Large Number Of Samples Research Articles

Automatic Road Crack Detection Using Convolutional Neural Network Based on Semi-Supervised Learning

Crack detection in pavements is a critical task for infrastructure maintenance, but it often requires extensive manual labeling of training samples, which is both time-consuming and labor-intensive. To address this challenge, this paper proposes a semi-supervised learning approach based on a DenseNet classification model to detect pavement cracks more efficiently. The primary objective is to leverage a small set of labeled samples to improve the model's performance by incorporating a large number of unlabeled samples through semi-supervised learning. This method enhances the DenseNet model's ability to generalize by iteratively learning from new unlabeled datasets. As a result, the proposed approach not only reduces the need for extensive manual labeling but also mitigates issues related to label inconsistency and errors in the original labels. The experimental results demonstrate that the semi-supervised DenseNet model achieves a prediction precision of 96.77% and a recall of 94.17%, with an F1 score of 95.45% and an Intersectidn over Union (IoU) of 91.30%. These metrics highlight the model's high accuracy and effectiveness in crack detection. The proposed method not only improves label quality and model performance but also offers practical value for engineering applications in the field of pavement maintenance, making it a valuable tool for infrastructure management.

A new adversarial malware detection method based on enhanced lightweight neural network

With the gradual expansion of Android systems from mobile phones to intelligent devices, a huge amount of malware has been found every year. To improve the malware detection performance and reduce its reliance on expert experience, deep learning technology has been widely used. However, as the complexity of deep learning models continues to increase, it rapidly increases the consumption of hardware resources. At the same time, anti-detection technology such as Generative Adversarial Networks (GANs) are widely used to evade Artificial Intelligence (AI)-based detection methods. In this paper, we propose a new classification model based on an improved lightweight neural network that can effectively improve the execution efficiency and detection performance of malware detection methods against adversarial malware samples. First, our method uses local-information-entropy-based image generation technology to construct effective image feature vectors. Then, the performance of the lightweight neural network model ESPNetV2 is improved from four aspects. Finally, a new adversarial malware generation model called Mal-WGANGP is proposed. It can automatically generate a large number of adversarial samples to robust our model. In order to evaluate our method, we construct several experiments and compare the detection performance of our method with 19 other novel efficient neural network detection models. Experimental results show that our image enhancement method and detection model have the highest detection accuracy of adversarial samples.

Maximizing efficiency in sedimentary ancient DNA analysis: a novel extract pooling approach

In the last few decades, the field of ancient DNA has taken a new direction towards using sedimentary ancient DNA (sedaDNA) for studying human and mammalian population dynamics as well as past ecosystems. However, the screening of numerous sediment samples from archaeological sites remains a time-consuming and costly endeavor, particularly when targeting hominin DNA. Here, we present a novel high-throughput method that facilitates the fast and efficient analysis of sediment samples by applying a pooled testing approach. This method combines multiple extracts, enabling early parallelization of laboratory procedures and effective aDNA screening. Pooled samples with detectable aDNA signals undergo detailed analysis, while empty pools are discarded. We have successfully applied our method to multiple sediment samples from Middle and Upper Paleolithic sites in Europe, Asia, and Africa. Notably, our results reveal that an aDNA signal remains discernible even when pooled with four negative samples. We also demonstrate that the DNA yield of double-stranded libraries increases significantly when reducing the extract input, potentially mitigating the effects of inhibition. By embracing this innovative approach, researchers can analyze large numbers of sediment samples for aDNA preservation, achieving significant cost reductions of up to 70% and reducing hands-on laboratory time to one-fifth.

