Search

Background: Multi‐organ segmentation is a critical task in medical imaging, with wide‐ranging applications in both clinical practice and research. Accurate delineation of organs from high‐resolution 3D medical images, such as CT scans, is essential for radiation therapy planning, enhancing treatment outcomes, and minimizing radiation toxicity risks. Additionally, it plays a pivotal role in quantitative image analysis, supporting various medical research studies. Despite its significance, manual segmentation of multiple organs from 3D images is labor‐intensive and prone to low reproducibility due to high interoperator variability. Recent advancements in deep learning have led to several automated segmentation methods, yet many rely heavily on labeled data and human anatomy expertise. Purpose: In this study, our primary objective is to address the limitations of existing semi‐supervised learning (SSL) methods for abdominal multi‐organ segmentation. We aim to introduce a novel SSL approach that leverages unlabeled data to enhance the performance of deep neural networks in segmenting abdominal organs. Specifically, we propose a method that incorporates a redrawing network into the segmentation process to correct errors and improve accuracy. Methods: Our proposed method comprises three interconnected neural networks: a segmentation network for image segmentation, a teacher network for consistency regularization, and a redrawing network for object redrawing. During training, the segmentation network undergoes two rounds of optimization: basic training and readjustment. We adopt the Mean‐Teacher model as our baseline SSL approach, utilizing labeled and unlabeled data. However, recognizing significant errors in abdominal multi‐organ segmentation using this method alone, we introduce the redrawing network to generate redrawn images based on CT scans, preserving original anatomical information. Our approach is grounded in the generative process hypothesis, encompassing segmentation, drawing, and assembling stages. Correct segmentation is crucial for generating accurate images. In the basic training phase, the segmentation network is trained using both labeled and unlabeled data, incorporating consistency learning to ensure consistent predictions before and after perturbations. The readjustment phase focuses on reducing segmentation errors by optimizing the segmentation network parameters based on the differences between redrawn and original CT images. Results: We evaluated our method using two publicly available datasets: the beyond the cranial vault (BTCV) segmentation dataset (training: 44, validation: 6) and the abdominal multi‐organ segmentation (AMOS) challenge 2022 dataset (training:138, validation:16). Our results were compared with state‐of‐the‐art SSL methods, including MT and dual‐task consistency (DTC), using the Dice similarity coefficient (DSC) as an accuracy metric. On both datasets, our proposed SSL method consistently outperformed other methods, including supervised learning, achieving superior segmentation performance for various abdominal organs. These findings demonstrate the effectiveness of our approach, even with a limited number of labeled data. Conclusions: Our novel semi‐supervised learning approach for abdominal multi‐organ segmentation addresses the challenges associated with this task. By integrating a redrawing network and leveraging unlabeled data, we achieve remarkable improvements in accuracy. Our method demonstrates superior performance compared to existing SSL and supervised learning methods. This approach holds great promise in enhancing the precision and efficiency of multi‐organ segmentation in medical imaging applications. [ABSTRACT FROM AUTHOR]

"Weatherscapes" is a new column in Weatherwise magazine by Contributing Editor Ed Darack that delves into the use of supercomputing and artificial intelligence in atmospheric science. The article explores the evolution of AI and its practical applications in meteorology, dispelling common misconceptions about the field. It traces the historical development of AI from its conceptual origins to its current role in enhancing weather forecasting and atmospheric research. The text highlights the growing importance of AI in atmospheric science and its potential to revolutionize weather prediction methods. [Extracted from the article]

This paper proposes a lightweight coal and gangue detection algorithm based on Yolov7-tiny for the current target detection algorithms, which have problems such as large model computation, low detection accuracy and low real-time detection performance. Firstly, we make cuts for the ShuffleNetV2 network and replace the Stage1 and Stage2 layers with CONV convolutional blocks to construct a new backbone network that improves the speed of feature extraction. Secondly, the neck of the original model is redesigned by fusing the original two paths of up-sampling and down-sampling into one path, introducing the CBAM attention mechanism to further improve the information extraction, fusion and interaction capabilities of the model, and the detection heads are reduced from three to two to improve the real-time inference capability of the model. Finally, SIoU is used to replace the loss function of the original network to improve the convergence ability and robustness of the model. The experimental results show that the accuracy and FPS of the lightweight model proposed in this paper reach 0.981 and 208, respectively. The improved model has higher real-time and accuracy, and the reduction of computation provides the possibility of deploying it on the edge devices. [ABSTRACT FROM AUTHOR]

Flotation is a physical and chemical process that uses the difference between mineral surface properties to separate valuable minerals from worthless minerals through the attachment of bubbles. The visual information of the flotation foam surface can reflect the flotation effect and will be closely related to the flotation conditions, directly indicating the degree of mineralization of the foam layer. Considering the problem that coal slurry foam images are difficult to segment due to the mutual adhesion of bubbles and fuzzy boundaries, in this work, we proposed a new and effective segmentation algorithm for extracting coal slurry flotation foam edge information. Based on the encoder and decoder structure, a selective multi-branch input segmentation deep neural network model was designed. We evaluated the proposed method and model using a self-built coal slurry foam image segmentation dataset. The intersection over union, mean intersection over union, and PA evaluation indicators of this method increased by 12.8376%, 7.2716%, and 1.457%, respectively, compared with those of the U-Net model. The experimental results showed that the selective branch segmentation network model had high accuracy and robustness in extracting coal slurry bubble edge information, and it could effectively solve the problem of difficult segmentation of coal slurry foam images. This study provided a foundation for optimizing flotation process parameters and realizing intelligent flotation site parameters. [ABSTRACT FROM AUTHOR]

Individuals with lower limb impairment after a stroke or spinal cord injury require rehabilitation, but traditional methods can be challenging for both patients and therapists. Robotic systems have been developed to help; however, they currently cannot detect the continuous gait phase in real time, hindering their effectiveness. To address this limitation, researchers have attempted to develop gait phase detection in general using fuzzy logic algorithms and neural networks. However, there is a paucity of research on real-time and continuous gait phase detection. In light of this gap, we propose an unsupervised learning method for real-time and continuous gait phase detection. This method employs windows of real-time trajectories and a pre-trained model, utilizing trajectories from treadmill walking data, to detect the real-time and continuous gait phase of human on overground locomotion. The neural network model that we have developed exhibits an average time error of less than 11.51 ms across all walking conditions, indicating its suitability for real-time applications. Specifically, the average time error during overground walking at different speeds is 11.20 ms, which is comparatively lower than the average time error observed during treadmill walking, where it is 12.42 ms. By utilizing this method, we can predict the real-time phase using a pre-trained model from treadmill walking data collected with a full motion capture system, which can be performed in a laboratory setting, thereby eliminating the need for overground walking data, which can be more challenging to obtain due to the complexity of the setting. [ABSTRACT FROM AUTHOR]

The solution of inverse problems is of fundamental interest in medical and astronomical imaging, geophysics as well as engineering and life sciences. Recent advances were made by using methods from machine learning, in particular deep neural networks. Most of these methods require a huge amount of data and computer capacity to train the networks, which often may not be available. Our paper addresses the issue of learning from small data sets by taking patches of very few images into account. We focus on the combination of model‐based and data‐driven methods by approximating just the image prior, also known as regularizer in the variational model. We review two methodically different approaches, namely optimizing the maximum log‐likelihood of the patch distribution, and penalizing Wasserstein‐like discrepancies of whole empirical patch distributions. From the point of view of Bayesian inverse problems, we show how we can achieve uncertainty quantification by approximating the posterior using Langevin Monte Carlo methods. We demonstrate the power of the methods in computed tomography, image super‐resolution, and inpainting. Indeed, the approach provides also high‐quality results in zero‐shot super‐resolution, where only a low‐resolution image is available. The article is accompanied by a GitHub repository containing implementations of all methods as well as data examples so that the reader can get their own insight into the performance. [ABSTRACT FROM AUTHOR]

Sumangala, D.; Abhinav, G.; Nagamani, P.V., and Warrior, H., 2024. Enhanced upwelling dynamics simulation in the Bay of Bengal: Integrating ANNs into hydrodynamic models. Journal of Coastal Research, 40(6), 1125–1136. Charlotte (North Carolina), ISSN 0749-0208. The Bay of Bengal exhibits a unique upwelling pattern, which is most pronounced during the summer monsoon months. This phenomenon is driven by a combination of open-ocean processes and seasonally reversing winds, leading to the transport of cooler, nutrient-rich waters to the coastal areas. The upwelling process is crucial for the marine ecosystem, supporting a diverse range of organisms that thrive in colder temperatures. To accurately capture the upwelling dynamics in the Bay of Bengal, a combination of the Delft3D model and an artificial neural network (ANN) was employed. The Delft3D model alone was unable to adequately represent the upwelling phenomenon, particularly the vertical velocity component (w). The incorporation of an ANN into the Delft3D framework significantly improved the accuracy of vertical velocity predictions. This modification provided a better capacity to forecast velocities in the Bay of Bengal, with a correlation of up to 0.92 (0.97) for the u (v) velocity components. The findings of this study highlight the importance of incorporating ANNs into hydrodynamic models to accurately represent complex processes such as upwelling. This approach provides a more comprehensive understanding of these critical processes, enabling better management and conservation of marine ecosystems. [ABSTRACT FROM AUTHOR]

