检索结果-南通市图书馆

Chaotic time series prediction using mean-field theory for support vector machine

在线全文

学校读者我要写书评

暂无评论

Chinese Physics B 2005年第5期14卷 922-929页

作者：崔万照朱长纯保文星刘君华 SchoolofElectronicsandInformationEngineering Xi'anJiaotongUniversityXi'an710049China SchoolofElectronicsandInformationEngineering Xi'anJiaotongUniversityXi'an710049China

This paper presents a novel method for predicting chaotic time series which is based on the support vector machines approach, and it uses the mean-field theory for developing an easy and efficient learning procedure for the support vector machine. The proposed method approximates the distribution of the support vector machine parameters to a Gaussian process and uses the mean-field theory to estimate these parameters easily, and select the weights of the mixture of kernels used in the support vector machine estimation more accurately and faster than traditional quadratic programming-based algorithms. Finally, relationships between the embedding dimension and the predicting performance of this method are discussed, and the Mackey-Glass equation is applied to test this method. The stimulations show that the mean-field theory for support vector machine can predict chaotic time series accurately, and even if the embedding dimension is unknown, the predicted results are still satisfactory. This result implies that the mean-field theory for support vector machine is a good tool for studying chaotic time series.

关键词： chaotic time series support vector machine mean-field theory

Large-scale self-normalizing neural networks

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Journal of Automation and Intelligence 2024年第2期3卷 101-110页

作者： Zhaodong Chen Weiqin Zhao Lei Deng Yufei Ding Qinghao Wen Guoqi Li Yuan Xie University of California Santa BarbaraSanta Barbara93106CAUnited States University of Hong Kong Hong Kong999077China Tsinghua University Beijing100084China University of California San DiegoSan Diego92093CAUnited States University of Sydney Sydney2006Australia Institute of Automation Chinese Academy of ScienceBeijing100190China University of Chinese Academy of Science Beijing100190China The Hong Kong University of Science and Technology Hong Kong999077China

Self-normalizing neural networks(SNN)regulate the activation and gradient flows through activation functions with the self-normalization *** SNNs do not rely on norms computed from minibatches,they are more friendly to data parallelism,kernel fusion,and emerging architectures such as ReRAM-based ***,existing SNNs have mainly demonstrated their effectiveness on toy datasets and fall short in accuracy when dealing with large-scale tasks like *** lack the strong normalization,regularization,and expression power required for wider,deeper models and larger-scale *** enhance the normalization strength,this paper introduces a comprehensive and practical definition of the self-normalization property in terms of the stability and attractiveness of the statistical fixed *** is comprehensive as it jointly considers all the fixed points used by existing studies:the first and second moment of forward activation and the expected Frobenius norm of backward *** practicality comes from the analytical equations provided by our paper to assess the stability and attractiveness of each fixed point,which are derived from theoretical analysis of the forward and backward *** proposed definition is applied to a meta activation function inspired by prior research,leading to a stronger self-normalizing activation function named‘‘bi-scaled exponential linear unit with backward standardized’’(bSELU-BSTD).We provide both theoretical and empirical evidence to show that it is superior to existing *** enhance the regularization and expression power,we further propose scaled-Mixup and channel-wise scale&*** these three techniques,our approach achieves 75.23%top-1 accuracy on the ImageNet with Conv MobileNet V1,surpassing the performance of existing self-normalizing activation *** the best of our knowledge,this is the first SNN that achieves comparable accuracy to batch normalization on ImageNet.

