Model Predictions Research Articles

J-PLUS: Bayesian object classification with a strum of BANNJOS

With its 12 optical filters, the Javalambre-Photometric Local Universe Survey (J-PLUS) provides an unprecedented multicolor view of the local Universe. The third data release (DR3) covers 3,192 deg$^2$ and contains 47.4 million objects. However, the classification algorithms currently implemented in the J-PLUS pipeline are deterministic and based solely on the morphology of the sources. Our goal is to classify the sources identified in the J-PLUS DR3 images as stars, quasi-stellar objects (QSOs), or galaxies. For this task, we present a machine learning pipeline that utilizes Bayesian neural networks to provide the full probability distribution function (PDF) of the classification. has been trained on photometric, astrometric, and morphological data from J-PLUS DR3 DR3, and using over 1.2 million objects with spectroscopic classification from DR18 DR9, the Early Data Release, and DR3. Results were validated on a test set of about $1.4 10^5$ objects and cross-checked against theoretical model predictions. outperforms all previous classifiers in terms of accuracy, precision, and completeness across the entire magnitude range. It delivers over $95<!PCT!>$ accuracy for objects brighter than $r = 21.5$ mag and $ 90<!PCT!>$ accuracy for those up to $r = 22$ mag, where J-PLUS completeness is $ 25<!PCT!>$. is also the first object classifier to provide the full PDF of the classification, enabling precise object selection for high purity or completeness, and for identifying objects with complex features, such as active galactic nuclei with resolved host galaxies. effectively classified J-PLUS sources into around 20 million galaxies, one million QSOs, and 26 million stars, with full PDFs for each, which allow for later refinement of the sample. The upcoming J-PAS survey, with its 56 color bands, will further enhance 's ability to detail the nature of each source.

CHEMOUT: CHEMical complexity in star-forming regions of the OUTer Galaxy. IV. ALMA observations of organic species at a galactocentric radius of ~ 23 kpc

Single-dish observations suggest that the abundances of organic species in star-forming regions of the outer Galaxy, which are characterised by sub-solar metallicities, are comparable to those found in the local Galaxy. To understand this counter-intuitive result and avoid a misleading interpretation due to beam dilution effects at these large distances, spatially resolved molecular emission maps are needed to correctly link the measured abundances and local physical properties. We observed several organic molecules with the Atacama Large Millimeter Array towards WB89-671, the source with the largest galactocentric distance (23.4 kpc) of the project CHEMical complexity in star-forming regions of the OUTer Galaxy (CHEMOUT) at a resolution of $ 15000$ au. We compared the observed molecular abundances with chemical model predictions. We detected emission of C$_4$H HCO HCS$^+$, CS, HN13C, and SO. The emission morphology is complex, extended, and different in each tracer. In particular, the most intense emission in and arises from two millimeter-continuum infrared-bright cores. The most intense and SO emission predominantly arises from the part of the filament that lacks continuum sources. The narrow line widths across the filament indicate quiescent gas in spite of the two embedded protostars. The derived molecular column densities are comparable with those in local star-forming regions, and they suggest an anti-correlation between hydrocarbons, ions, HCO, and on the one hand, and and SO on the other. The static chemical models that match the observed column densities best favour low-energy conditions that are expected at large galactocentric radii, but they also favour carbon elemental abundances that exceed those derived by extrapolating the C/H galactocentric gradient at 23 kpc by three times. This would indicate a flatter C/H trend at large galactocentric radii, which is in line with a flat abundance of organics. However, to properly reproduce the chemical composition of each region, models should include dynamical evolution.

Physiologically based pharmacokinetic modeling of long-acting extended-release naltrexone in pregnant women with opioid use disorder.