Background: Mavacamten is a first-in-class cardiac myosin inhibitor approved by the US Food and Drug Administration (FDA) for symptomatic obstructive hypertrophic cardiomyopathy (HCM). This pharmacovigilance study aimed to assess mavacamten-related adverse drug reactions (ADRs) in the real world as reported in the FDA Adverse Event Reporting System (FAERS). Methods: We conducted disproportionality analyses with four signal detection algorithms—reporting odds ratio (ROR), proportional reporting ratio (PRR), Bayesian confidence propagation neural network, and the multi-item gamma Poisson shrinker to identify mavacamten-related ADRs. Results: Out of 4,500,131 reports from the FAERS database, 1004 mavacamten-related ADRs were identified from 1 January 2022 to 30 September 2023. A total of 26 significant disproportionality preferred terms (PTs) conforming to the four signal detection algorithms were noted. Some of the statistically significant cardiac ADRs at PT level include decreased ejection fraction (EF) [ROR 33.60 (95% confidence interval, CI 21.79–51.82), PRR 32.86 (χ2 615.96), information component (IC) 5.03, IC025 4.61, empirical Bayesian geometric mean (EBGM) 32.77, EBGM05 21.25], cardiac failure [ROR 9.39 (95% CI 6.49–13.60), PRR 9.13 (χ2 202.42), IC 3.19, IC025 2.83, EBGM 9.12, EBGM05 6.30], and atrial fibrillation [ROR 16.63 (95% CI 12.72–21.75), PRR 15.66 (χ2 769.93), IC 3.97, IC025 3.71, EBGM 15.64, EBGM05 11.96]. Conclusions: The results of our study were consistent with the safety data of clinical trials, including reduced ejection fraction, atrial fibrillation, dyspnea, and syncope. We also found potential new and unexpected ADR signals, such as urinary tract infection, gout, and peripheral edema. [ABSTRACT FROM AUTHOR]

Developing non-contact, non-destructive monitoring methods for marine life is crucial for sustainable resource management. Recent monitoring technologies and machine learning analysis advancements have enhanced underwater image and acoustic data acquisition. Systems to obtain 3D acoustic data from beneath the seafloor are being developed; however, manual analysis of large 3D datasets is challenging. Therefore, an automatic method for analyzing benthic resource distribution is needed. This study developed a system to estimate benthic resource distribution non-destructively by combining high-precision habitat data acquisition using high-frequency ultrasonic waves and prediction models based on a 3D convolutional neural network (3D-CNN). The system estimated the distribution of asari clams (Ruditapes philippinarum) in Lake Hamana, Japan. Clam presence and count were successfully estimated in a voxel with an ROC-AUC of 0.9 and a macro-average ROC-AUC of 0.8, respectively. This system visualized clam distribution and estimated numbers, demonstrating its effectiveness for quantifying marine resources beneath the seafloor. [ABSTRACT FROM AUTHOR]

Arsenic (As) contamination in drinking water has been highlighted for its environmental significance and potential health implications. Iron-based filters are cost-effective and sustainable solutions for As removal from contaminated water. Applying Machine Learning (ML) models to investigate and optimize As removal using iron-based filters is limited. The present study developed Deep Learning Neural Network (DLNN) models for predicting the removal of As and other contaminants by iron-based filters from groundwater. A small Original Dataset (ODS) consisting of 20 data points and 13 groundwater parameters was obtained from the field performances of 20 individual iron-amended ceramic filters. Cubic-spline interpolation (CSI) expanded the ODS, generating 1600 interpolated data points (IDPs) without duplication. The Bayesian optimization algorithm tuned the model hyper-parameters and IDPs in a Stratified fivefold Cross-Validation (CV) setup trained all the models. The models demonstrated reliable performances with the coefficient of determination (R2) 0.990–0.999 for As, 0.774–0.976 for Iron (Fe), 0.934–0.954 for Phosphorus (P), and 0.878–0.998 for predicting manganese (Mn) in the effluent. Sobol sensitivity analysis revealed that As (total order index (ST) = 0.563), P (ST = 0.441), Eh (ST = 0.712), and Temp (ST = 0.371) are the most sensitive parameters for the removal of As, Fe, P, and Mn. The comprehensive approach, from data expansion through DLNN model development, provides a valuable tool for estimating optimal As removal conditions from groundwater. [ABSTRACT FROM AUTHOR]

Deep Neural Networks (DNNs) have achieved remarkable accuracy for numerous applications, yet their complexity often renders the explanation of predictions a challenging task. This complexity contrasts with easily interpretable statistical models, which, however, often suffer from lower accuracy. Our work suggests that this underperformance may stem more from inadequate training methods than from the inherent limitations of model structures. We hereby introduce the Synced Explanation-Enhanced Neural Network (SEE-Net), a novel architecture integrating a guiding DNN with a shallow neural network, functionally equivalent to a two-layer mixture of linear models. This shallow network is trained under the guidance of the DNN, effectively bridging the gap between the prediction power of deep learning and the need for explainable models. Experiments on image and tabular data demonstrate that SEE-Net can leverage the advantage of DNNs while providing an interpretable prediction framework. Critically, SEE-Net embodies a new paradigm in machine learning: it achieves high-level explainability with minimal compromise on prediction accuracy by training an almost "white-box" model under the co-supervision of a "black-box" model, which can be tailored for diverse applications. [ABSTRACT FROM AUTHOR]

Quantum entanglement acts as a crucial part in quantum computation and quantum information, hence quantifying unknown entanglement is an important task. Due to the fact that the amount of entanglement cannot be achieved directly by measuring any physical observables, it remains an open problem to quantify entanglement experimentally. In this work, we provide an effective way to quantify entanglement for the unknown quantum states via artificial neural networks. By choosing the expectation values of measurements as input features and the values of entanglement measures as labels, we train artificial neural network models to predict the entanglement for new quantum states accurately. Our method does not require the full information about unknown quantum states, which highlights the effectiveness and versatility of machine learning in exploring quantum entanglement. [ABSTRACT FROM AUTHOR]

Clinical sleep diagnosis traditionally relies on polysomnography (PSG) and expert manual classification of sleep stages. Recent advancements in deep learning have shown promise in automating sleep stage classification using a single PSG channel. However, variations in PSG acquisition devices and environments mean that the number of PSG channels can differ across sleep centers. To integrate a sleep staging method into clinical practice effectively, it must accommodate a flexible number of PSG channels. In this paper, we proposed FlexSleepTransformer, a transformer-based model designed to handle varying number of input channels, making it adaptable to diverse sleep staging datasets. We evaluated FlexSleepTransformer using two distinct datasets: the public SleepEDF-78 dataset and the local SleepUHS dataset. Notably, FlexSleepTransformer is the first model capable of simultaneously training on datasets with differing number of PSG channels. Our experiments showed that FlexSleepTransformer trained on both datasets together achieved 98% of the accuracy compared to models trained on each dataset individually. Furthermore, it outperformed models trained exclusively on one dataset when tested on the other dataset. Additionally, FlexSleepTransformer surpassed state-of-the-art CNN and RNN-based models on both datasets. Due to its adaptability with varying channels numbers, FlexSleepTransformer holds significant potential for clinical adoption, especially when trained with data from a wide range of sleep centers. [ABSTRACT FROM AUTHOR]