A New Look at Diffusion in Vapor Intrusion Assessments; Passive Adsorptive Diffusion Samplers

AbstractVapor intrusion of toxic volatile organic compounds (VOCs) from subsurface soil vapor, through a building slab/floor, and into the indoor air is an important environmental contaminant transport mechanism. It is widely believed that advective flow, driven by the pressure differential between the subslab and indoor air, is the primary mechanism of subslab soil vapor entry into buildings. This paper explores the hypothesis that molecular diffusion through the slab may potentially play a larger role in vapor intrusion than previously believed and may even be the predominant vapor intrusion mechanism when the subslab vapor source strength is sufficiently high or the pressure differential is relatively low. A novel sampling device, referred to as a Passive Adsorptive Diffusion Sampler (PADS), is presented for the purpose of directly measuring the diffusion of VOCs through a building slab. A vacant warehouse was identified as a case study site where historical sampling had determined that vapor intrusion of trichloroethene (TCE) was adversely impacting the indoor air. Calculations using Fick's First Law of Diffusion are presented which demonstrate that diffusion alone can theoretically account for all the TCE observed in the indoor air at this building based on an effective diffusion coefficient for concrete that was calculated from the Johnson and Ettinger Model. Two groups of nine replicate PADS were deployed at two areas on the slab and used to measure the flux and effective diffusion coefficient at each of the 18 total points, which showed an order of magnitude variability within each area and over two orders of magnitude variability overall. These results indicate that diffusion through concrete is inherently variable when measured at a sub‐meter scale. However, when combined over both areas, the overall average approached that calculated from the Johnson and Ettinger Model. An additional 12 PADS were deployed across the building slab (for a total of 30) to quantify the overall building‐wide diffusive flux. This area‐weighted average diffusive flux was consistent with the predicted diffusive flux as calculated from Fick's First Law and the vapor intrusion mass input required to achieve the observed indoor air TCE concentration. The results of this study show that PADS provides a simple way to measure diffusive flux directly without having to drill through the slab. However, significant variability in the measured flux should be expected and will need to be accounted for by the inclusion of a relatively large number of samples including replicates. When using PADs at a new site, the collection of traditional subslab vapors at a select number of locations is recommended for the verification of a building‐specific effective diffusion coefficient, which may not necessarily be the same as for this building.

Characterization of bovine vaginal microbiota using 16S rRNA sequencing: associations with host fertility, longevity, health, and production

Due to their potential impact on the host’s phenotype, organ-specific microbiotas are receiving increasing attention in several animal species, including cattle. Specifically, the vaginal microbiota of ruminants is attracting growing interest, due to its predicted critical role on cows’ reproductive functions in livestock contexts. Notably, fertility disorders represent a leading cause for culling, and additional research would help to fill relevant knowledge gaps. In the present study, we aimed to characterize the vaginal microbiota of a large cohort of 1171 female dairy cattle from 19 commercial herds in Northern France. Vaginal samples were collected using a swab and the composition of the microbiota was determined through 16S rRNA sequencing targeting the V3–V4 hypervariable regions. Initial analyses allowed us to define the core bacterial vaginal microbiota, comprising all the taxa observed in more than 90% of the animals. Consequently, four phyla, 16 families, 14 genera and a single amplicon sequence variant (ASV) met the criteria, suggesting a high diversity of bacterial vaginal microbiota within the studied population. This variability was partially attributed to various environmental factors such as the herd, sampling season, parity, and lactation stage. Next, we identified numerous significant associations between the diversity and composition of the vaginal microbiota and several traits related to host’s production and reproduction performance, as well as reproductive tract health. Specifically, 169 genera were associated with at least one trait, with 69% of them significantly associated with multiple traits. Among these, the abundances of Negativibacillus and Ruminobacter were positively correlated with the cows’ performances (i.e., longevity, production performances). Other genera showed mixed relationships with the phenotypes, such as Leptotrichia being overabundant in cows with improved fertility records and reproductive tract health, but also in cows with lower production levels. Overall, the numerous associations underscored the complex interactions between the vaginal microbiota and its host. Given the large number of samples collected from commercial farms and the diversity of the phenotypes considered, this study marks an initial step towards a better understanding of the intimate relationship between the vaginal microbiota and the dairy cow’s phenotypes.