Opioid use disorders (OUD) are a major issue in the U.S. Current treatments for pregnant women, like methadone and buprenorphine require daily dosing and have adverse effects. Monthly injectable naltrexone (XR-NTX) mitigates these adverse effects but is not recommended during pregnancy due to limited pharmacokinetic and safety data. This study developed a physiologically based pharmacokinetic (PBPK) model to describe XR-NTX pharmacokinetics during pregnancy, and to predict dosing recommendations. Model predictions were successfully validated with observed data. Maternal plasma XR-NTX profiles were simulated for 400 non-pregnant virtual females at the approved dose of 380 mg, then randomized to continue with either 380, 285, 190, or 95 mg during pregnancy. The non-pregnant virtual females had a mean predicted Cmax, AUC0-7days, and AUC0-28days of 23.3 ng/mL, 142 ng·d/mL, and 148 ng·d/mL, respectively. Maternal XR-NTX exposure (AUC0-28days) were predicted to increase by 1.37, 1.43, and 1.72 times during the first, second, and third trimester of pregnancy. However, the fetal-to-maternal exposure (AUC0-28days) was lower in the first (15%), second (7%), and third (9%) trimesters. A dose of 285 mg of XR-NTX in pregnancy during the first/second trimester and dose of 190 mg in the third trimester were predicted to provide maternal exposures that were comparable to non-pregnant levels at the standard dose. This study provides crucial insights into XR-NTX pharmacokinetics and proposes a dosing strategy during pregnancy, potentially aiding further clinical investigations and decision making regarding XR-NTX use during pregnancy.

Experimental and Modeling Analysis of the Tensile Properties of Heavy-Duty Coatings for Steel Structures

Coatings are essential for protecting steel structures from corrosion and mechanical stresses, especially under challenging environmental conditions. To this end, this study systematically examines the effects of temperature (20 °C to 50 °C), strain rate (6.67 × 10−4 s−1 to 1.67 × 10−2 s−1), and intermediate coat thickness (140 μm to 700 μm, the layer between the primer and topcoat) on the uniaxial tensile properties of heavy-duty coatings for steel structures. Experimental and theoretical analyses were conducted to quantitatively assess the influence of these factors on the mechanical properties of the coatings. A multifactor constitutive model was developed based on the Sherwood–Frost model by integrating material characteristics and fitting experimental data, incorporating response functions for temperature, strain rate, and intermediate coat thickness. The results reveal that increased temperature causes temperature-induced softening, while higher strain rates lead to strain rate-dependent strengthening of the coatings. In contrast, the effect of layer thickness on mechanical properties follows a non-monotonic trend, influenced by the structural and material characteristics of the coatings, with the most significant mechanical response occurring at 560 μm thickness. These findings suggest that optimal coating design must consider multiple factors to enhance mechanical performance. Additionally, the correlation coefficients (r) between the model predictions and experimental results are 0.97 or higher, indicating the model’s effectiveness in predicting and optimizing the mechanical performance of heavy-duty coatings under complex conditions.

Impacts of age and BMI on vancomycin model choice in a Bayesian software: Lessons from a very large multi-site retrospective study.

Model-informed precision dosing (MIPD) optimizes drug doses based on pharmacokinetic (PK) model predictions, necessitating careful selection of models tailored to patient characteristics. This study evaluates the predictive performance of various vancomycin PK models across diverse age and BMI categories, drawing insights from a large multi-site database. Adults receiving vancomycin intravenous therapy at United States health systems between January 1, 2022, and December 31, 2023, were included. Patient demographics, vancomycin administration records, and therapeutic drug monitoring levels (TDMs) were collected from the InsightRX database. Age and body mass index (BMI)-based subgroups were formed to assess model performance, with predictions made iteratively. The optimal model for each age-BMI subgroup was chosen based on predefined criteria: models were filtered for mean percentage error (MPE) ≤ 20% and normalized root mean squared error (RMSE) < 8 mg/L, and then the most accurate among them was selected. A total of 384,876 treatment courses across 155 US health systems were analyzed, contributing 841,604 TDMs. Eleven models were compared, showing varying accuracy across age-BMI categories (41%-73%), with higher accuracy observed once TDMs were available for Bayesian estimates of individual PK parameters. Models performed more poorly in younger adults compared to older adults, and the optimal model differed depending on age-BMI categories and prediction methods. Notably, in the a priori period, the Colin model performed best in adults aged 18-64 years across most BMI categories; the Goti/Tong model performed best in the older, non-obese adults; and the Hughes model performed best in many of the obese categories. Our study identifies specific vancomycin PK models that demonstrate superior predictions across age-BMI categories in MIPD applications. Our findings underscore the importance of tailored model selection for vancomycin management, especially highlighting the need for improved models in younger adult patients. Further research into the clinical implications of model performance is warranted to enhance patient care outcomes.