Accurate detection plays a critical role in improving the safety situation of vulnerable road users. This study extends infrastructure-based LiDAR application to all three major vulnerable road user groups including pedestrians, cyclists, and wheelchair users. Two critical problems for accurate detection of small-sized road users are scanning angle variability and feature fluctuation. To address these issues, a feature-based classification method combined with prior LiDAR trajectory information is developed. Effective dimension-related features are proposed and five classifiers including artificial neural network (ANN), random forest (RF), adaptive boosting (AdaBoost), random under-sampling boosting (RUSBoost), and long short-term memory (LSTM) are tested with a novel feature engineering process. A total of seven features are selected from the point cloud of clusters for vehicle/pedestrian/cyclist/wheelchair classification. By updating these significant features based on prior information of the entire trajectory, the performance of road user classification (imbalanced datasets) has been significantly improved. Experimental study is conducted to examine the recall rate, F1-score, and AUC of vehicles, pedestrians, cyclists, and wheelchairs before and after integration with prior trajectory information. The result shows the trained AdaBoost, RUSBoost, and LSTM classifiers with prior trajectory information can achieve recall/F1-score/AUC: (1) Low traffic volumes – vehicles (100%/99.96%/99.96%), pedestrians (99.96%/99.96%/99.97%), cyclists (99.74%/99.45%/99.67%), and wheelchairs (99.22%/99.68%/99.01%) and (2) Moderate traffic volumes – vehicles (99.39%/99.44%/99.69%), pedestrians (98.33%/97.99%/98.64%), and cyclists (95.41%/94.29%/94.40%), using 32-laser LiDAR sensors (10 Hz). [ABSTRACT FROM AUTHOR]

Soybean (Glycine max) is the world's most cultivated legume; currently, most of its varieties are Bt. Spodoptera spp. (Lepidoptera: Noctuidae) are important pests of soybean. An artificial neural network (ANN) is an artificial intelligence tool that can be used in the study of spatiotemporal dynamics of pest populations. Thus, this work aims to determine ANN to identify population regulation factors of Spodoptera spp. and predict its density in Bt soybean. For two years, the density of Spodoptera spp. caterpillars, predators, and parasitoids, climate data, and plant age was evaluated in commercial soybean fields. The selected ANN was the one with the weather data from 25 days before the pest's density evaluation. ANN forecasting and pest densities in soybean fields presented a correlation of 0.863. It was found that higher densities of the pest occurred in dry seasons, with less wind, higher atmospheric pressure and with increasing plant age. Pest density increased with the increase in temperature until this curve reached its maximum value. ANN forecasting and pest densities in soybean fields in different years, seasons, and stages of plant development were similar. Therefore, this ANN is promising to be implemented into integrated pest management programs in soybean fields. [ABSTRACT FROM AUTHOR]

Temperature-related mortality is the leading cause of weather-related deaths in the United States. Herein, we explore the effect of air masses (AMs) – a relatively novel and holistic measure of environmental conditions – on human mortality across 61 cities in the United States. Geographic and seasonal differences in the effects of each AM on deseasonalized and detrended anomalous lagged mortality are examined using simple descriptive statistics, one-way analyses of variance, relative risks of excess mortality, and regression-based artificial neural network (ANN) models. Results show that AMs are significantly related to anomalous mortality in most US cities, and in most seasons. Of note, two of the three cool AMs (Cool and Dry-Cool) each show a strong, but delayed mortality response in all seasons, with peak mortality 2 to 4 days after they occur, with the Dry-Cool AM having nearly a 15% increased risk of excess mortality. Humid-Warm (HW) air masses are associated with increases in deaths in all seasons 0 to 1 days after they occur. In most seasons, these near-term mortality increases are offset by reduced mortality for 1–2 weeks afterwards; however, in summer, no such reduction is noted. The Warm and Dry-Warm AMs show slightly longer periods of increased mortality, albeit slightly less intensely as compared with HW, but with a similar lag structure by season. Meanwhile, the most seasonally consistent results are with transitional weather, whereby passing cold fronts are associated with a significant decrease in mortality 1 day after they occur, while warm fronts are associated with significant increases in mortality at that same lag time. Finally, ANN modeling reveals that AM-mortality relationships gleaned from a combined meta-analysis can actually lead to more skillful modeling of these relationships than models trained on some individual cities, especially in the cities where such relationships might be masked due to low average daily mortality. [ABSTRACT FROM AUTHOR]

Objectives: This study aimed to explore the endothelialization process and assess the potential association between endothelialization and peri-device leak (PDL) following Watchman implantation via a quantitative method. Methods: This is a single-center retrospective study of consecutive patients undergoing LAAO between December 2015 and November 2021. Device endothelialization, compared between PDL and non-PDL group, were quantitatively analyzed based on hypoattenuated thickening in cardiac computed tomography angiography (CCTA). Advancement in endothelialization over time were explored using the Cochran-Armitage test and generalized estimating equation approach. Potential risk factors of delayed endothelialization were analyzed using the Cox proportional-hazards model. Results: A total of 172 patients (mean age, 68 years ± 10 [standard deviation], 114 men) were finally included. The average endothelialization ratio of the study population was 89.8 ± 7.2 percent. In the follow-up period of postprocedural 3 months to more than 12 months, an incremental trend of endothelialization over time was observed with the ratio of 85.8 ± 8.0, 89.6 ± 7.6, 92.2 ± 4.5, 94.3 ± 2.9 percent, respectively (p < 0.0001). Notably, patients without PDL exhibited a swifter advancement in endothelialization compared to those with PDL, irrespective of device size. The multivariable Cox regression model showed that PDL (HR = 2.113, 95%CI: 1.300–3.435, p = 0.003), DSP (HR = 1.717, 95%CI: 1.113–2.647, p = 0.014) were independent risk factors of delayed endothelialization. Conclusion: CCTA holds promise as an effective means of quantitatively assessing device endothelialization. Endothelialization advanced gradually over time, with PDL potentially impeding device endothelialization. Clinical relevance statement: A comprehensive understanding of the correlation between endothelialization ratio, time, and residual shunt can establish a more dependable foundation for determining the appropriate anticoagulation treatment following left atrial appendage closure. Key Points: Current recommendations for postleft atrial appendage occlusion anti-platelet and anticoagulation therapy are based on animal studies. Cardiac computed tomography angiography (CCTA) combined with the UNet neural network model enables the quantitative assessment of device endothelialization. This technique will allow for additional studies to better understand device endothelialization to optimize treatments in this population. [ABSTRACT FROM AUTHOR]

Projects related to Geotechnical Engineering require thorough ground study to know the soil behaviour thus making it necessary to determine the strength and stability parameters of soil. The soils of Assam are primarily of four types viz., Alluvial (old and new), Piedmont, Hill and Lateritic soils with the alluvium abundantly available in the state. The most important soil parameters are the shear strength parameters, namely, cohesion (c) and angle of internal friction (Φ) which can be obtained using triaxial tests with different drainage conditions. But one triaxial test may take weeks to complete and are very expensive. Moreover, for large projects it is practically preposterous to conduct large number of triaxial tests. Hence, an alternate way is to develop a relation between different parameters which enable us to determine others. The main objective of this study is establishing prediction models that allow the shear parameters to be predicted from soil index properties. For this, three prediction models are developed by combining various soil parameters viz., Liquid Limit (LL), Plastic Limit (PL), Specific Gravity (G) and Optimum Moisture Content (OMC) using Multivariate Regression Analysis (MRA) and Artificial Neural Network (ANN). The results show better performance for ANN than MRA. [ABSTRACT FROM AUTHOR]

The study presents analysis of 1 year 15 min cycle radon data using artificial neural network (ANN) in an attempt to estimate earthquake prediction time near Indo-Burman subduction line. The region was found to be seismically active and radon anomalies respond well to it, when majority of the radon anomalies peaks and selected earthquakes correlated. The observation also shows that application of non-linear technique ANN to a non-linear data like radon seems a promising approach in predicting geophysical phenomena at one of the most seismically active region of the world. [ABSTRACT FROM AUTHOR]

Modern deep neural networks are prone to learn domain-dependent shortcuts and thus usually suffer from severe performance degradation when tested in unseen target domains due to their poor ability of out-of-distribution generalization, which significantly limits the real-world applications. The main reason is the domain shift lying in the large distribution gap between source and unseen target data. To this end, this paper takes a step towards training robust models for domain generalizable visual tasks, which mainly focuses on learning domain-invariant visual representation to alleviate the domain shift. Specifically, we first propose an effective Hierarchical Visual Transformation (HVT) network to (1) first transform the training sample hierarchically into new domains with diverse distributions from three levels: Global, Local, and Pixel, (2) then maximize the visual discrepancy between the source domain and new domains, and minimize the cross-domain feature inconsistency to capture domain-invariant features. Besides, we further enhance the HVT network by introducing the environment-invariant learning. To be specific, we enforce the invariance of the visual representation across automatically inferred environments by minimizing invariant learning loss that considers the weighted average of environmental losses. In this way, we can prevent the model from relying on the spurious features for prediction, thus helping the model to effectively learn domain-invariant representation and narrow the domain gap in various visual matching and recognition tasks, such as stereo matching, pedestrian retrieval, and image classification. We term our extended HVT as EHVT to show distinction. We integrate our EHVT network into different models and evaluate its effectiveness and compatibility on several public benchmark datasets. Extensive experiments clearly show that our EHVT can substantially enhance the generalization performance in various tasks. Our codes are available at https://github.com/cty8998/EHVT-VisualDG. [ABSTRACT FROM AUTHOR]