Long-Term Monitoring of Microplastics in a German Municipal Wastewater Treatment Plant

Wastewater treatment plants (WWTPs) have been identified as important point sources for microplastics (MPs) in the environment; monitoring MP emissions in the WWTP effluent is therefore essential for contamination control. The aim of this study is to acquire a large number of samples (320) over a period of two years and three months to determine the temporal variations in microplastic contamination in the outlet of the municipal WWTP Landau-Mörlheim. The effluent of the third cleaning stage is sampled with a 10 µm filter cartridge, processed in the laboratory using a hydrogen peroxide treatment, and MPs are then detected by fluorescence staining. The results show high temporal variations in the microplastic concentrations in the effluent of the WWTP. This indicates that high numbers of samples are necessary to obtain a representative assessment of the microplastic emissions; single samples are not representative. The average microplastic concentration in the effluent was 27.8 ± 29.8 MP/L, ranging from 0.6 MP/L to 194.0 MP/L. This leads to a yearly emission of 1.5 × 1011 MP for the WWTP Landau-Mörlheim, corresponding to an emission of 2.8 × 106 MP/inhabitant and year. Statistically significant seasonal variations could not be observed, although there is a trend towards lower MP concentrations in summer. Further, no correlations with other wastewater or weather parameters could be found.

Scientific analysis of folk contract documents from Tianshui region: insights of fiber use and preservation state

Folk contract documents (FCD) are valuable materials for studying social history, and the paper they use reflects the social realities of different eras and social classes. Research and scientific analysis of numerous FCD samples after the fourteenth century are rare. We conducted a study on 96 Tianshui folk contract documents (TFCD, 107 paper samples) from the Tianshui area of Gansu Province, Northwest China, taking into account both the textual content and the materiality of paper carriers, and interpreted the results from multiple lines of evidence and discussion. Physical performance analysis revealed that the paper used by the northern folk exhibits a lower apparent density, which is not conducive to the long-term preservation of paper. The preservation status investigation, curtain pattern analysis, and fiber analysis show that the paper used in the TFCD differs from traditional cultural paper regarding disease types, production precision, and fiber materials, providing a basis for its protection and restoration. The analysis of chemical components indicated that the aging and yellowing of paper can be correlated with the content of sulfur and carbonyl groups. The feasibility of using pyrolysis gas chromatography-mass spectrometry (Py-GC/MS) analysis to quickly identify papermaking fibers in a large number of paper samples was proposed. By utilizing various techniques to inspect the paper of documents, this study helps to enhance the academic understanding of FCD materials. In addition, it expands the knowledge base of Northwest handmade paper.

Rotary barrel tumbling as a method of surface preparation for pin-on-disc wear testing samples

Surface preparation is essential to ensure sample homogeneity in terms of surface roughness and mechanical properties, as these factors can significantly affect wear behavior and test repeatability. Although conventional semi-automatic and automatic grinding and polishing processes are efficient and well established, limitations arise for the preparation of large numbers of samples of a hard material such as high-carbon steel, including issues with sample geometry, the need for on-demand sample holders, cost-related limitations, and even considerable human workload and expertise. This study aims to evaluate the use of rotary barrel tumbling—a polishing method that relies on the sliding of abrasive media over the samples’ surface inside a rotating barrel—as an alternative method to prepare the surface of wheel and rail steel samples for pin-on-disc wear testing. A 4-stage tumbling procedure was employed, using different compositions of tumbling media in each stage. Surface roughness and hardness were evaluated throughout the process via 2D and 3D profilometry, microhardness Vickers measurements, and optical microscopy. The proposed method resulted in significant reductions in the surface roughness and hardness of the samples, along with improved homogeneity between samples of different materials and within the same material. These findings suggest that rotary barrel tumbling is an effective alternative method for the surface preparation of pins and discs made of high-carbon steel, enhancing the samples’ suitability for subsequent wear tests.

