检索结果-南通市图书馆

Generalized Embedding Machines for Recommender Systems

在线全文

学校读者我要写书评

暂无评论

Machine Intelligence Research 2024年第3期21卷 571-584页

作者： Enneng Yang xin xin Li shen Yudong Luo Guibing Guo Software College Northeastern UniversityShenyang110000China School of Computer Science and Technology Shandong UniversityQingdao266000China JD Explore Academy JD Explore AcademyBeijing100000China

Factorization machine (FM) is an effective model for feature-based recommendation that utilizes inner products to capture second-order feature interactions. However, one of the major drawbacks of FM is that it cannot capture complex high-order interaction signals. A common solution is to change the interaction function, such as stacking deep neural networks on the top level of FM. In this work, we propose an alternative approach to model high-order interaction signals at the embedding level, namely generalized embedding machine (GEM). The embedding used in GEM encodes not only the information from the feature itself but also the information from other correlated features. Under such a situation, the embedding becomes high-order. Then we can incorporate GEM with FM and even its advanced variants to perform feature interactions. More specifically, in this paper, we utilize graph convolution networks (GCN) to generate high-order embeddings. We integrate GEM with several FM-based models and conduct extensive experiments on two real-world datasets. The results demonstrate significant improvement of GEM over the corresponding baselines.

关键词： Feature interactions high-order interaction factorization machine(FM) recommender system graph neural network(GNN)

High-precision X-ray polarimeter based on channel-cut crystals

维普期刊数据库评论

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Nuclear Science and Techniques 2024年第3期35卷 30-39页

作者： Shang-Yu Si Zhong-Liang Li Wen-hong Jia Lian Xue hong-xin Luo Jian-Cai Xu Bai-Fei shen Lin-Gang Zhang Liang-Liang Ji Yu-xin Leng Ren-Zhong Tai Shanghai Institute of Applied Physics Chinese Academy of SciencesShanghai 201800China University of Chinese Academy of Sciences Beijing 100049China Shanghai Synchrotron Radiation Facility Shanghai Advanced Research InstituteChinese Academy of SciencesShanghai 201204China Shanghai Institute of Optics and Fine Mechanics Chinese Academy of SciencesShanghai 201800China

We report on using synthetic silicon for a high-precision X-ray polarimeter comprising a polarizer and an analyzer,each based on a monolithic channel-cut crystal used at multiple Brewster reflections with a Bragg angle very close to 45°.Experiments were performed at the BL09B bending magnet beamline of the Shanghai Synchrotron Radiation Facility using a Si(800)crystal at an X-ray energy of 12.914 keV.A polarization purity of 8.4×10^(-9)was *** result is encouraging,as the measured polarization purity is the best-reported value for the bending magnet ***,this is the firstly systematic study on the hard X-ray polarimeter in China,which is crucial for exploring new physics,such as verifying vacuum birefringence.

关键词： X-ray polarimeter Vacuum birefringence Synchrotron radiation Channel-cut crystal

Modeling ocean wave propagation under sea ice covers

在线全文

学校读者我要写书评

暂无评论

Acta Mechanica Sinica 2015年第1期31卷 1-15页

作者： xin Zhao Hayley H. shen Sukun Cheng Department of Civil and Environmental Engineering Clarkson University

Operational ocean wave models need to work globally, yet current ocean wave models can only treat ice covered regions crudely. The purpose of this paper is to provide a brief overview of ice effects on wave propagation and different research methodology used in studying these effects. Based on its proximity to land or sea, sea ice can be classified as： landfast ice zone, shear zone, and the marginal ice zone. All ice covers attenuate wave energy. Only long swells can penetrate deep into an ice cover. Being closest to open water, wave propagation in the marginal ice zone is the most complex to model. The physical appearance of sea ice in the marginal ice zone varies. Grease ice, pancake ice, brash ice, floe aggregates, and continuous ice sheet may be found in this zone at different times and locations. These types of ice are formed under different thermalmechanical forcing. There are three classic models that describe wave propagation through an idealized ice cover： mass loading, thin elastic plate, and viscous layer models. From physi cal arguments we may conjecture that mass loading model is suitable for disjoint aggregates of ice floes much smaller than the wavelength, thin elastic plate model is suitable for a con tinuous ice sheet, and the viscous layer model is suitable for grease ice. For different sea ice types we may need different wave ice interaction models. A recently proposed viscoelas tic model is able to synthesize all three classic models into one. Under suitable limiting conditions it converges to the three previous models. The complete theoretical framework for evaluating wave propagation through various ice covers need to be implemented in the operational ocean wave models. In this review, we introduce the sea ice types, previous wave ice interaction models, wave attenuation mechanisms,the methods to calculate wave reflection and transmission between different ice covers, and the effect of ice floe breaking on shaping the sea ice morphology. Labora