Functionally graded graphene nanoplatelet reinforced composite (FG-GNPRC) have exhibited significant potential for the development of high-performance and multifunctional structures. In this paper, we present a machine learning (ML) assisted uncertainty analysis of nonlinear vibration of FG-GNPRC membranes under the influence of multi-factor coupling. Effective medium theory (EMT), Mori-Tanaka (MT) model and rule of mixture are utilized to evaluate the effective material properties of the composite membrane. Governing equations are derived via an energy method with the frameworks of the hyperelastic membrane theory, Neo-Hookean constitutive model and the couple dielectric theory. Randomly generated inputs after data pre-processing are fed into governing equations, which are solved by numerical methods for outputs. Three ML models, including artificial neural network (ANN), support vector regression (SVR) and AutoGluon-Tabular (AGT), are adopted to capture the complex relationship between the systematic inputs (i.e., structural dimensions, attributes of GNPs and pores, external electric field, etc.), frequency ratio and dimensionless amplitude of FG-GNPRC membranes. The results demonstrate that all three ML models demonstrate exceptional computational efficiency, and AGT presents higher prediction accuracy compared to the other two models. Based on the Shapley additive explanations (SHAP) approach, the effects of uncertainties of system parameters and multi-factor coupling on the nonlinear vibration of the FG-GNPRC membrane are analyzed. It is found that the uncertainty of structural parameters has the greatest impact on the nonlinear vibration of FG-GNPRC membranes, particularly when the membrane is subjected to a voltage of 10 V and smaller stretching ratio. [ABSTRACT FROM AUTHOR]

Purpose: To evaluate the diagnostic performance (DP) of the high-resolution contrast computed tomography (HR-contrast-CT) based Neck-Persistency-Net in distinguishing vital from non-vital persistent cervical lymph nodes (pcLNs) in patients with advanced head and neck squamous cell carcinoma (HNSCC) following primary concurrent chemoradiotherapy (CRT) with [18F]-fluorodeoxyglucose positron emission tomography and high-resolution contrast-enhanced computed tomography ([18F]FDG-PET-CT). Furthermore, the Neck-Persistency-Net's potential to justify omitting post-CRT neck dissection (ND) without risking treatment delays or preventing unnecessary surgery was explored. Methods: All HNSCC patients undergoing primary CRT followed by post-CRT-ND for pcLNs recorded in the institutional HNSCC registry were analyzed. The Neck-Persistency-Net DP was explored for three scenarios: balanced performance (BalPerf), optimized sensitivity (OptSens), and optimized specificity (OptSpec). Histopathology of post-CRT-ND served as a reference. Results: Among 68 included patients, 11 were female and 32 had vital pcLNs. The Neck-Persistency-Net demonstrated good DP with an area under the curve of 0.82. For BalPerf, both sensitivity and specificity were 78%; for OptSens (90%), specificity was 62%; for OptSpec (95%), sensitivity was 54%. Limiting post-CRT-ND to negative results would have delayed treatment in 27%, 40%, and 7% for BalPerf, OptSens and OptSpec, respectively, versus 23% for [18F]FDG-PET-CT. Conversely, restricting post-CRT-ND to positive results would have prevented unnecessary post-CRT-ND in 78%, 60%, and 95% for BalPerf, OptSens and OptSpec, respectively, versus 55% for [18F]FDG-PET-CT. Conclusion: The DP of the Neck-Persistency-Net was comparable to [18F]-FDG-PET-CT. Depending on the chosen decision boundary, the potential to justify the omission of post-CRT-ND without risking treatment delays in false negative findings or reliably prevent unnecessary surgery in false positive findings outperforms the [18F]-FDG-PET-CT. [ABSTRACT FROM AUTHOR]

Accurately predicting drug-target affinity is crucial in expediting the discovery and development of new drugs, which is a complex and risky process. Identifying these interactions not only aids in screening potential compounds but also guides further optimization. To address this, we propose a multi-perspective feature fusion model, MFF-DTA, which integrates chemical structure, biological sequence, and other data to comprehensively capture drug-target affinity features. The MFF-DTA model incorporates multiple feature learning components, each of which is capable of extracting drug molecular features and protein target information, respectively. These components are able to obtain key information from both global and local perspectives. Then, these features from different perspectives are efficiently combined using specific splicing strategies to create a comprehensive representation. Finally, the model uses the fused features to predict drug-target affinity. Comparative experiments show that MFF-DTA performs optimally on the Davis and KIBA data sets. Ablation experiments demonstrate that removing specific components results in the loss of unique information, thus confirming the effectiveness of the MFF-DTA design. Improvements in DTA prediction methods will decrease costs and time in drug development, enhancing industry efficiency and ultimately benefiting patients. • The MFF-DTA model is proposed to accurately predict the affinity relationship between drugs and target proteins. • The model includes multiple feature extractors to capture global and local key features. • We create a comprehensive and information-rich representation. • The model has demonstrated strong performance on public datasets. [ABSTRACT FROM AUTHOR]

The anions C |$_7$| N |$^-$| and C |$_{10}$| H |$^-$| are the two longest of the linear (C,N)-bearing and (C,H)-bearing chains that have so far been detected in the interstellar medium (ISM). In order to glean information on their collision-induced rotational state-changing processes, we analyse the general features of new ab initio potentials describing the interaction of both linear anions with H |$_2$|⁠ , one of the most abundant partners in their ISM environment. We employ an artificial neural network fit of the reduced-dimensionality potential energy surface for C |$_7$| N |$^-$|⁠...H |$_2$| interaction and discuss in detail the spatial features in terms of multipolar radial coefficients. For the C |$_{10}$| H |$^-$|⁠...H |$_2$| interaction, we use the initial grid of two-dimensional raw points to generate by quadrature the Legendre expansion directly, further including the long-range terms as discussed in the main text. Quantum scattering calculations are employed to obtain rotationally inelastic cross-sections, for collision energies in the range of 10 |$^{-4}$| to 400 cm |$^{-1}$|⁠. From them we generate the corresponding inelastic rate coefficients as a function of temperature covering the range from 10 to 50 K. The results for the rate coefficients for the longest cyanopolyyne are compared with the earlier results obtained for the smaller terms of the same series, also in collision with H |$_2$|⁠. We obtain that the inelastic rate coefficients for the long linear anions are all fairly large compared with the earlier systems. The consequences of such findings on their non-equilibrium rotational populations in interstellar environments are illustrated in our conclusions. [ABSTRACT FROM AUTHOR]

We present an artificial neural network (ANN) model of photospheric lithium depletion in cool stars (⁠|$3000\lt T_{\rm eff}/{\rm K} \lt 6500$|⁠), producing estimates and probability distributions of age from |$^7$| Li 6708 Å equivalent width (LiEW) and effective temperature data inputs. The model is trained on the same sample of 6200 stars from 52 open clusters, observed in the Gaia -ESO spectroscopic survey, and used to calibrate the previously published analytical eagles model, with ages 2–6000 Myr and |$-0.3 \lt $| [Fe/H] |$\lt 0.2$|⁠. The additional flexibility of the ANN provides some improvements, including better modelling of the 'lithium dip' at ages |$\lt 50$|  Myr and |$T_{\rm eff}\sim 3500$|  K, and of the intrinsic dispersion in LiEW at all ages. Poor age discrimination is still an issue at ages >1 Gyr, confirming that additional modelling flexibility is not sufficient to fully represent the LiEW–age–T |$_{\text{eff}}$| relationship, and suggesting the involvement of further astrophysical parameters. Expansion to include such parameters–rotation, accretion, and surface gravity–is discussed, and the use of an ANN means these can be more easily included in future iterations, alongside more flexible functional forms for the LiEW dispersion. Our methods and ANN model are provided in an updated version 2.0 of the eagles software. [ABSTRACT FROM AUTHOR]

In the realm of marine intelligence, effective ship classification across vast oceanic expanses is pivotal. Despite strides in conventional identification techniques, existing methods exhibit limitations in terms of effectiveness, robustness, and overall performance. This paper introduces a novel Elephant Herd optimization based Spiking Neural Networks (EHO-SNN) for discerning large ships, small vessels, and the absence of ships. Initial satellite images from the Airbus dataset capture ships at sea, subjected to preprocessing via the wavelet transform-based Retinex algorithm (WRA) to eliminate noise and fog artifacts. Deep learning mobile net facilitates feature extraction, while the Elephant Herd algorithm culls irrelevant features, honing in on the most pertinent ones. Finally, the classification through a spiking neural network, distinguishing between large ships, small vessels, and the absence of ships. Detected large and small ships are accurately positioned within a selected scene, while the absence of a ship terminates the process. The Proposed EHO-SNN model attains an impressive classification accuracy of 99.10%. Notably, it surpasses OMRCNN-SHD, Efficient Net, and AN-YOLOv4 by 1.15%, 12.30%, and 7.79%, respectively, thereby advancing overall accuracy in ship classification. [ABSTRACT FROM AUTHOR]