PERBEDAAN KECEPATAN BERJALAN PADA PENGGUNA TRANSTIBIAL PROSTHESIS TIPE SACH FOOT DAN SINGLE AXIS FOOT

Background: Amputation can be defined as the act of separating part of the body, part or all of an extremity. Currently, lower limb amputations account for 85-90% of all amputations and transtibial amputation is the most frequently performed type of amputation surgery. One of the problems with transtibial amputation is hyperextension and weakness of the ligaments in the knee. Knee hyperextension is a condition that occurs when the knee joint is extended (straightening the knee) excessively, thus putting pressure on the structure of the knee and the back of the knee joint. Ligament weakness in the knee is weakness in the anterior cruciate ligament, posterior cruciate ligament, medial collateral ligament and lateral collateral ligament. Choosing the right prosthesis components can overcome this. One component that plays an important role is the foot component. Methods: This type of research is observational analytic with a cross-sectional design. The research subjects were 17 transtibial prosthesis users. Data analysis in this study used Spearman Rank Test. Results: Based on the Spearman Rank results, the significance result was 0.653 for the type of foot with hyperextension, while for the type of foot with ligament weakness, the significance result was 0.434. Both variables had variable values > 0.05, meaning there was no correlation between foot type and hyperextension and ligament weakness in transtibial prosthesis users. Conclusion: There is no correlation type of foot to hyperextension and ligament weakness. In future research, it is hoped that research can be carried out with a larger number of samples and a more varied type of foot.

Automatic Foreign Matter Segmentation System for Superabsorbent Polymer Powder: Application of Diffusion Adversarial Representation Learning

In current industries, sampling inspections of the quality of powders, such as superabsorbent polymers (SAPs) still are conducted via visual inspection. The size of samples and foreign matter are around 500 μm, making them difficult for humans to identify. An automatic foreign matter detection system for powder has been developed in the present study. The powder samples can be automatically delivered, distributed, and recycled, and images of them are captured through the hardware of the system, while the identification software of this system was developed based on diffusion adversarial representation learning (DARL). The background image is a foreign-matter-free powder image with an input image size of 1024 × 1024 × 3. Since DARL includes adversarial segmentation, a diffusion process, and synthetic image generation, the DARL model was trained using a diffusion block with the employment of a U-Net attention mechanism and a spatial-adaptation de-normalization (SPADE) layer through the adoption of a loss function from a vanilla generative adversarial network (GAN). This model was then compared with supervised models such as a fully convolutional network (FCN), U-Net, and DeepLABV3+, as well as with an unsupervised Otsu threshold segmentation. It should be noted that only 10% of the training samples were utilized for the DARL to learn and the intersection over union (IoU) of the DARL can reach up to 80.15%, which is much higher than the 59.00%, 53.47%, 49.39%, and 30.08% for the Otsu threshold segmentation, FCN, U-Net, and DeepLABV3+ models. Therefore, the performance of the model developed in the present study would not be degraded due to an insufficient number of samples containing foreign matter. In practical applications, there is no need to collect, label, and design features for a large number of foreign matter samples before using the developed system.

Comparative analysis of SNaPshot and massively parallel sequencing for body fluid-specific DNA methylation markers.

The identification of tissue-specific differentially methylated regions has significantly contributed to the field of forensic genetics, particularly in body fluid identification crucial for linking evidence to crimes. Among the various approaches to analyzing DNA methylation, the SNaPshot assay has been popularly studied in numerous researches. However, there is a growing interest in exploring alternative methods such as the use of massively parallel sequencing (MPS), which can process a large number of samples simultaneously. This study compares SNaPshot and MPS multiplex assays using nine cytosine-phosphate-guanine markers for body fluid identification. As a result of analyzing 112 samples, including blood, saliva, vaginal fluid, menstrual blood, and semen, both methods demonstrated high sensitivity and specificity, indicating their reliability in forensic investigations. A total of 92.0% samples were correctly identified by both methods. Although both methods accurately identified all blood, saliva, and semen samples, some vaginal fluid samples showed unexpected methylation signals at nontarget loci in addition to the target loci. In the case of menstrual blood samples, due to their complexity, independent typing criteria were applied, and successful menstrual blood typing was possible, whereas a few samples showed profiles similar to vaginal fluid. The MPS method worked better in vaginal fluid samples, and the SNaPshot method performed better in menstrual blood samples. This study offers valuable insights into body fluid identification based on the characteristics of the SNaPshot and MPS methods, which may help in more efficient forensic applications.

