检索结果-内蒙古大学图书馆

Synthetic Data-Empowered Supervised Machine Learning Modeling for a Binary Fracture Risk Assessment from 2D-DXA Analysis

学校读者我要写书评

暂无评论

SSRN 2023年

作者： Choi, Heesun Quagliato, Luca Chung, Yoon-Sok Lee, Taeyong Division of Mechanical and Biomedical Engineering Ewha Womans University Seoul03760 Korea Republic of Division of Mechanical and Biomedical Engineering Graduate Program in System Health Science and Engineering Ewha Womans University Seoul03760 Korea Republic of Department of Endocrinology and Metabolism Ajou University School of Medicine Suwon16499 Korea Republic of

Background and Objectives: Bone fracture risk assessment for osteopenia and osteoporotic patients is essential for the adoption of early countermeasures and avoiding discomfort and hospitalization. Currently employed methodologies, such as FRAX®, provides a risk assessment over a five to ten years period without showing the main variables influencing the prediction nor how they can be targeted in the short term. Thus, a lesser black-box approach where the fracture risk can be assessed in real-time from a commonly employed analysis, id est the dual-energy X-ray absorptiometry (2D-DXA), would be of help to clinicians and patients alike. Accordingly, this study presents three real-time machine learning (ML) assessment models, with distinct complexity, architectures, and performances, capable of predicting a binary fracture risk based on a femur head-hip joint 2D-DXA scan. Methods: A ~10,000 adult Korean gathered between 2017 and 2021 and composed of ~90% female and ~10% male ranging from 50 to 99 years of age was considered. 10% of the data is relevant to subjects who experienced skeletal fractures, among which 245 cases are femur fractures. The 2D-DXA analyses of the femur head-hip joint region carried out on patients allowed collecting 23 parameters, including the patient’s age, BMI, and gender, associated with one binary variable, defined in terms of non-fracture (NFX) or fracture (FX), to be employed as the training dataset for the three ML models employed and optimized in this research. The 2D-DXA results’ database (DB) was employed to train three ML classifiers with a binary (NFX/FX) output layer, namely the Extreme Gradient Boosting (XGB), the K-Nearest Neighbor (KNN), and Deep Neural Network (DNN). To avoid overfitting due to the higher number of NFX data with respect to FX one, two features augmentation techniques based on the Synthetic Minority Over-sampling Technique (SMOTE) and Oversample using Adaptive Synthetic (ADASYN) oversamplers were employed. For all

关键词： Fracture

Constructing High Efficient Carrier Channels in Double Ternary Compounds Mosse/Cdsse Heterojunction for Photoelectrochemistry and Electrocatalytic Hydrogen Production

学校读者我要写书评

暂无评论

SSRN 2024年

作者： Zhang, Yujia Yang, Han Zhang, Jun Lin, Bao Xu, Jing Yang, Shuai New Energy Materials Research Center College of Materials and Environmental Engineering Hangzhou Dianzi University Hangzhou310018 China School of Physics and Electronic Engineering Hubei University of Arts and Science Xiangyang441053 China Hubei Key Laboratory of Low Dimensional Optoelectronic Materials and Devices Hubei University of Arts and Science Xiangyang441053 China Key Laboratory of Novel Materials for Sensor of Zhejiang Province Hangzhou Dianzi University Hangzhou310018 China Department of Mechanical Engineering Hangzhou Dianzi University Hangzhou310018 China Hangzhou Dianzi University China