Objectives: Exhaled breath analysis is an attractive lung cancer diagnostic tool. However, various factors that are not related to the disease status, comorbidities, and other diseases must be considered to obtain a reliable diagnostic model. Methods: Exhaled breath samples from 646 individuals including 273 patients with lung cancer (LC), 90 patients with cancer of other localizations (OC), 150 patients with noncancer lung diseases (NLD), and 133 healthy controls (HC) were analyzed using gas chromatography–mass spectrometry (GC–MS). The samples were collected in Tedlar bags. Volatile organic compounds (VOCs) were preconcentrated on Tenax TA sorbent tubes with subsequent two-stage thermal desorption followed by GC–MS analysis. The influence of age, gender, smoking status, time since last food consumption, and comorbidities on exhaled breath were evaluated. Also, the effect of histology, TNM, tumor localization, treatment status, and the presence of a tumor on VOC profile of patients with lung cancer were assessed. Intergroup statistics were estimated, diagnostic models were created using artificial neural networks (ANNs) and gradient boosted decision trees (GBDTs). Results: Smoking status and food consumption affect exhaled breath VOC profile: benzene, ethylbenzene, toluene, 1,3-pentadiene 1,4-pentadiene acetonitrile, and some ratios are significantly different in exhaled breath of smokers and nonsmokers; the ratios 2,3-butandione/2-pentanone, 2,3-butandione/dimethylsulfide, and 2-butanone/2-pentanone are affected by time since last food consumption. Exhaled breath of LC is affected by the form of the disease and comorbidities. One-pentanol and 2-butanone were different in exhaled breath of patients with various tumor localization; 2-butanone was different in exhaled breath of patients before and during treatment. Diabetes as a comorbidity affects the pentanal level in exhaled breath; obesity affects the ratios of 2,3-butanedione/dimethylsulfide and 2-butanone/isoprene. Sensitivity and specificity of diagnostic models aimed to discriminate LC and HC, OC, and NLD were 78.7% and 51.0%, 62.2% and 53.4%, and 60.4% and 58.0%, respectively. HC and patients, regardless of the disease, can be classified with sensitivity of 76.6% and specificity of 68.2%. Conclusions: The models created to diagnose lung cancer can also classify OC and NLD as patients with lung cancer. Additionally, the influence of comorbidities and factors not related to the disease status must be considered before the creation of diagnostic models to avoid false results. [ABSTRACT FROM AUTHOR]

This paper presents a new approach to determine more robust shear failure criteria, focusing on rough concrete-rock interfaces. The proposed method could also be applied to rock joints. The new approach is based on the distribution and variety of interfaces numerically tested in terms of surface roughness. For this reason, random field simulations are performed using the turning bands method to generate an extensive database of synthetic rough rock surfaces. With this database of synthetic rough rock surfaces, numerical simulations of direct shear tests are carried out. Finally, analytical and neural network models are proposed using the database of shear strength obtained from the finite element simulations to estimate the peak shear strength of concrete-rock interfaces. The performances of the analytical and neural network models in estimating this peak shear strength are evaluated by computing the percentage error and the mean absolute error (MAE) between the predicted and the numerically obtained values. Both models lead to satisfactory predictions. Nevertheless, it is worth noting that the neural network model mildly outperforms the analytical model regarding the magnitude of the error. Furthermore, the neural network model reproduces the possible non-bijective aspect of the correlation roughness-peak shear strength. Highlights: New methodology to investigate the influence of roughness on the shear behavior of interfaces Methodology to generate databases of synthetic rough rock surfaces with controlled roughness values Establishment of numerical simulations as a strategy to perform extensive virtual experimental studies The correlation between interface roughness and shear strength is not bijective. Assessment of artificial neural network modeling as a complementary alternative to failure criteria [ABSTRACT FROM AUTHOR]

A great understanding of the fracture behavior of rocks is very important for ensuring the successful and efficient design, implementation, completion, and structural stability of important large-scale mining, tunneling, oil and gas, and civil engineering projects. This study employed four soft computing models of an artificial neural network trained using the Levenberg–Marquardt algorithm (ANN–LM), grasshopper optimization algorithm-optimized ANN (ANN–GOA), salp swarm algorithm-optimized ANN (ANN–SSA), and arithmetic operation algorithm-optimized ANN (ANN–AOA) to predict the Mode-I fracture toughness (KIc) of rock. For this purpose, a database comprising 121 experimental datasets obtained from the KIc test on a semi-circular bend (SCB) rock samples were used to train and validate the models. Four important parameters affecting KIc, namely, the uniaxial tensile strength, disc specimen radius and thickness, and notch or crack length, were selected as the input parameters. The ANN–GOA 4-9-1 model was adjudged to be the optimum of the generated KIc models as determined by the error metrics used to evaluate model performance. The ANN–GOA 4-9-1 had the lowest error metrics and highest coefficient of correlation for the overall dataset, with R = 0.98498, MSE = 0.0036, VAF = 97.02%, and a20-index = 0.96694. To ensure easy implementation of the optimum ANN–GOA 4-9-1, the model was transformed into a tractable closed-form explicit equation. Furthermore, the impact of each of the four KIc effective parameters on predicted KIc is evaluated and the Brazilian tensile strength and rock specimen radius are determined to be the most sensitive parameters to KIc. Hence, the proposed models can provide a robust and functional reliable alternative to the laborious and costly experimental method for the determination of KIc of rocks. Highlights: An ANN-based model for KIc of SCB specimen prediction is presented. 121 experimental data of SCB specimen was used for model development. Transformation of the optimal ANN-based model into closed-form equation. Variable importance analysis was performed to evaluate the impacts of predictors. [ABSTRACT FROM AUTHOR]

Efficiently determining the properties of the rock mass is essential for evaluating tunnel stability in tunneling projects. The back-analysis technique is commonly employed as an indirect approach for deriving rock mass parameters from field information. However, many back-analysis approaches are often time-consuming in numerical simulations and rely exclusively on monitoring data. This can lead to the identification of rock mass parameters that may not fully represent the characteristics of the surrounding rock. To improve the accuracy of prediction on the tunnel displacement during the tunnel excavation process, this study proposes a probabilistic back-analysis framework to update geo-mechanical parameters. A model based on the matrix-variate deep Gaussian process (MVDGP) and Nesterov-accelerated adaptive moment estimation algorithm (Nadam) is proposed to replace the finite element model to improve its computational efficiency. Initially, a training framework based on ABAQUS–MVDGP is used to conduct the training and evaluation. Subsequently, considering the uncertainty of the geo-mechanical parameters, a probabilistic back analysis is implemented by using the Bayesian theory. Finally, variational inference is used to solve the joint posterior distribution of the geo-mechanical parameters. A rock tunnel is chosen to illustrate the efficacy of the proposed framework. The findings indicate that utilizing the mixed types of monitoring data for vault settlement and convergence enhances the forecasting performance of MVDGP compared to models relying on a single type of monitoring data. The more stable the field monitoring data, the more conducive it is to conduct the back analysis of geo-mechanical parameters of rocks. Furthermore, the comparative results reveal that the presented method demonstrates superior prediction accuracy compared to existing back-analysis models. Highlights: A novel probabilistic back-analysis framework based on matrix-variate deep Gaussian process (MVDGP), Nadam and Bayesian theory provides a significant method for identifying and determining rock mass parameters. Compared to single types of data, mixed types of data have better predictive performance in the MVDGP. The more stable the field monitoring data, the better is the performance of probabilistic back-analysis framework. MVDGP performs better than backpropagation neural network (BPNN), multiple output Gaussian process (MOGP), multi-output support vector regression, and proper orthogonal decomposition and artificial neural network (POD–ANN). [ABSTRACT FROM AUTHOR]