Improving DNA recovery and sample throughput using the PrepFiler™ Automated Forensic DNA Extraction Kit on two customised Tecan Fluent® 1080 Automated Workstations

One of the primary goals of forensic laboratories performing DNA recovery and profiling is obtaining a high-quality DNA extract from crime scene samples. This is largely dependent on the sensitivity and reliability of the DNA extraction chemistry utilised, as well as the liquid handling and contamination minimisation techniques employed. Automation of DNA extraction methods on large liquid handling platforms allows high-throughput laboratories to apply sensitive chemistries to reliably process a large number of samples, while minimising manual processes and cross-contamination.This study describes the first known implementation of the PrepFiler™ Automated Forensic DNA Extraction Kit on a Tecan Fluent® Gx 1080 Automation Workstation. Two Workstations were customised with the addition of novel “Safe Pipetting Modules” to eliminate sample crossover between wells, which is important in a forensic biology setting to reduce inadvertent DNA transfer. A comparison of DNA extraction efficiency between an optimised PrepFiler™ method and the DNA IQ™ System performed on a Perkin Elmer Janus® Integrator platform showed the optimised PrepFiler™ method consistently extracted a higher yield of DNA from a range of blood inputs, as well as blood and buccal swabs. The PrepFiler™ chemistry also more efficiently removed humic acid and haematin, reducing subsequent PCR inhibition. The subsequent implementation of the optimised PrepFiler™ method onto the Tecan Fluent® Gx workstations showed a further increase in sensitivity, with no evidence of DNA cross-contamination observed. However, the optimised PrepFiler™ method encountered difficulties extracting DNA from fabric substrates, with the PrepFiler Express™ chemistry extracting higher yields on the cartridge-based AutoMate Express™ System.Overall, this study demonstrated the Tecan Fluent® Gx 1080 Automation Workstation is a sensitive, reliable and robust method for DNA extraction using the PrepFiler™ Automated Forensic DNA Extraction Kit, and the addition of the novel Safe Pipetting Module makes this platform an attractive option for forensic biology laboratories where minimising inadvertent DNA transfer is of paramount importance.

Intelligent Microcontroller-Based Infrared Attenuated Total Reflection Spectroscopy for High–Throughput Screening and Discrimination of Foodborne Fungi

Food safety became one of the most critical issues owing to the large expansion of international trading and emission of various pollutants in air, water and soil. Fungal contamination of food and feed has attracted most of the attention in the last decade because of the emerging analytical tools that facilitate the detection and discrimination of fungal species in imported foodstuff, seeds, grains, plants, meats …etc. In this work, we give an insight on the application of integrated attenuated total reflection–Fourier transform infrared (ATR–FTIR) spectroscopy and artificial–intelligence algorithms to the determination and discrimination of fungal species/strains which potentially infect plants, seeds and grains. The proposed method is based on a microcontroller which allows the PC to analyze a large number of samples via serial connection with an UART module. Penicillium chrysogenum, Aspergillus niger, Aspergillus fumigatus, Aspergillus solani, Aspergillus flavus and two different strains of Fusarium oxysporum were used as model microorganisms. The use of artificial–intelligence algorithms herein provides the advantage of automation enabling high throughput screening of large numbers of food samples in less than 5 s. In addition, the classification accuracy is enhanced by applying these machine–learning classification techniques. Principle component analysis (PCA) was used in order to extract the spectral discriminative features from the recorded fungal FTIR spectra. Three intelligent methods of classification; namely, artificial neural network (ANN), support–vector machine (SVM) and k–nearest neighbor (KNN), were used in this study in order to prove that integration of spectroscopic measurements with varying machine–learning methods give a simple analytical tool for detection and classification of foodborne pathogens. All the utilized classifiers gave an accuracy of 100 % and were able to discriminate different species and/or strains of the investigated fungi in few milliseconds.

Pharmacokinetic Simulation and Area under the Curve Estimation of Drugs Subject to Enterohepatic Circulation.

Enterohepatic circulation (EHC) is a complex process where drugs undergo secretion and reabsorption from the intestinal lumen multiple times, resulting in pharmacokinetic profiles with multiple peaks. The impact of EHC on area under the curve (AUC) has been a topic of extensive debate, questioning the suitability of conventional AUC estimation methods. Moreover, a universal model for accurately estimating AUC in EHC scenarios is lacking. To address this gap, we conducted a simulation study evaluating five empirical models under various sampling strategies to assess their performance in AUC estimation. Our results identify the most suitable model for EHC scenarios and underscore the critical role of meal-based sampling strategies in accurate AUC estimation. Additionally, we demonstrate that while the trapezoidal method performs comparably to other models with a large number of samples, alternative models are essential when sample numbers are limited. These findings not only illuminate how EHC influences AUC but also pave the way for the application of empirical models in real-world drug studies.