关键词： Ocean waves Sea ice Cryospheric change

Effect of surface roughness element on near wall turbulence with zero-pressure gradient

在线全文

学校读者我要写书评

暂无评论

Science China(Physics,Mechanics & Astronomy) 2015年第6期58卷 95-102页

作者： ZHANG xin PAN Chong shen JunQi WANG JinJun School of Aeronautic Northwestern Polytechnical University Fluid Mechanics Key Laboratory of Ministry of Education Beijing University of Aeronautics and Astronautics

In the present paper,we present an investigation on the effect of roughness elements onto near-wall kinematics of a zeropressure-gradient turbulent boundary *** array of spanwisely-aligned cylindrical roughness elements was attached to the wall surface to regulate the near-wall low-speed streaky *** both qualitative visualization and quantitative measurement,we found that the regularization only occurs in the region below the height of the roughness *** analysis on the probability distribution of the streak spanwise spacing showed that the mean spanwise streak spacing is dominated by the roughness elements;however,the latter's effect is in competition with the intrinsic streak generation mechanisms of smooth wall *** the top of the roughness elements,local streamwise turbulent fluctuation intensity can be reduced by about 10%.We used POD analysis to depict such regularization effect in terms of near-wall structure *** further found that if the spanwise spacing of roughness elements increased to be larger than the mean streak spacing in the smooth wall turbulence,there is no streak-regularization effect in the buffer region,so that the near-wall streamwise turbulent fluctuation intensity doesn't reduce.

关键词： turbulent coherent structures low-speed streak flow control

Driving mechanisms of nitrogen transport and transformation in lacustrine wetlands

在线全文

学校读者我要写书评

暂无评论

Science China Earth Sciences 2016年第3期59卷 464-476页

作者： ZHAO Shan ZHOU Nian Qing shen xin Ping Department of Hydraulic Engineering Tongji UniversityKey Laboratory of Yangtze River Water Environment Ministry of Education Shanghai 200092 China School of Civil and Environmental Engineering Georgia Institute of Technology Atlanta Georgia 30332 USA Dongting Lake Water Resources Administration Bureau of Hunan Province Changsha 410007 China

As an essential component of proteins and genetic material for all organisms, nitrogen(N) is one of the major limiting factors that control the dynamics, biodiversity and functioning of lacustrine wetlands, in which intensified N biogeochemical activities take place. Reactive N loaded into wetland ecosystems has been doubled due to various human activities, including industrial, agricultural activities and urbanization. The main driving mechanisms of N transport and transformation in lacustrine wetlands are categorized to pushing forces and pulling forces in this study. Geomorphology, wetland age, N concentrations, and temperature are the main pushing forces(passive forces); whereas water table variation, oxygen concentration, other elements availability, oxidation-reduction potential(Eh) and p H, and microorganisms are the predominant pulling forces(active forces). The direction and kinetic energy of reactions are determined by pulling forces and then are stimulated by pushing forces. These two types of forces are analyzed and discussed separately. Based on the analysis of driving mechanisms, possible solutions to wetland N pollutions are proposed at individual, regional and global scales, respectively. Additional research needs are addressed to obtain a thorough understanding of N transport and transformations in wetlands and to reduce detrimental impacts of excessive N on such fragile ecosystems.

关键词： Nitrogen Driving mechanisms Transport and transformation Excessive loading Lacustrine wetlands

A high-order splitting scheme for the advection diffusion equation

在线全文

学校读者我要写书评

暂无评论

Journal of Environmental Sciences 2001年第4期13卷 444-448页

作者： ZHENG Yong-hong shen Yong-ming QIU Da-hong Guangzhou Institute of Energy Conversion the Chinese Academy of SciencesGuangzhou 510070China State Key Laboratory of Coastal and Offshore Engineering Dalian University of TechnologyDalian 116023China

A high-order splitting scheme for the advection-diffusion equation of pollutants is proposed in this paper. The multidimensional advection-diffusion equation is splitted into several one-dimensional equations that are solved by the scheme. Only three spatial grid points are needed in each direction and the scheme has fourth-order spatial accuracy. Several typically pure advection and advection-diffusion problems are simulated. Numerical results show that the accuracy of the scheme is much higher than that of the classical schemes and the scheme can he efficiently solved with little programming effort.