The buried diagenesis and dolomitization in the deep stratum of coral reef island enable the Coral reef limestone (CRL) to exhibit a significant distinction compared with the calcified coral skeleton in mechanical properties. This paper utilized the Split Hopkinson Pressure Bar and X-ray computed tomography to investigatethe dynamic mechanical characteristics of deep CRL, revealing the effect of free water on stress–strain behaviors and the energy evolution law of CRL under high-rate impact. Furthermore, the concept of pore tensor proposed in our previous investigation was introduced to modify the dynamic constitutive model. The results indicate that the dynamic compressive strength (DCS) of CRL specimens is strongly related to stress rate, whereas the dynamic elastic modulus and DCS are independent of strain rate. The peak strains of dry specimens are significantly greater than those of saturated ones at comparative loading rates, and the viscosity effect of free water at a high loading rate impedes the deformation of CRL specimens. Moreover, the energy absorption efficiency of CRL is significantly higher than that of Bohus granite and Solnhofen limestone. Dry specimens have higher energy absorption efficiency under the same energy input, but the dynamic increase factor (DIF) of saturated specimens is larger under the same energy absorption efficiency. For the determination of DIF, this paper adopted the Back Propagation Neural Network model trained by pore tensor to predict the quasi-static compressive strength of CRL specimens, which solved the problem of considerable discreteness in mechanical parameters of CRL. The critical transition value of input energy proportion that the DIF increases rapidly is about 30%. In addition, the constitutive model modified by pore tensor reflects the influence of pore structure and water content of CRL specimens, and well describes the characteristics of dynamic elastic modulus and peak strain of deep CRL independent of loading rate. Highlights: Dynamic compressive strength of coral reef limestone strongly correlates with stress rate and weakly correlates with strain rate. The machine learning-based method for predicting quasi-static strength improves the determination accuracy of dynamic increase factor of coral reef limestone. The critical transition value of input energy proportion that the dynamic increase factor increases rapidly is about 30%. The modified constitutive model reflects the effects of pore structure and water content of coral framework limestone. [ABSTRACT FROM AUTHOR]

Medical anomaly identification using machine learning is a significant subject that has received a lot of attention. Artificial neural networks' successor, deep learning, is a well-developed technology with strong computational capabilities. Its popularity has increased in recent years due to the availability of rapid data storage and hardware parallelism. Numerous, sizeable medical imaging datasets have recently been made available to the public, which has sparked interest in the field and increased the number of research studies and publications. The main goal of this study is to give a complete theoretical examination of prominent deep learning algorithms for detecting medical anomalies. The study further presents the architecture of current methodologies, compare and contrasts training algorithms, and gives a robust assessment of current methodologies. A thorough analysis of the state-of-the-art is provided, covering the benefits and limitations associated with using open-source data, and the specifications for clinically relevant systems. This study further identifies the gaps in the body of existing knowledge and suggests future research directions. [ABSTRACT FROM AUTHOR]

In this paper, we focus on comparing the imputation performance of different deep and machine learning techniques on nine related datasets containing different missing rates ranging from 10% to 50%. Moreover, since each feature value ranges differently, such as total liability/equity ratio and earnings per share, the effect of feature normalisation on the imputation results is also examined to see whether normalising the feature values after missing value imputation can improve the prediction model performance. The experimental results show that the deep neural network technique does not necessarily perform better than the traditional machine learning technique for missing value imputation. In particular, the random forest imputation model performs the best, whereas the k-nearest neighbour method is the second best imputation model in terms of the AUC rates and type II errors. The performance in improvement of prediction models after performing feature normalisation is heavily dependent on the chosen classification technique. There is thus no need to consider the normalisation step when the random forest classifier is used. It is found that the deep neural network and support vector machine classifiers can significantly outperform those without feature normalisation. [ABSTRACT FROM AUTHOR]

Purpose: The main focus is to provide a non-similar solution for the magnetohydrodynamic (MHD) flow of Casson fluid over a curved stretching surface through the novel technique of the artificial intelligence (AI)-based Lavenberg–Marquardt scheme of an artificial neural network (ANN). The effects of joule heating, viscous dissipation and non-linear thermal radiation are discussed in relation to the thermal behavior of Casson fluid. Design/methodology/approach: The non-linear coupled boundary layer equations are transformed into a non-linear dimensionless Partial Differential Equation (PDE) by using a non-similar transformation. The local non-similar technique is utilized to truncate the non-similar dimensionless system up to 2nd order, which is treated as coupled ordinary differential equations (ODEs). The coupled system of ODEs is solved numerically via bvp4c. The data sets are constructed numerically and then implemented by the ANN. Findings: The results indicate that the non-linear radiation parameter increases the fluid temperature. The Casson parameter reduces the fluid velocity as well as the temperature. The mean squared error (MSE), regression plot, error histogram, error analysis of skin friction, and local Nusselt number are presented. Furthermore, the regression values of skin friction and local Nusselt number are obtained as 0.99993 and 0.99997, respectively. The ANN predicted values of skin friction and the local Nusselt number show stability and convergence with high accuracy. Originality/value: AI-based ANNs have not been applied to non-similar solutions of curved stretching surfaces with Casson fluid model, with viscous dissipation. Moreover, the authors of this study employed Levenberg–Marquardt supervised learning to investigate the non-similar solution of the MHD Casson fluid model over a curved stretching surface with non-linear thermal radiation and joule heating. The governing boundary layer equations are transformed into a non-linear, dimensionless PDE by using a non-similar transformation. The local non-similar technique is utilized to truncate the non-similar dimensionless system up to 2nd order, which is treated as coupled ODEs. The coupled system of ODEs is solved numerically via bvp4c. The data sets are constructed numerically and then implemented by the ANN. [ABSTRACT FROM AUTHOR]

In this study, microbial siderophore-mediated Li bioleaching from cathode active material (CAM) of spent LIBs was attempted. Pseudomonas aeruginosa was used to produce siderophores. For maximum siderophore production, favourable conditions like initial pH of Fe-free medium (5–9) and glucose (0.5–5 g/L) were assessed. Under optimal siderophore production conditions, Li bioleaching from CAM powder was studied at varying pulp densities (PD: 1–5 g/L), using bacterial cultures cultivated for 48 h and 120 h. Maximum siderophore production of ∼52 psu occurred at pH 7 and 1 g/L glucose after 48 h. Qualitative and quantitative analyses confirmed the production of mixed-type siderophores (catecholate and hydroxamate), which was further corroborated by FTIR analysis. Produced siderophores using 120 h old bacterial culture showed maximum Li extraction of >99 % at 1 g/L PD in 144 h. Artificial neural network model accurately predicted the response (R2=0.9997; RMSE=0.4955). Following the shrinking core model, kinetics analysis revealed the microbial siderophore-mediated Li bioleaching mechanism followed the diffusion of lixiviant into CAM particles and dissolution of product layer away from the particle surface, which further substantiated by XRD and FESEM analyses. This study offers an alternative bioleaching approach for sustainable Li extraction from spent LIBs employing microbial siderophores. [Display omitted] • Siderophore-mediated lithium (Li) extraction from spent Li-ion batteries was studied. • Pseudomonas aeruginosa produced mixed-type siderophores: catecholate and hydroxamate. • Maximum Li extraction of more than 99 % achieved in 144 h at 1 g/L of pulp density. • Diffusion is the rate-controlling step of Li extraction from cathode active material. • Artificial neural network model reliably predicted siderophore-mediated Li extraction. [ABSTRACT FROM AUTHOR]

Machine learning (ML) has made significant advancements in predictive modelling across many engineering sectors. However, predicting the bearing capacity of pre-bored grouted planted nodular (PGPN) piles remains a relatively unexplored area due to the complexity of the load-bearing mechanism, pile-soil interactions, and multiple variables involved. The study utilises state-of-the-art ML techniques such as extreme gradient boosting (XGBoost), random forest (RF), gradient boosting machines (GBMs), and deep learning-based simulation models. The dataset fed into the model comprises 81 case histories of static pile load tests conducted in various regions of Vietnam. The data was validated using descriptive statistics, sensitivity analysis, correlation matrix displays, SHAP plot analysis, and regression curves, with predictive performance validated through k-fold cross-validation. Among all the models tested, XGBoost (R2 = 0.91, RMSE = 0.09) and RF (R2 = 0.82, RMSE = 0.09) performed the best, while the deep neural network also yielded satisfactory results. However, GBM was found not to be a robust model for this analysis. The performance of the models was visually analysed using Violin plot comparisons and Taylor diagrams. The outcome of this study facilitates the safe and economical designs of the eco-friendly pile. [ABSTRACT FROM AUTHOR]