关键词： Self-normalizing neural network mean-field theory Block dynamical isometry Activation function

Effective diffusion of a tracer in active bath: A path-integral approach

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

National Science Open 2024年第4期3卷 29-40页

作者： Mengkai Feng Zhonghuai Hou Hefei National Research Center for Physical Sciences at the Microscale & Key Laboratory of Precision and Intelligent Chemistry University of Science and Technology of China

We investigate the effective diffusion of a tracer immersed in an active particle bath consisting of self-propelled particles. Utilising the Dean's method developed for the equilibrium bath and extending it to the nonequilibrium situation, we derive a generalized Langevin equation(GLE) for the tracer particle. The complex interactions between the tracer and bath particles are shown as a memory kernel term and two colored noise terms. To obtain the effective diffusivity of the tracer, we use path integral technique to calculate all necessary correlation functions. Calculations show the effective diffusion decreases with the persistent time of active force, and has rich behavior with the number density of bath particles, depending on different activities. All theoretical results regarding the dependence of such diffusivity on bath parameters have been confirmed by direct computer simulation.

关键词： active matter nonequilibrium statistical mechanics active bath mean-field theory path integral tracer diffusivity

Successive defects asymmetric simple exclusion processes with particles of arbitrary size

在线全文

同方期刊数据库

学校读者我要写书评

暂无评论

Chinese Physics B 2009年第12期18卷 5097-5102页

作者：蔡九菊肖松王若翚刘飞 SEPA Key Laboratory on Eco-Industry Northeastern University College of Science Institute of Information Science and TechnologyMassey UniversityPalmerston NorthNew Zealand

This paper uses various mean-field approaches and the Monte Carlo simulation to calculate asymmetric simple exclusion processes with particles of arbitrary size in the successive defects system. In this system, the hopping probability p （p 〈 1） and the size d of particles are not constant, Through theoretical calculation and computer simulation, it obtains the exact theoretical results and finds that the theoretical results are in agreement with the computer simulation. These results are helpful in analysing the effect of traffic with different hopping probabilities p and sizes d of particle.

关键词： successive defects asymmetric simple exclusion processes arbitrary size Monte Carlo simulation mean-field theory

Comparison of Properties of the Simplest Neutron Stars in Three RMF Models

在线全文

学校读者我要写书评

暂无评论

Chinese Physics Letters 2008年第8期25卷 2837-2840页

作者：王国华付伟杰刘玉鑫 Department of Physics and the State Key Laboratory of Nuclear Physics and Technology Peking University Belling 100871 Genter of Theoretical Nuclear Physics National Laboratory of Heavy Ion Accelerator Lanzhou 730000

We study some properties of the simplest neutron stars （NSs） in the Glendenning Moszkowski （GM） model, the hybrid derivative coupling （HD） model and the Zimanyi Moszkowski （ZM） model in the framework of relativistic mean field （RMF） theory with and without the interaction by exchanging the ^-meson. We show that the maximal mass of the NSs becomes smaller, but the redshift becomes larger from the GM model to the HD model, then to the ZM model. The interaction with the 6-meson exchange enlarges the maximal mass of neutron stars, increases the relative population of charged particles （proton, electron and muon） and descends the relative population of neutron.

关键词： ISOVECTOR-SCALAR MESON mean-field theory NUCLEAR-MATTER SYMMETRY ENERGY HYPERNUCLEI STATE

Ground-State Properties of Ca Isotopes and the Density Dependence of the Symmetry Energy

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Chinese Physics Letters 2006年第7期23卷 1719-1722页

作者：梁钧马中玉陈宝秋 China Institute of Atomic Energy Beijing 102413 Department of Physics University of Technology of Taiyuan Taiyuan 030024 Center of Theoretical Nuclear Physics National Laboratory of Heavy Ion Accelerator of Lanzhou Lanzhou 730000

A relativistic mean field model is used to study the ground-state properties of neutron-rich nuclei in Ca isotopes. An additional isoscalar and isovector nonlinear coupling has been introduced in the relativistic mean field model, which could soften the symmetry energy, while keep the agreement with the experimental data. The sensitivity of proton and neutron density distributions and single particle states in Ca isotopes to the additional isoscalarisovector nonlinear coupling term is investigated. We found that the binding energies, the density distributions of single particle levels are strongly correlated with the density dependence of the symmetric energy in nuclear matter.