Implementation and Efficient Analysis of Preprocessing Techniques in Deep Learning for Image Classification

Deep learning models have recently been preferred to perform certain image-processing tasks. Recently, with the increasing radiation, heat, and poor lighting conditions, the raw image samples may contain noisy and ambiguous information. To process these images, the deep learning model requires a large number of data samples to learn the missing information from other clear data samples. This necessitates training the neural network with a huge dataset. The researchers are now attempting to filter and improve such noisy images via preprocessing in order to provide valid and accurate feature information to the neural network layers. However, certain research studies claim that some useful information may be lost when the image is not preprocessed with an appropriate filter or enhancement technique. The MSA (meta-synthesis and analysis) approach is utilized in this work to present the impact of the image processing applications done with and without preprocessing steps. Also, this work summarizes the existing deep learning-based image processing models utilizing or not preprocessing steps in their implementation. This work has also found that 85% of the existing techniques involve a preprocessing step while developing a deep learning model. However, a maximum accuracy of 96.89% is observed on Sine-Net when it is implemented without a preprocessing and the same model gave 96.85% when implemented with preprocessing. This research provides various research insights on the requirement and non-requirement of preprocessing steps in a deep learning-based implementation.

Investigation of different methods in rapid microbial identification directly from positive blood culture bottles by MALDI-TOF MS.

Sepsis is a life-threatening condition, and rapid and accurate identification of the causative microorganisms from blood cultures is crucial for timely and effective treatment. Although there are many studies on direct identification from blood cultures with MALDI-TOF MS, further standardization is still needed. In our study, we analyzed the performance of three different preparation methods and compared by using two analysis modules of the Bruker Biotyper MALDI-TOF MS for direct identification of bacteria from numerous positive blood culture bottles. The literature reports a limited number of studies that compare different preparation methods for direct blood culture identification, processing a large number of blood samples concurrently and evaluating the same samples as in our study. Moreover, although SDS is used very frequently in medical laboratories, there are few studies on direct identification from blood culture bottles. In our study, the highest correct identification rate was observed with the SDS method.

Two new qPCR assays for detecting and quantifying the Aspergillus flavus and Aspergillus parasiticus clades in maize kernels

AbstractThe fungi of Aspergillus section Flavi can produce carcinogenic mycotoxins—aflatoxins (AFs)—of two types: types B and G (AFBs and AFGs). AFs are highly hazardous for human and animal health. Their levels in food and feed are therefore highly regulated, with a low acceptable limit for AF content. In France, climate change has led to the detection of AFs in maize harvests since 2015. Mycoflora analyses have identified two species, A. flavus (producing AFBs) and A. parasiticus (producing both AFBs and AFGs), as responsible for this AF contamination. However, mycoflora analysis is a time‐consuming method that cannot readily be applied to large numbers of samples. We propose here an alternative clade‐specific functional TaqMan quantitative PCR method based on the calmodulin gene for distinguishing between the A. flavus clade (AfC) and the A. parasiticus clade (ApC). We applied this method to 553 maize samples collected in three different harvest years (2018–2020). Both clades were detected in about 40% of the samples tested. As expected, we observed significant positive correlations between AFBs and AfC DNA (R2 = 0.708), and between AFGs and ApC DNA (R2 = 0.885). This method will be useful for the rapid, simple and cheap characterization of maize grain contamination with Aspergillus section Flavi. This method will make it possible to study the relationship between agroclimatic conditions, AF content and species prevalence, to facilitate the anticipation of AF risks due to global warming in France.