关键词： pollutants advecton-diffusion equation high-order scheme numerical modelling

The magnetic properties and magnetocaloric effects in binary R-T(R = Pr,Gd,Tb,Dy,Ho,Er,Tm;T = Ga,Ni,Co,Cu)intermetallic compounds

在线全文

学校读者我要写书评

暂无评论

Chinese Physics B 2017年第2期26卷 1-41页

作者： xin-Qi Zheng Bao-Gen shen School of Materials Science and Engineering University of Science and Technology Beijing Beijing 100083 China Institute of Physics Chinese Academy of Sciences and University of Chinese Academy of Sciences Beijing 100190 China

In this paper, we review the magnetic properties and magnetocaloric effects（MCE） of binary R–T（R = Pr, Gd, Tb,Dy, Ho, Er, Tm; T = Ga, Ni, Co, Cu） intermetallic compounds（including RGa series, RNi series, R_（12）Co_7 series, R_3 Co series and RCu_2series）, which have been investigated in detail in the past several years. The R–T compounds are studied by means of magnetic measurements, heat capacity measurements, magnetoresistance measurements and neutron powder diffraction measurements. The R–T compounds show complex magnetic transitions and interesting magnetic *** types of magnetic transitions are investigated and confirmed in detail by multiple approaches. Especially, most of the R–T compounds undergo more than one magnetic transition, which has significant impact on the magnetocaloric effect of R–T compounds. The MCE of R–T compounds are calculated by different ways and the special shapes of MCE peaks for different compounds are investigated and discussed in detail. To improve the MCE performance of R–T compounds,atoms with large spin（S） and atoms with large total angular momentum（J） are introduced to substitute the related rare earth atoms. With the atom substitution, the maximum of magnetic entropy change（？SM）, refrigerant temperature width（Twidth）or refrigerant capacity（RC） is enlarged for some R–T compounds. In the low temperature range, binary R–T（R = Pr, Gd,Tb, Dy, Ho, Er, Tm; T = Ga, Ni, Co, Cu） intermetallic compounds（including RGa series, RNi series,R_（12）Co_7 series, R_3 Co series and RCu_2series） show excellent performance of MCE, indicating the potential application for gas liquefaction in the future.

关键词： magnetocaloric effect magnetic entropy change magnetic property neutron diffraction magnetic structure

Environment map building and localization for robot navigation based on image sequences

在线全文

学校读者我要写书评

暂无评论

Journal of Zhejiang University-Science A(Applied Physics & Engineering) 2008年第4期9卷 489-499页

作者： Ye-hu shen Ji-lin LIU xin DU Department of Information Science and Electronic Engineering Zhejiang University Hangzhou 310027 China

SLAM is one of the most important components in robot navigation. A SLAM algorithm based on image sequences captured by a single digital camera is proposed in this paper. By this algorithm, SIFT feature points are selected and matched between image pairs sequentially. After three images have been captured, the environment’s 3D map and the camera’s positions are initialized based on matched feature points and intrinsic parameters of the camera. A robust method is applied to estimate the position and orientation of the camera in the forthcoming images. Finally, a robust adaptive bundle adjustment algorithm is adopted to optimize the environment’s 3D map and the camera’s positions simultaneously. Results of quantitative and qualitative experiments show that our algorithm can reconstruct the environment and localize the camera accurately and efficiently.

关键词： Monocular vision Digital elevation map (DEM) SIFT Robust adaptive bundle adiustment SLAM algorithm

Multi-scale thermodynamic analysis method for 2D SiC/SiC composite turbine guide vanes

在线全文

学校读者我要写书评

暂无评论

Chinese Journal of Aeronautics 2018年第1期31卷 117-125页

作者： xin LIU Xiuli shen Longdong GONG Peng LI School of Energy and Power Engineering Beihang University Beijing 100083 China Navigation and Control Technology Research Institute of China Ordnance Industries Beijing 100089 China

Ceramic Matrix Composite （CMC） turbine guide vanes possess multi-scale stress and strain with inhomogeneity at the microscopic scale. Given that the macroscopic distribution cannot reflect the microscopic stress fluctuation, the macroscopic method fails to meet the requirements of stress and strain analysis of CMC turbine guide vanes. Furthermore, the complete thermodynamic properties of 2D woven SiC/SiC-CMC cannot be obtained through experimentation, Accordingly, a method to calculate the thermodynamic properties of CMC and analyze multi-scale stress and strain of the turbine guide vanes should be established. In this study, the multi-scale thermodynamic analysis is investigated. The thermodynamic properties of Chemical Vapor Infiltration （CVI） pro- cessed SiC/SiC-CMC are predicted by a Representative Volume Element （RVE） model with porosity, leading to the result that the relative error between the calculated in-plane tensile modulus and the experimental value is 4.2%. The macroscopic response of a guide vane under given conditions is predicted. The relative error between the predicted strain on the trailing edge and the experimental value is 9.7%. The calculation of the stress distribution of micro-scale RVE shows that the maximum value of microscopic stress, which is located in the interlayer matrix, is more than 1.5 times that of macroscopic stress in the same direction and the microscopic stress distribution of the interlayer matrix is related to the pore distribution of the composite.

关键词： Ceramic matrix composites Multi-scale Representative volume element Thermal-mechanical coupling Turbine guide vane