High- accuracy chaotic time series prediction of the flexible beam-ring model based on P$$\_$$CNN-BiLSTM ED network

High- accuracy chaotic time series prediction of the flexible beam-ring model based on P$$\_$$CNN-BiLSTM ED network

Nonintrusive identification of boiling regimes enabled by deep learning based on flow boiling acoustics

Monitoring two-phase cooling systems and identifying corresponding boiling regimes are vital to avoid device failure owing to thermal runaway. Boiling acoustic characteristics can be used to develop a nonintrusive monitoring tool because of its close relation to the bubble behavior. Compared with bubble image-based techniques, acoustic sensing is cheaper and easier to implement. However, the acoustic data available is very limited, especially for flow boiling. Also, real-time strategies are still lacking since the prediction models based on boiling acoustics are usually offline. Additionally, noise interference typically degrades model prediction performance. To promote the implementation of an automated boiling system amidst these challenges, we present a deep learning framework for real-time identification of boiling regimes using acoustic signals of flow boiling. Flow boiling experiments are first performed using HFE7100 under different conditions and then combined with deionized water data from our previous experiments to construct the acoustic dataset. We use the corresponding acoustic spectrograms to train and test a convolutional neural network which shows a prediction accuracy of 99.97 % without noise interference. A transfer learning strategy is further employed to improve the original model performance under noise interference. Finally, we add a decision logic to adapt the CNN model to acoustic data streams for real-time identification of boiling regimes. The CNN model with the decision logic can accurately identify the boiling regimes in real-time by scanning the corresponding spectrograms irrespective of working fluids and conditions. In practice, our strategy can be utilized to initiate the necessary precautions to avoid overheating-induced failure.

Integration of FAERS, DrugBank and SIDER Data for Machine Learning-based Detection of Adverse Drug Reactions

AbstractTraditionally, disproportionality analysis (DPA) methods are employed for signal detection in pharmacovigilance, but these methods utilize only a limited portion of the data available from spontaneous event reports (SERs). This research aims to enhance signal detection by applying machine learning (ML) methods that can leverage additional data. We create a dataset by integrating SER data from the FDA Adverse Event Reporting System (FAERS) with biological and chemical data from DrugBank, and information on known adverse drug reactions (ADRs) from Side Effect Resource (SIDER). The known ADRs from SIDER are used to label the dataset for ML training. Using the AutoML library TPOT, ML models are trained on this dataset. Our findings indicate that ML models, even when trained with the same features as DPA methods, achieve higher recall and precision. Moreover, incorporating additional features related to drugs and events significantly boosts the performance of ML models. Analysis using the explainable AI (XAI) technique SHAP reveals that the drug name, event name, and fifth-level ATC code are the most influential features for model predictions. These ML models offer a promising alternative or supplement to conventional DPA methods for signal detection in pharmacovigilance.

Global fit to b→cτν anomalies as of Spring 2024

Recently, several new experimental results of the test of lepton flavor universality (LFU) in B→D(*) semileptonic decays were announced: the first result of RD from the LHCb run 1 data, the first results of RD and RD* from the LHCb run 2 data, and the first result of RD* from the Belle II collaboration. Including these new data, a global analysis still prefers the violation of the LFU between the tau and light leptons. A new world average of the data from the , LHCb, Belle, and Belle II collaborations is RD=0.342±0.026 and RD*=0.287±0.012. Including this new data, we update a circumstance of the b→cτν¯ measurements and their implications for new physics (NP). Incorporating recent developments for the B→D(*) form factors in the Standard Model, we observe a 4.4σ deviation from the Standard Model prediction. Our updates also include model-independent NP formulas for the related observables and the global fittings of parameters for leptoquark scenarios as well as single NP operator scenarios. Furthermore, we show future potential to indirectly distinguish different NP scenarios with the use of the precise measurements of the polarization observables in B→D(*)τν¯ at the Belle II and the high-pT flavored-tail searches at the LHC. We also discuss an impact on the LFU violation in ϒ→l+l−. Published by the American Physical Society 2024

Two-species model for nonlinear flow of wormlike micelle solutions. Part II: Experiment