Here is a demonstration of constructing MoSSe/CdSSe heterojunction catalyst from bimetallic selenium sulfides. CdSSe has a high conduction band position of -0.65 eV, resulting in a strong reduction potential of photogenerated electrons. However, the high resistance of CdSSe affects the transport of charge carriers. In addition, experiments have confirmed that the presence of S and Se in MoSSe and CdSSe enhances their chemical activity and facilitates bonding. A better connection can be formed between MoSSe and CdSSe, resulting in high carrier transport efficiency. In the X-ray photoelectron spectroscopy results, the Mo and Se signals showed a large shift nearly 1 eV which confirms the strong binding between MoSSe and CdSSe. In the photoelectrochemical tests, the MoSSe/CdSSe heterojunction demonstrated a photocurrent density of 3.097 mA/cm2 at 0V (RHE) which is 2.7 and 2.9 times that of MoSSe (1.227 mA/cm2) and CdSSe (1.127 mA/cm2) respectively with a low resistance of 161 Ω. Moreover, the MoSSe/CdSSe also has good electrocatalytic hydrogen evolution performance with an overvoltage of 180 mV and Tafel slope of 69 mV/dec. Fluorescence and photocurrent response tests confirmed its higher carrier separation efficiency. This work demonstrates the application of ternary metal selenium sulfide heterojunctions in catalytic energy materials. © 2024, The Authors. All rights reserved.

关键词： Sulfur compounds

NOMAD: nonlinear manifold decoders for operator learning 22

学校读者我要写书评

暂无评论

NOMAD: nonlinear manifold decoders for operator learning

Proceedings of the 36th International Conference on Neural Information Processing Systems

作者： Jacob H. Seidman Georgios Kissas Paris Perdikaris George J. Pappas Graduate Program in Applied Mathematics and Computational Science University of Pennsylvania Department of Mechanical Engineering and Applied Mechanics University of Pennsylvania Department of Electrical and Systems Engineering University of Pennsylvania

ISBN: (纸本)9781713871088

Supervised learning in function spaces is an emerging area of machine learning research with applications to the prediction of complex physical systems such as fluid flows, solid mechanics, and climate modeling. By directly learning maps (operators) between infinite dimensional function spaces, these models are able to learn discretization invariant representations of target functions. A common approach is to represent such target functions as linear combinations of basis elements learned from data. However, there are simple scenarios where, even though the target functions form a low dimensional submanifold, a very large number of basis elements is needed for an accurate linear representation. Here we present NOMAD, a novel operator learning framework with a nonlinear decoder map capable of learning finite dimensional representations of nonlinear submanifolds in function spaces. We show this method is able to accurately learn low dimensional representations of solution manifolds to partial differential equations while outperforming linear models of larger size. Additionally, we compare to state-of-the-art operator learning methods on a complex fluid dynamics benchmark and achieve competitive performance with a significantly smaller model size and training cost.

关键词：

Comparison of Coherent and Incoherent Donut Beams for Deep Tissue STED Microscopy

学校读者我要写书评

暂无评论

Comparison of Coherent and Incoherent Donut Beams for Deep T...

Conference on Lasers and Electro-Optics (CLEO)

作者： Thariq Shanavas Robert R. McLeod Mark E. Siemens Juliet T. Gopinath Department of Physics University of Colorado Boulder CO USA Department of Electrical Computer and Energy Engineering University of Colorado Boulder CO USA Department of Physics University of Denver Denver CO Materials Science Engineering Program University of Colorado Boulder CO USA

We numerically compare the null quality for STED microscopy generated by Laguerre-Gaussian beams with orbital angular momentum and donut beams generated by incoherent addition of orthogonal Hermite Gaussian beams when... 详细信息

ISBN: (纸本)9781957171258

关键词： Biological imaging Gaussian beams Microscopy Spatial light modulators Spiral phase Step index fibers

Unveiling the Strengthening and Toughening Effects of Copper-Coated Carbon Nanotubes for the Allicu Alloy Matrix Composite

学校读者我要写书评

暂无评论

SSRN 2024年

作者： Yu, Weilin Luo, Siwei Zhou, Weiwei Yang, Yong Song, Min Chen, Biao Jia, Lei Liu, Yangzhen Yu, Zhentao Li, Wei Guo, Baisong Institute of Advanced Wear & Corrosion Resistant and Functional Materials Jinan University Guangzhou510632 China Department of Materials Processing Graduate School of Engineering Tohoku University Japan Department of Mechanical Engineering College of Engineering City University of Hong Kong Kowloon Tong Kowloon Hong Kong State Key Laboratory of Powder Metallurgy Central South University Changsha410083 China State Key Laboratory of Solidification Processing Northwestern Polytechnical University Xi’an710072 China School of Materials Science and Engineering Xi’an University of Technology Xi’an710048 China