Although there are many sleep-related complaints in chronic obstructive pulmonary disease (COPD) patients, nocturnal leg cramps have not been adequately and extensively studied. This study fills a significant gap in the literature by determining the prevalence and influencing factors of nocturnal leg cramps in COPD patients. However, our findings also underscore the need for further research, inspiring future studies and interventions in this area. This study was conducted with a rigorous methodology, employing a comprehensive approach to evaluate the probability of experiencing nocturnal leg cramps in 215 COPD and 215 control group patients matched for age and gender. Logistic regression analysis was used, supplemented by artificial neural networks, to identify the influencing factors. This robust methodology ensures the reliability and validity of our findings. The findings of this study are not only significant but also enlightening, shedding light on the prevalence and influencing factors of nocturnal leg cramps in COPD patients. The frequency of experiencing these cramps was found to be 40.9% in chronic obstructive pulmonary patients and 21.9% in the control group (p <.05). In COPD patients, factors such as milk group food consumption, blood erythrocyte level, the cover used while sleeping, blood creatinine level, the presence of coronary artery disease, the diagnosis of diabetes mellitus, the upper mid-arm muscle area, and use of drugs with methylxanthine active ingredient methylxanthine were found to affect the occurrence of these cramps. Our findings not only call for further research but also have immediate practical implications. They highlight the crucial role of nurses in managing nocturnal leg cramps in COPD patients. By controlling patients' cold stress, the bed covers they use, and their dairy product consumption, nurses can significantly contribute to managing these cramps, thereby improving the quality of life for these patients. This underscores the importance of their role in patient care and management. [ABSTRACT FROM AUTHOR]

Alzheimer's disease (AD) patients admitted to intensive care units (ICUs) exhibit varying survival outcomes due to the unique challenges in managing AD patients. Stratifying patient mortality risk and understanding the criticality of nursing care are important to improve the clinical outcomes of AD patients. This study aimed to leverage machine learning (ML) and electronic health records (EHRs) only consisting of demographics, disease history, and routine lab tests, with a focus on nursing care, to facilitate the optimization of nursing practices for AD patients. We utilized Medical Information Mart for Intensive Care III, an open-source EHR dataset, and AD patients were identified based on the International Classification of Diseases, Ninth Revision codes. From a cohort of 453 patients, a total of 60 features, encompassing demographics, laboratory tests, disease history, and number of nursing events, were extracted. ML models, including XGBoost, random forest, logistic regression, and multi-layer perceptron, were trained to predict the 30-day mortality risk. In addition, the influence of nursing care was analyzed in terms of feature importance using values calculated from both the inherent XGBoost module and the SHapley Additive exPlanations (SHAP) library. XGBoost emerged as the lead model with a high accuracy of 0.730, area under the curve (AUC) of 0.750, sensitivity of 0.688, and specificity of 0.740. Feature importance analyses using inherent XGBoost module or SHAP both indicated the number of nursing care within 14 days post-admission as an important denominator for 30-day mortality risk. When nursing care events were excluded as a feature, stratifying patient mortality risk was also possible but the model's AUC of receiver operating characteristic curve was reduced to 0.68. Nursing care plays a pivotal role in the survival outcomes of AD patients in ICUs. ML models can be effectively employed to predict mortality risks and underscore the importance of specific features, including nursing care, in patient outcomes. Early identification of high-risk AD patients can aid in prioritizing intensive nursing care, potentially improving survival rates. [ABSTRACT FROM AUTHOR]

Impaired numerosity perception in developmental dyscalculia (low "number acuity") has been interpreted as evidence of reduced representational precision in the neurocognitive system supporting non‐symbolic number sense. However, recent studies suggest that poor numerosity judgments might stem from stronger interference from non‐numerical visual information, in line with alternative accounts that highlight impairments in executive functions and visuospatial abilities in the etiology of dyscalculia. To resolve this debate, we used a psychophysical method designed to disentangle the contribution of numerical and non‐numerical features to explicit numerosity judgments in a dot comparison task and we assessed the relative saliency of numerosity in a spontaneous categorization task. Children with dyscalculia were compared to control children with average mathematical skills matched for age, IQ, and visuospatial memory. In the comparison task, the lower accuracy of dyscalculics compared to controls was linked to weaker encoding of numerosity, but not to the strength of non‐numerical biases. Similarly, in the spontaneous categorization task, children with dyscalculia showed a weaker number‐based categorization compared to the control group, with no evidence of a stronger influence of non‐numerical information on category choice. Simulations with a neurocomputational model of numerosity perception showed that the reduction of representational resources affected the progressive refinement of number acuity, with little effect on non‐numerical bias in numerosity judgments. Together, these results suggest that impaired numerosity perception in dyscalculia cannot be explained by increased interference from non‐numerical visual cues, thereby supporting the hypothesis of a core number sense deficit. Research Highlights: A strongly debated issue is whether impaired numerosity perception in dyscalculia stems from a deficit in number sense or from poor executive and visuospatial functions.Dyscalculic children show reduced precision in visual numerosity judgments and weaker number‐based spontaneous categorization, but no increasing reliance on continuous visual properties.Simulations with deep neural networks demonstrate that reduced neural/computational resources affect the developmental trajectory of number acuity and account for impaired numerosity judgments.Our findings show that weaker number acuity in developmental dyscalculia is not necessarily related to increased interference from non‐numerical visual cues. [ABSTRACT FROM AUTHOR]

In the field of forensic anthropology, researchers aim to identify anonymous human remains and determine the cause and circumstances of death from skeletonized human remains. Sex determination is a fundamental step of this procedure because it influences the estimation of other traits, such as age and stature. Pelvic bones are especially dimorphic, and are thus the most useful bones for sex identification. Sex estimation methods are usually based on morphologic traits, measurements, or landmarks on the bones. However, these methods are time-consuming and can be subject to inter- or intra-observer bias. Sex determination can be done using dry bones or CT scans. Recently, artificial neural networks (ANN) have attracted attention in forensic anthropology. Here we tested a fully automated and data-driven machine learning method for sex estimation using CT-scan reconstructions of coxal bones. We studied 580 CT scans of living individuals. Sex was predicted by two networks trained on an independent sample: a disentangled variational auto-encoder (DVAE) alone, and the same DVAE associated with another classifier (Crecon). The DVAE alone exhibited an accuracy of 97.9%, and the DVAE + Crecon showed an accuracy of 99.8%. Sensibility and precision were also high for both sexes. These results are better than those reported from previous studies. These data-driven algorithms are easy to implement, since the pre-processing step is also entirely automatic. Fully automated methods save time, as it only takes a few minutes to pre-process the images and predict sex, and does not require strong experience in forensic anthropology. [ABSTRACT FROM AUTHOR]

Investigation of the biological sex of human remains is a crucial aspect of physical anthropology. However, due to varying states of skeletal preservation, multiple approaches and structures of interest need to be explored. This research aims to investigate the potential use of distances between bifrontal breadth (FMB), infraorbital foramina distance (IOD), nasal breadth (NLB), inter-canine width (ICD), and distance between mental foramina (MFD) for combined sex prediction through traditional statistical methods and through open-access machine-learning tools. Ethical approval was obtained from the ethics committee, and out of 100 cone beam computed tomography (CBCT) scans, 54 individuals were selected with all the points visible. Ten extra exams were chosen to test the predictors developed from the learning sample. Descriptive analysis of measurements, standard deviation, and standard error were obtained. T-student and Mann–Whitney tests were utilized to assess the sex differences within the variables. A logistic regression equation was developed and tested for the investigation of the biological sex as well as decision trees, random forest, and artificial neural networks machine-learning models. The results indicate a strong correlation between the measurements and the sex of individuals. When combined, the measurements were able to predict sex using a regression formula or machine learning based models which can be exported and added to software or webpages. Considering the methods, the estimations showed an accuracy rate superior to 80% for males and 82% for females. All skulls in the test sample were accurately predicted by both statistical and machine-learning models. This exploratory study successfully established a correlation between facial measurements and the sex of individuals, validating the prediction potential of machine learning, augmenting the investigative tools available to experts with a high differentiation potential. [ABSTRACT FROM AUTHOR]

The hysteresis model simplifies the relationship between resilience and deformation in seismic performance analysis of structural components. However, accurately capturing structural performance poses a challenge due to the intricate nature of hysteresis data. To enhance precision, this study introduces a multi-physics-guided neural network model. This innovative approach integrates explicit hysteresis models with an implicit network to create a robust hybrid model. The network architecture is grounded in physical principles, with the loss function incorporating physical parameter relationships for effective training. Through quasi-static testing, this approach demonstrates its robust adaptability and precision to the unique hysteretic curves of diverse components. [ABSTRACT FROM AUTHOR]

Similar waveforms and overlapping frequency ranges make distinguishing blasting vibrations and microseismic signals challenging, causing interference with coal mine microseismic monitoring systems. To address this problem, we propose a spectrum dataset (MSData) reflecting the spectrum features of both signal types and present a signal classification network (SCNet) combining CNNs and Transformers for signal classification. The network can learn multi-dimensional features of both signals from MSData and automatically and efficiently identify the two signal types. Experimental results yield F1-scores of 0.991 for microseismic signals and 0.993 for blasting vibration signals, meeting engineering application requirements. [ABSTRACT FROM AUTHOR]

The article focuses on using deep learning to accurately identify fetal head positions during the second stage of labor through transperineal ultrasound imaging. It highlights how an artificial intelligence model demonstrated a high accuracy rate of 94.5 percent in detecting various fetal positions, which is crucial for preventing complications such as prolonged labor and operative deliveries.