Applications often expose wormlike micelle solutions to a very wide range of shear and temperature conditions. The two-species model presented in Part I [Salipante et al., J. Rheol. 68 (2024)] describes the nonlinear rheology over a wide range of shear rates. Here, we compare the model predictions to measurements using a combination of microcapillary and rotational rheology to measure the viscosity of surfactant solutions across seven decades of shear rate and five decades of viscosity. The effect of temperature is studied between 20 and 60 °C for different surfactant concentrations. Model parameters are determined from both small-amplitude shear measurements and fitting to the nonlinear data. Under shear stress, the model predicts due to hindered combination kinetics that the average micelle length decreases from several micrometers to a few hundred nanometers. At sufficiently high stress, the micelle shear rheology exhibits a transition from entangled wormlike behavior to a dilute rod rheology in agreement with the model. Transient stress-growth measurements exhibit a large overshoot, which is rather well predicted by the model with hindered combination rate. Microcapillary flow birefringence also is adequately predicted by the model, confirming the accuracy of its predicted micelle lengths and exhibiting a marked change in stress-optic response at the transition between entangled polymers and dilute rods. The relaxation of retardance after flow cessation follows model predictions that include micelle-micelle interactions, which are sensitive to the rotational diffusivity and length. These methods can be applied broadly to explore relationships between composition and performance.

Corrosion prediction model for long-distance pipelines based on NLFE-NGO-ELM

Abstract To improve the accuracy of buried pipeline corrosion rate prediction, we developed a Northern Goshawk Optimization-based Extreme Learning Machine (NLFE-NGO-ELM) model enhanced by Non-linear Feature Expansion. Feature selection was used to expand the small dataset and reduce the dimensionality of factors affecting pipeline corrosion. The Northern Goshawk algorithm was used, and 106 sets of pipeline external corrosion detection data were simulated in MATLAB as a case study. The ELM optimizer was analyzed using Grey Wolf Optimization (GWO-ELM), Northern Goshawk Optimization (NGO-ELM), and BP optimization as comparison models. Using NLFE-NGO-ELM significantly improved the model's prediction accuracy and speed. The NGO-ELM model was applied to predict the corrosion rates of 21 pipelines, achieving an absolute error of only 2.40% and an R value of 0.97255, surpassing the predictive performance of NGO-BP and GWO-ELM models. These results indicate that the current model exhibits superior learning performance and stronger fitting capabilities.

Ethnicity Detection with Convolutional Neural Network (CNN): Bangladesh Perspectives

Ethnicity detection, or the automatic determination of people's ethnic backgrounds using visual data, has become important in a number of disciplines, including social sciences, demographics, and computer vision. With a focus on Bangladesh's complex cultural landscape, we describe in this paper a novel method for ethnicity detection using convolutional neural networks (CNNs). We gathered a dataset of photographs depicting members of the three main ethnic groups living in the Chittagong hill neighborhood: Chakma, Marma, and Tripura. Utilizing the DenseNet121 model's robust feature extraction skills, we customized the architecture to meet the needs of our particular ethnicity detection application. A wide range of performance indicators were used to train and assess our unique CNN model. The outcomes illustrate accurate ethnicity detection technology's potential for identifying and resolving social inequities while showcasing a promising level of accuracy. To address possible biases in the model's prediction, ethical considerations are also covered. Overall, this work advances knowledge of Bangladesh's ethnic variety and demonstrates the potential of CNNs for ethnicity detection. The results provide up new avenues for investigation and applications that advance a more diverse and inclusive society.

Natural Seawater Impact on Crack Propagation and Fatigue Behavior of Welded Nickel Aluminum Bronze

ABSTRACTNickel aluminum bronze (NAB) is a complex alloy used extensively in the marine environment. Fatigue strength of NAB is reduced by welding and prior seawater corrosion. This study investigated the combined effect of corrosion and plasma welding on the fatigue behavior of NAB. Natural seawater corroded samples were used in tension‐tension cyclic loading tests to observe fatigue crack initiation, propagation, and failure. Fatigue cracks initiated from corrosion pits at the weld toe. Stress corrosion and fatigue cracks propagated along the path of β′ and κIII phases. A short crack growth model (SCGM) predicted fatigue strength using experimentally obtained material properties and corrosion pit dimensions. Model predictions were used to develop S‐N curves and were within 30% of experimental results. The SCGM produced accurate and reliable fatigue life results that could be applied by industry to aid in revalidation decision making and inspection scheduling.