Carbon nanotubes (CNTs) reinforced Al-Li-Cu alloy matrix composite is a promising candidate for the lightweight structural materials demanded in advanced industrial areas such as aerospace. However, how to fully exploit the combined strengthening effects of CNTs and precipitates of the Al-Li-Cu matrix is still an open issue. In the present study, the Cu coating was firstly precoated on CNTs before fabricating CNTs reinforced Al-Li-Cu alloy matrix composite. Benefiting from the Cu coating, the improved CNTs dispersion and Al-CNTs interfacial bonding was achieved, and the precipitation of Al2Cu, Al2CuLi, and Al3Li was enhanced in the AlLiCu alloy matrix, resulting in the enhanced strength-ductility synergy for Cu-coated CNTs reinforced AlLiCu alloy matrix composite. Moreover, strengthening and toughening mechanisms analyzation reveals that the Orowan strengthening of CNTs and precipitates, as well as load transfer strengthening of CNTs contribute mostly to the improved strength, while the increased ductility dominantly derives from the improved CNTs dispersion and Al-CNTs interfacial bonding, and the formation of low-density dislocation nanoregion (LDDN) in the interface area. This study not only opens up a new way for developing the AlLiCu alloy matrix composite reinforced by Cu-coated CNTs but also guides the microstructure regulation of other Al alloy matrix composites. © 2024, The Authors. All rights reserved.

关键词： Precipitation (chemical)

mechanical Cycling-Induced Evolution of Structure and Local mechanical Properties in a Pdcunip Bulk Metallic Glass

学校读者我要写书评

暂无评论

SSRN 2023年

作者： Zhang, Shuhan Wang, Xinzhe Hay, Jennifer Schwarz, Udo D. Datye, Amit Department of Mechanical Engineering and Materials Science Yale University New HavenCT06511 United States Department of Chemical and Environmental Engineering Yale University New HavenCT06511 United States KLA Inc MilpitasCA United States

To improve the ductility of bulk metallic glasses (BMGs), external mechanical stimulations can be applied. Thereby, cyclically loading the samples at stress levels below the macroscopic elastic limit leads to permanent structural changes, which can alter the plasticity in bulk metallic glasses. Here, the effect of systematic variations of [[EQUATION]] BMG’s structural state, achieved through cyclically loading the material to 70, 80, and 90% of the yield strength, on a range of mechanical properties like modulus, hardness, yield strength, strain rate sensitivity, activation volume, maximum shear stress and stiffness at different probing volume is studied. It is observed that samples rejuvenated from the non-cycled state to the 70% of yield stress state, relaxed when cycled up to 80% of yield stress, and finally rejuvenated again when cycled up to 90% of yield stress. This rejuvenation to relaxation transition might be due to the random atomic rearrangement during the cycling. © 2023, The Authors. All rights reserved.

关键词： Yield stress

Large-eddy simulation of Rayleigh-Bénard convection at extreme Rayleigh numbers

学校读者我要写书评

暂无评论

arXiv

arXiv 2022年

作者： Samuel, Roshan Samtaney, Ravi Verma, Mahendra K. Department of Mechanical Engineering Indian Institute of Technology Kanpur208016 India Mechanical Engineering Program Physical Science and Engineering Division King Abdullah University of Science and Technology Thuwal 23955 Saudi Arabia Department of Physics Indian Institute of Technology Kanpur208016 India

