Neural Network Based on Mean-Field Theory Approximation

Xi Guangcheng

您当前的位置：

首页 >

文章列表页 >

Neural Network Based on Mean-Field Theory Approximation

更新时间：2025-12-08

- Neural Network Based on Mean-Field Theory Approximation
- Acta Electronica Sinica Issue 8, (1995)
- 作者机构：
  
  中国科学院自动化研究所
- 作者简介：
- 基金信息：
- DOI：
  CLC： TP18
- Published：1995
- 稿件说明：
移动端阅览
[1]西广成.基于平均场理论逼近的神经网络[J].电子学报,1995(08):62-64.

Xi Guangcheng. Neural Network Based on Mean-Field Theory Approximation[J]. Acta Electronica Sinica, 1995, (8).
[1]西广成.基于平均场理论逼近的神经网络[J].电子学报,1995(08):62-64. DOI：

Xi Guangcheng. Neural Network Based on Mean-Field Theory Approximation[J]. Acta Electronica Sinica, 1995, (8). DOI：

摘要

本文试图从信息论观点导出平均场理论逼近的神经网络及其退火算法，证明ＡＭＦＴＡ（ＡｌｇｏｒｉｔｈｍｓｏｆＭｅａｎ－ｆｉｅｌｄｔｈｅｏｒｙａｐｐｒｏｘｉｍａｔｉｏｎ）和ＡＢＭ（ＡｌｇｏｒｉｔｈｍｏｆＢｏｌｔｚｍａｎｎｍａｃｈｉｎｅ）在一定条件下的等价性，既保留传统Ｂｏｌｔｚｍａｎｎ机算法的优点又提高了网络趋于稳态的收敛速度。

Abstract

This paper intends to derive the neural network based on mean-field theory approximation and its annealing algorithm from the information-theoretic view point

to prove equivalence between AMFTA and ABM under certain condition.both keeping advantage of ABM and speeding up rate of convergence for the neural network to approach to equilibrium state.

关键词

Keywords

references

Views

200

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

A Novel Edge Detection Algorithm

A Minimum Relative Entropy Approach to Parameter Estimation for Gaussian Random Fields

Related Author

Wang Tao

Xing Xiaoliang & Zhuang Xinhua

庄新华

Kuang Jinyu

Related Institution

Dept. of Computer Science & Engineering, Zhejiang University

浙江大学计算机科学与工程系杭州 310027

Dept. of Radio-Electronics, Beijing Normal University

⁰