Interpretable and Physicochemical-Intuitive Deep Learning Approach for the Design of Thermal Resistance of Energetic Compounds.

Thermal resistance of energetic materials is critical due to its impact on safety and sustainability. However, developing predictive models remains challenging because of data scarcity and limited insights into quantitative structure-property relationships. In this work, a deep learning framework, named EM-thermo, was proposed to address these challenges. A data set comprising 5029 CHNO compounds, including 976 energetic compounds, was constructed to facilitate this study. EM-thermo employs molecular graphs and direct message-passing neural networks to capture structural features and predict thermal resistance. Using transfer learning, the model achieves an accuracy of approximately 97% for predicting the thermal-resistance property (decomposition temperatures above 573.15 K) in energetic compounds. The involvement of molecular descriptors improved model prediction. These findings suggest that EM-thermo is effective for correlating thermal resistance from the atom and covalent bond level, offering a promising tool for advancing molecular design and discovery in the field of energetic compounds.

Model‐ versus data‐uncertainty for concrete members and connections in cyclic loading

AbstractLarge databases of cyclic tests on flexure‐ or shear‐critical concrete members and shear‐critical connections of columns to beams or slabs are used to estimate the uncertainty inherent in experimental data in literature—as read by users. To this end, the predictions of two different, presumably independent, design‐oriented models for the properties of interest are used to establish the “central tendency” of data, against which individual tests or small groups thereof are assessed. Properties considered are: (a) the cyclic ultimate chord‐rotation of flexure‐controlled members with continuous or lap‐spliced deformed bars, (b) the cyclic shear strength of shear‐critical members, (c) the chord‐rotation at yielding of rectangular columns with plain bars, and (d) the cyclic shear strength of shear‐controlled beam‐column and slab‐column joints. Results suggest that the data from each test campaign have a certain degree of bias, specific to it. Test campaigns with ratio of estimated average deviation from the “central tendency” to the standard deviation of campaign deviations (called “data uncertainty”) which is far into the tail of the Normal distribution may be excluded as questionable. This systematic bias, along with other types of “data uncertainty” addressed in this work, seem to contribute to the apparent scatter of model predictions with respect to cyclic test results the equivalent of a coefficient of variation of model‐to‐test‐ratio of at least 10% and possibly as high as 25%–30%. Model uncertainty seems to contribute to this scatter the equivalent of a coefficient of variation of at least 15% in shear‐controlled connections, or as much as 25% in the case of flexural deformation capacity of members with deformed bars; the cyclic shear resistance of members and—with the reservation of the small number of tests—the chord‐rotation at yielding of members with plain bars, are in‐between.

Role of protein-protein interactions on organization and dynamics of a model chromatin.

The three-dimensional organization of chromatin is influenced by chromatin-binding proteins through both specific and non-specific interactions. However, the roles of chromatin sequence and the interactions between binding proteins in shaping chromatin structure remain elusive. By employing a simple polymer-based model of chromatin that explicitly considers sequence-dependent protein binding and protein-protein interactions, we elucidate a mechanism for chromatin organization. We find that tuning protein-protein interactions and protein concentration is sufficient to either promote or inhibit chromatin compartmentalization. Moreover, chromatin sequence and protein-protein attraction strongly affect the structural and dynamic exponents that describe the spatiotemporal organization of chromatin. Strikingly, our model's predictions for the exponents governing chromatin structure and dynamics successfully capture experimental observations, in sharp contrast to previous chromatin models. Overall, our findings have the potential to reinterpret data obtained from various chromosome conformation capture technologies, laying the groundwork for advancing our understanding of chromatin organization.

Research on Dynamic Mechanical Behavior of Reinforcing Steel Under Ultrahigh Strain Rate

The strain rate effect significantly influences the dynamic mechanical behavior of reinforcing steel under impact loads. To investigate the effect of strain rate on the dynamic mechanical behavior of reinforcing steel, this study conducts dynamic compression tests on four types of reinforcing steel using a Split Hopkinson pressure bar. The results show that there is a significant strain rate effect on the reinforcing steel when the strain rate reaches 103–104 s−1. The experimental results are analyzed using the Johnson–Cook model (J–C model), and the original model is modified based on the error between the model predictions and the experimental results. The results indicate that the modified J–C model, which includes an enhanced strain rate effect term, better predicts the dynamic mechanical behavior of reinforcing steel at ultrahigh strain rates.