Summary: Cocoa butter is the most important ingredient of chocolate, which determines its melting behaviour. Variations in the melting characteristics of cocoa butter can profoundly affect the performance and suitability for their industrial utilisation. Over time, researchers have been attempting to establish a logical relationship between cocoa butter's unique thermal properties and the amount of saturated to unsaturated fatty acids in mono, di and triglycerides, and fatty acids (as major components), and free fatty acids, soap, primary oxidation products, minerals, moisture, phospholipids, tocopherols, unsaponifiable matters and metals (as minor components) found in cocoa butter. In this research, the thermal behaviours of thirteen samples of cocoa butter with different origins were investigated using isothermal differential scanning calorimetry. The cocoa butters starting temperature of crystallisation, temperature of maximum heat release, temperature of completed crystallisation and the enthalpy of heat release during recrystallisation were evaluated. In addition, the chemical composition (moisture, acidity, peroxide, minerals and soap content), fatty acid and triacylglycerol composition were used to establish an MLP‐ANN with fourteen input neurons connected by two flexible, sigmoid activation function layers. The back‐propagation was used to train the artificial neural network (ANN) structure and optimise the error of prediction. The study showed that the MLP algorithm can predict the thermal behaviour of CB samples with trace error, regardless of plant growth and extract process condition. [ABSTRACT FROM AUTHOR]

This article introduces a novel method for geological fault interpretation utilizing a 2D convolutional neural network approach with a focus on coalbed horizons, dealing with the problem as an image classification task. At the beginning, linear annotations, reflecting geological fault features, are applied to multiple seismic sections. By considering texture differences between fault and non‐fault areas, we construct samples that represent these distinct zones for training deep neural networks. Initially, fault annotations are transformed into single dots to facilitate pixel‐based processing. To depict a specific dot's geological structure, we employ a matrix clipped around the point, determined by a combination of range and step parameters. Convolutional layers generate filters equivalent to seismic data transformation, streamlining the need for analysis and selection of seismic attributes. The article discusses enhancing the efficiency of 2D convolutional neural network–based fault interpretation by optimizing sample selection, data extraction and model construction procedures. Through the incorporation of data from two mining areas (totalling 27.09 km2) in sample creation, the overall accuracy exceeds 0.99. Recognition extends seamlessly to unlabelled sections, showcasing the innovative technical route and methodology of fault interpretation with linear annotation and pixel‐based thinking. This study presents a method that integrates planar and raster thinking, transitioning from vision‐oriented geological structure annotation to algorithm‐oriented pixel location. The proposed 2D convolutional neural network–based matrix‐oriented fault/non‐fault binary classification demonstrates feasibility and reproducibility, offering a new automated approach for fault detection in coalbeds through convolutional neural network algorithms. [ABSTRACT FROM AUTHOR]

High-throughput technologies and machine learning (ML), when applied to a huge pool of medical data such as omics data, result in efficient analysis. Recent research aims to apply and develop ML models to predict a disease well in time using available omics datasets. The present work proposed a framework, 'OmicPredict', deploying a hybrid feature selection method and deep neural network (DNN) model to predict multiple diseases using omics data. The hybrid feature selection method is developed using the Analysis of Variance (ANOVA) technique and firefly algorithm. The OmicPredict framework is applied to three case studies, Alzheimer's disease, Breast cancer, and Coronavirus disease 2019 (COVID-19). In the case study of Alzheimer's disease, the framework predicts patients using GSE33000 and GSE44770 dataset. In the case study of Breast cancer, the framework predicts human epidermal growth factor receptor 2 (HER2) subtype status using Molecular Taxonomy of Breast Cancer International Consortium (METABRIC) dataset. In the case study of COVID-19, the framework performs patients' classification using GSE157103 dataset. The experimental results show that DNN model achieved an Area Under Curve (AUC) score of 0.949 for the Alzheimer's (GSE33000 and GSE44770) dataset. Furthermore, it achieved an AUC score of 0.987 and 0.989 for breast cancer (METABRIC) and COVID-19 (GSE157103) datasets, respectively, outperforming Random Forest, Naïve Bayes models, and the existing research. [ABSTRACT FROM AUTHOR]

Seasonal rainfall plays a vital role in both environmental dynamics and decision-making for rainfed agriculture in Ethiopia, a country often impacted by extreme climate events such as drought and flooding. Predicting the onset of the rainy season and providing localized rainfall forecasts for Ethiopia is challenging due to the changing spatiotemporal patterns and the country's rugged topography. The Climate Hazards Group InfraRed Precipitation with Station Data (CHIRPS), ERA5-Land total precipitation and temperature data are used from 1981–2022 to predict spatial rainfall by applying an artificial neural network (ANN). The recurrent neural network (RNN) is a nonlinear autoregressive network with exogenous input (NARX), which includes feed-forward connections and multiple network layers, employing the Levenberg Marquait algorithm. This method is applied to downscale data from the European Centre for Medium-range Weather Forecasts fifth-generation seasonal forecast system (ECMWF-SEAS5) and the Euro-Mediterranean Centre for Climate Change (CMCC) to the specific locations of rainfall stations in Ethiopia for the period 1980–2020. Across the stations, the results of NARX exhibit strong associations and reduced errors. The statistical results indicate that, except for the southwestern Ethiopian highlands, the downscaled monthly precipitation data exhibits high skill scores compared to the station records, demonstrating the effectiveness of the NARX approach for predicting local seasonal rainfall in Ethiopia's complex terrain. In addition to this spatial ANN of the summer season precipitation, temperature, as well as the combination of these two variables, show promising results. [ABSTRACT FROM AUTHOR]

Environmental awareness has driven a crucial search for biodegradable materials. Over the past few decades, the machining of these materials, particularly through drilling, has garnered significant research interest. This study presents an evaluation of the drilling performance of a hybrid composite (jute/glass), considering factors such as the concentration of chemical treatment with sodium bicarbonate, the type of glass fiber (random or woven), the degree of hybridization between the jute and glass fibers, and cutting parameters like rotation speed, feed rate, and drill diameter. Drilling operations were conducted to determine the maximum temperature in the cutting zone, the surface roughness of the drilled holes, and the maximum delamination factor. A Sobol sensitivity analysis was performed to identify the input parameters influencing these responses studied. Predictive modeling of the results based on the input factors was then conducted using the deep neural network (DNN) and Adaptive Neuro-Fuzzy Inference System (ANFIS) approaches. The results show that the type of glass fiber significantly affects surface roughness, while rotation speed primarily influences cutting temperature. Similarly, the drill diameter directly impacts the maximum delamination factor. Finally, the predictive modeling demonstrates that the ANN-IGWO models provide the best predictions of drilling parameters compared to the ANN-GA and ANFIS models, with mean absolute errors ranging from 1.22 to 12%. [ABSTRACT FROM AUTHOR]

Thermal errors account for a significant part of the dimensional errors of components produced by precision machine tools. These errors are commonly compensated using predictions from temperature-based empirical models. The accuracy and robustness of these predictions are affected by the locations of temperature sensors used to obtain the model's input data. Methods for sensor selection found in literature are often difficult to replicate and automate because they require tuning of several hyperparameters. This work presents a sensor and model selection approach that uses proper orthogonal decomposition (POD) and QR pivoting to select a subset of sensors that have been preinstalled in the machine tool as possible model inputs. These sensors are then sorted according to their correlation with the thermal error being modelled. The final set of inputs and thermal error model structure is chosen using Bayesian Information Criterion (BIC) to limit model overfitting. The approach was tested by modelling the Y-axis thermal error measured from air-cutting experiments performed under different spindle speeds and feed rates. This enabled determination of the structure and the choice inputs for Adaptive Neuro Fuzzy Inference System (ANFIS) and Artificial Neural Network (ANN) thermal error models. The accuracy of these models was compared to that of models trained using inputs selected by conventional approaches: the least absolute shrinkage and selection operator (LASSO), fuzzy c-means clustering (FCM), and principal component regression (PCR). The presented approach had better or comparable results to the conventional approaches while using fewer inputs. The presented approach is also well suited for automation compared to conventional approaches that require expert input. [ABSTRACT FROM AUTHOR]