We adopt the stretched spiral vortex sub-grid model for large-eddy simulation (LES) of turbulent convection at extreme Rayleigh numbers. We simulate Rayleigh-Bénard convection (RBC) for Rayleigh numbers ranging from 106 to 1015 and for Prandtl numbers 0.768 and 1. We choose a box of dimensions 1:1:10 to reduce computational cost. Our LES yields Nusselt and Reynolds numbers that are in good agreement with the direct-numerical simulation (DNS) results of Iyer et al.1, albeit with a smaller grid size and at significantly reduced computational expense. For example, in our simulations at Ra = 1013, we use grids that are 1/120 times the grid-resolution as that of the DNS1. The Reynolds numbers in our simulations span 3 orders of magnitude from 1,000 to 1,700,000. Consistent with the literature, we obtain scaling relations for Nusselt and Reynolds numbers as Nu ∼ Ra0.317 and Re ∼ Ra0.497. We also perform LES of RBC with periodic side-walls, for which we obtain the corresponding scaling exponents as 0.343 and 0.477 respectively. Our LES is a promising tool to push simulations of thermal convection to extreme Rayleigh numbers, and hence enable us to test the transition to ultimate convection regime. © 2022, CC BY.

关键词： Large eddy simulation

Accelerate the Design of New Superhard Carbon Allotropes in the Pca21 Space Group: High-Throughput Screening and Machine Learning Strategies

学校读者我要写书评

暂无评论

SSRN 2023年

作者： Fan, Qingyang Min, Gege Liu, Li Zhao, Yingbo Yu, Xinhai Yun, Sining College of Information and Control Engineering Xi’an University of Architecture and Technology Xi’an710055 China School of Mechanical and Electrical Engineering Xi’an University of Architecture and Technology Xi’an710055 China Department of Mechanical and Electrical Engineering Hetao College Bayannur Inner Mongolia015000 China School of Materials Science and Engineering Xi’an University of Architecture and Technology Xi’an710055 China

As an efficient calculation and screening method, high-throughput methods can discover and optimize new materials and shorten the development cycle and cost of new materials. However, using high throughput for material screening and computation, a large amount of computing resources and storage space are indispensable. To accelerate the design of novel superhard carbon materials, we combined machine learning methods with high-throughput computing to construct three machine learning models: support vector machine regression, random forests, and artificial neural networks, and mined data from existing material databases to select 1,276 structures as datasets for the model to predict the volume modulus and shear modulus. Through comparative analysis, the optimal model was selected to predict the bulk modulus and shear modulus of the structure obtained through high-throughput calculations, and the prediction results of the model were verified by density functional theory calculations. Finally, 8 superhard carbon allotropes were found in 49 Pca21 space groups. © 2023, The Authors. All rights reserved.

关键词： Elastic moduli

An Experimental-Computational Approach for Measuring Charged Solute Diffusivity Through Human Synovium

学校读者我要写书评

暂无评论

SSRN 2025年

作者： Davis, Alexandra Lenamarie Vaidyanathan, Ashish Sachie, Sarah Owusu Rohanifar, Milad Akanda, Shamimur Rahman Setton, Lori A. Department of Biomedical Engineering Washington University in St. Louis St. LouisMO United States Department of Orthopedic Surgery Washington University School of Medicine St. LouisMO United States Department of Mechanical Engineering and Materials Science Washington University in St. Louis St. LouisMO United States

Intra-articular drug delivery has shown promise in targeting arthritic joints, but its therapeutic efficacy is hindered by the synovium, a multilayered connective tissue that rapidly clears locally delivered drugs from the joint space. To better understand the mechanisms behind synovial drug clearance, we previously developed a finite element model of synovium as a multiphasic tissue and used an inverse method to determine the effective diffusivity (Deff) of neutral solutes through synovium, which was found to decrease with increasing molecular weight. Here, we adapted this experimental-computational approach to measure Deff of charged dextrans through human synovium. The fixed charge density of synovium was found to be negligible and orders of magnitude lower than that of other soft tissues, and Deff was significantly affected by charge, molecular weight, and the interaction between them. According to FEM predictions and single exponential fitting of experimental data, Deff and t1/2 of cationic dextrans were higher and lower, respectively, than their anionic and neutral counterparts. Apart from cationic dextrans, 4 kDa dextrans diffused through synovium faster than 20 kDa dextrans as expected. These data are among the first to explore charged solute-matrix interactions in synovium and will guide future experimental and computational studies on charged drug transport. © 2025, The Authors. All rights reserved.

关键词： Tissue