Small data reliability analysis in concrete three-point bending tests: A Weibull mixture model approach based on Weibull fracture theory

Due to the high costs and time-consuming nature of experiments, data from concrete three-point bending tests are usually limited. Additionally, the performance of concrete is influenced by various factors such as mix composition and sample preparation methods, leading to significant variability in test results. This poses a major challenge for the reliability analysis of small data from repeated tests. In reliability analysis, statistical models are particularly important as they allow for the prediction of performance and failure probabilities based on existing data. However, traditional statistical methods require a large number of similar samples for data processing and may need specific adjustments due to the diversity of practical applications. To address these issues, this study proposes a mixture model that combines domain knowledge of concrete with Weibull fracture theory and uses the Clustering-Circular Block Bootstrap method for sample augmentation. By analyzing real data obtained from concrete three-point bending tests, we accurately estimated the failure probability of concrete. Additionally, the study uses Gaussian process regression with a spectral mixture kernel function to fit the threshold fracture strength of concrete and estimate failure probabilities. The results show that the proposed method can accurately fit small datasets, with a RMSE of 0.320 and a MAPE of 0.100. The nominal bending strength range at a 5% failure probability is 2.0 MPa to 6.5 MPa, depending on the sample height and pre-crack length. Statistical learning demonstrates that the proposed method excels in fitting small datasets, providing a solid framework for predicting concrete strength and reliability.

Semi-supervised adaptive anti-noise meta-learning for few-shot industrial gearbox fault diagnosis

Abstract Real-time and accurate predictive maintenance of industrial equipment is fundamental for ensuring the safety and stability of advanced manufacturing processes. Current fault diagnosis methods based on data mining rely on a large number of labeled samples, and obtaining sufficient labeled data for diagnosing industrial equipment faults is challenging. Meta-learning can achieve the diagnosis of few-shot samples to a certain extent, but the effect is not ideal. Semi-supervision can effectively leverage a large number of unlabeled samples, which is of great practical significance for handling scenarios involving limited labeled samples. However, noise interference can occur when unlabeled samples appear that do not belong to known categories. Therefore, this study proposes adaptive semi-supervised meta-learning networks for noisy few-shot gearbox fault diagnosis. Firstly, a residual network with a Morlet Wavelet layer is used to extract signal features. Next, sample-level attention is defined to select unlabeled samples that are more similar to labeled sample prototypes, thereby reducing the influence of noisy samples. The adaptive metric is used to obtain the relational distance functions of labeled samples and unlabeled samples. Adaptive semi-supervised meta-learning networks uses unlabeled data to refine prototypes for better fault diagnosis. The effectiveness and anti-noise performance of the proposed method are verified by using two gearbox datasets with various few-shot noise scenarios.

Serum calcium as a candidate marker in detecting stunting in toddler

<p><strong><em>Background: </em></strong><em>Monitoring the nutritional status of toddlers is a very important thing to do. Calcium is one of the blood parameters that can be assessed, because calcium deficiency will affect the child's linear growth. So far, monitoring of nutritional status is only based on anthropometry and is not followed by assessment of blood parameters. </em><strong><em>Objectives:</em></strong><strong><em> </em></strong><em>This study aims to measure serum calcium levels as a candidate marker in detecting stunting in toddlers. </em></p><p><strong><em>Method:</em></strong><em> The design of this research is cross sectional with 62 samples of toddlers aged 2-5 years. Data on sample characteristics was obtained through a questionnaire, as for serum calcium levels measurement, venous blood samples were taken and measured with the colorimetric method. The incidence of stunting is obtained based on TB/U measurements and Z-score was calculated using Anthro 1.02 software. Data were then analyzed univariately (frequency distribution) and bivariately (</em><em>independent t-test and </em><em>chi-square) using SPSS version 22. </em></p><p><strong><em>Results:</em></strong><em> Based on anthropometric measurements, it was found that 21 (33.9%) toddlers had stunting, 28 (45.1%) toddlers had low Serum calcium levels and the majority of toddlers (54.9%) had normal serum calcium levels. There was no significant difference (p=0.989) in mean serum calcium levels between stunting and non-stunting toddlers (1.961 ± 0.223 µq/dL vs 1.960 ± 0.175 µq/dL). Bivariate results also showed that there was no significant relationship (p=0.414) between serum calcium levels and the incidence of stunting. </em></p><p><strong><em>Conclusion:</em></strong><em> Measuring serum calcium levels cannot be used as a candidate marker for detecting stunting in toddlers. Further research is needed with a larger number of samples and measurements of various parameters related to calcium balance.</em><strong></strong></p>