 0.68, and the ‘Temporality’ model achieved a mean PPV of 0.89 and a mean sensitivity of 0.84 (Figure 2). The lower metrics of the ‘Presence’ model reflect the imbalance in the training data, as "Present" appeared around 20 times more frequently than the labels of "Not Present" and "Hypothetical" combined, biasing the algorithm to the majority class.ConclusionWe present proof of concept for the use of NLP as a novel patient identification tool in DCM using EHRs. Automated generation of large-scale cohorts with this pipeline can support clinical audit and observational research, recruitment for trials, and targeted genetic screening initiatives to improve disease characterisation.

Ball bearings are vital for rotating machinery, requiring reliable fault diagnosis. Traditional approaches rely on manual feature extraction, ‎requiring specialised expertise. Deep learning reduces human input but struggles with capturing global input context, integrating statistical ‎features, and computational costs. This work introduces a CNN-based fault diagnosis method with Mix-up augmentation. Vibration signals ‎are transformed into 2D time-frequency images via Continuous Wavelet Transform (CWT) to retain temporal-spectral information. Mix-up ‎enhances dataset diversity, improving model robustness. CNNs then classify fault type and severity using these augmented inputs. Evaluat-‎ed on experimental and CWRU datasets, the approach surpasses state-of-the-art methods in accuracy and stability. Combining CWT’s de-‎tailed analysis, Mix-up’s data enrichment, and CNNs’ automated feature extraction resolves prior limitations, delivering an efficient solution ‎for industrial fault detection. The framework ensures reliable, resource-effective automation, advancing predictive maintenance in rotating ‎machinery‎.

Much medical data is only available in unstructured electronic health records (EHR). These data can be obtained through manual (human) extraction or programmatic natural language processing (NLP) methods. We estimate that NLP only becomes economically competitive with manual extraction when there are ~6500 EHRs records. We have found that there is interest from clinicians and researchers in using NLP on projects with fewer records. We examine whether a large language model (LLM) can be used to reduce the cost of NLP to make it economically competitive for such projects, and study the feasibility of such framework for accuracy. We developed an NLP pipeline using an off-the-shelf open LLM to extract breast cancer ER, PR, and HER2 biomarker data. Pipeline development stopped when the prompts performances were competitive with manual extraction. The development time and extraction performance were compared to those for an existing rule-based (RB) NLP pipeline. The code for the extraction portion of the LLM pipeline is available at https://github.com/sehagler/llm_biomarker_extraction . The LLM pipeline produced performance competitive with manual data extraction with a hands-on development time that was ~38% that of the RB pipeline. LLMs exhibit lower hands-on development costs compared to standard NLP techniques, but require significant and potentially costly computation resources. LLMs may potentially allow the economically competitive application of NLP to smaller projects if computation costs can be managed.

ABSTRACT This study addresses limitations in the non‑destructive evaluation of grouted rock bolts, notably their susceptibility to ambient noise interference and reliance on manual feature extraction by experts. An intelligent framework integrating impact acoustics and deep learning is proposed. A theoretical model, derived from dynamic equation analysis and experimental fitting, establishes a strong negative correlation between the unbonded length and the fundamental frequency. Experiments were conducted on three‑layer rock bolt models with five distinct bonding conditions, where impulse excitation was applied to acquire acoustic signals. Wavelet transform was employed to convert one‑dimensional time‑domain signals into two‑dimensional time–frequency representations, thereby enriching feature representation for model input. Seven classification algorithms were evaluated; the hybrid CNN–SVM model attained optimal classification accuracy (100%) under controlled laboratory conditions, significantly outperforming other methods and establishing a performance benchmark for ideal scenarios. Although real‑world deployment may be affected by environmental variability, the proposed approach exhibits robust performance within the experimental scope. By integrating theoretical modeling, wavelet‑based signal dimension enhancement, and deep learning, this method provides an automated, high‑precision, and interpreter‑independent solution for detecting grouting defects, demonstrating potential for industrial quality assurance in rock bolt systems.

A pretrained foundation model for headache disorders based on magnetoencephalography