Numerical Analysis and Safety Assessment of the Former Site of Shanghai South Downtown Library Reinforced by U-Steel During Integral Translation

ABSTRACT The utilization of overall translation techniques in buildings offers numerous advantages, including cost-effectiveness, shortened construction timelines, and environmental sustainability. These techniques are increasingly being employed for the conservation and renovation of existing historic structures. Taking the former site project of the South Downtown Library in Shanghai, which requires the overall translation due to the requirements of urban renewal, as an example, a systematic detection and identification of the damage and safety states of the overall structure were conducted. The damaged condition and overall structural safety were evaluated, and the corresponding reinforcement and repair suggestions were proposed. The former site of the library, a brick-timber mixed structure, was reinforced using U-steel through internal support and external pulling. The structure was also lifted and translated using the overall translation technique. The displacement responses of the key parts of the structure were monitored using the displacement sensors during the translation. The detailed finite element model of the former site of the Shanghai south downtown library, strengthened by U-steel, was established using the ABAQUS. The displacement responses of the structure’s key components during the jacking and translation were analyzed, and compared with the actual monitoring results. Good agreements were observed between the model predictions and test results. Utilizing the validated finite element model, parametric analyses were carried out. The effects of the rate of jacking and translation as well as the ground flatness on the displacement responses of a structure were analyzed, the mechanical and safety states of the structure reinforced with U-steel were further evaluated. This study can provide a reference for the overall translation technique and safety assessment of similar historic buildings in the urban renewal processes.

Development of optimized ensemble machine learning-based prediction models for wire electrical discharge machining processes

This paper proposes development of optimized heterogeneous ensemble models for prediction of responses based on given sets of input parameters for wire electrical discharge machining (WEDM) processes, which have found immense applications in many of the present-day manufacturing industries because of their ability to generate complicated 2D and 3D profiles on hard-to-machine engineering materials. These ensembles are developed combining predictions of the three base models, i.e. random forest, support vector machine and ridge regression. These three base models are first framed utilizing the training datasets, providing predictions for all the responses under consideration. Based on these predictions, two optimization problems are formulated for each of the responses, while minimizing root mean squared error and mean absolute error, for subsequent development of two optimized ensembles whose predictions are the weighted sum of the predictions of the base models. The prediction performance of all the five models is ascertained through nine statistical metrics, after which a cumulative quality loss-based multi-response signal-to-noise (MRSN) ratio for each model is computed, for each of the responses, where a higher MRSN ratio indicates greater accuracy in prediction. This study is conducted using two experimental datasets of WEDM process. Overall, the optimized ensemble models having higher MRSN ratios than the base models are indicated to deliver better prediction accuracy.

Novel hybrid data-driven models for enhanced renewable energy prediction

Global power grid management depends on accurate solar energy estimation, yet present prediction techniques frequently suffer from unreliability as a result of abnormalities in solar energy data. Solar radiation projections are affected by variables such as anticipated horizon length, meteorological classification, and power measuring techniques. Therefore, a Solar Wind Energy Prediction System (SWEPS) is proposed as a solution to these problems. It improves renewable energy projections by taking sun trajectories and atmospheric characteristics into account. In addition to using a variety of optimization methods and pre-processing techniques, such as Principal Component Analysis (PCA), Recursive Feature Elimination (RFE), Least Absolute Shrinkage Selection Operator (LASSO), and recursive feature addition processes (RFA), complemented by a genetic algorithm for feature selection (GAFS). The SWEPS also makes use of sophisticated machine learning algorithms and Statistical Correlation Analysis (SCA) to find important connections. Neural Network Algorithms (NNA) and other metaheuristic techniques like Cuckoo Search Optimization (CSO), Social Spider Optimization (SSO), and Particle Swarm Optimization (PSO) are adopted in this work to increase the predictability and accuracy of models. Utilizing the strengths of machine learning and deep learning techniques (Artificial Neural Networks (ANN), Decision Trees, Support Vector Machine (SVM), Recurrent Neural Networks (RNN), and Long Short Term Memory (LSTM)) for robust forecasting, as well as meta-heuristic optimization techniques to fine-tune hyper-parameters and achieve near-optimal values and significantly improve model performance, are some of this work contributions to the development of a comprehensive prediction system.