关键词： mean-field theory NEUTRON RADII NUCLEI SCATTERING EQUATION PB-208 MODELS

Kinetic Ising model in a time-dependent oscillating external magnetic field:effective-field theory

在线全文

维普期刊数据库

学校读者我要写书评

暂无评论

Chinese Physics B 2010年第5期19卷 187-194页

作者： Bayram Deviren Osman Canko Mustafa Keskin Institute of Science Erciyes University38039 KayseriTurkey Department of Physics Nevsehir University50300 NevsehirTurkey Department of Physics Erciyes University38039 KayseriTurkey

Recently, Shiet al. [2008 Phys. Left. A 372 5922] have studied the dynamical response of the kinetic Ising model in the presence of a sinusoidal oscillating field and presented the dynamic phase diagrams by using an effective-field theory （EFT） and a mean-field theory （MFT）. The MFT results are in conflict with those of the earlier work of Tome and de Oliveira, [1990 Phys. Rev. A 41 4251]. We calculate the dynamic phase diagrams and find that our results are similar to those of the earlier work of Tome and de Oliveira; hence the dynamic phase diagrams calculated by Shiet al. are incomplete within both theories, except the low values of frequencies for the MFT calculation. We also investigate the influence of external field frequency （w） and static external field amplitude （h0） for both MFT and EFT calculations. We find that the behaviour of the system strongly depends on the values of w and h0.

关键词： kinetic Ising model effective-field theory mean-field theory

Effects of electron correlation on superconductivity in the Hatsugai-Kohmoto model

在线全文

学校读者我要写书评

暂无评论

Chinese Physics B 2021年第10期30卷 541-547页

作者： Huai-Shuang Zhu Qiang Han Department of Physics Renmin University of ChinaBeijing 100872China

Understanding how electrons form pairs in the presence of strong electron correlations demands going beyond the BCS *** study a correlated superconducting model where the correlation effects are accounted for by a U term local in momentum *** electron correlation is treated exactly while the electron pairing is treated approximately using the mean-field *** self-consistent equation for the pair potential is derived and *** contrary to expectation,a weak attractive U comparable to the pair potential can destroy the superconductivity,whereas for weak to intermediate repulsive U,the pair potential can be *** fidelity of the mean-field ground state is calculated to describe the strength of the elelectron *** show that the pair potential is not equal to the single-electron superconducting gap for the strongly correlated superconductors,in contrast to the uncorrelated BCS limit.

关键词： strong correlated superconductivity exactly solvable model mean-field theory

Less is more: wiring-economical modular networks support self-sustained firing-economical neural avalanches for efficient processing

在线全文

学校读者我要写书评

暂无评论

National Science Review 2022年第3期9卷 19-31页

作者： Junhao Liang Sheng-Jun Wang Changsong Zhou Department of Physics Centre for Nonlinear Studies and Beijing-Hong Kong-Singapore Joint Centre for Nonlinear and Complex Systems (Hong Kong) Institute of Computational and Theoretical Studies Hong Kong Baptist University Department of Physics Shaanxi Normal University Department of Physics Zhejiang University Beijing Computational Science Research Center

The brain network is notably cost-efficient, while the fundamental physical and dynamic mechanisms underlying its economical optimization in network structure and activity have not been determined. In this study, we investigate the intricate cost-efficient interplay between structure and dynamics in biologically plausible spatial modular neuronal network models. We observe that critical avalanche states from excitation-inhibition balance under modular network topology with less wiring cost can also achieve lower costs in firing but with strongly enhanced response sensitivity to stimuli. We derive mean-field equations that govern the macroscopic network dynamics through a novel approximate theory. The mechanism of low firing cost and stronger response in the form of critical avalanches is explained as a proximity to a Hopf bifurcation of the modules when increasing their connection density. Our work reveals the generic mechanism underlying the cost-efficient modular organization and critical dynamics widely observed in neural systems, providing insights into brain-inspired efficient computational designs.

关键词： neural network cost efficiency critical avalanche modular network mean-field theory