摘要
生物神经系统是一种脉冲发放神经网络系统。当前流行的人工神经网络模型与生物神经系统相比仍存在着较大差异。本文根据电生理学结果建立了适合于工程应用的脉冲发放神经网络模型。给出了输入－输出关系、突触联接关系、动态方程和学习功力学以及各种统计量的均值和方差。
The biological nerve system is an impulse-firing neural network. However, there is a great difference between the available neural networks and the biological nerve system. According to the results obtained in the nerve physiology, the paper models the'impulsefiring neural networks that can be extensively applied in engineering. The relation between input and output, the synaptic connection relation, and their dynamic equations and learning rules, as well as the mean value and mean-square error of some stochastic variables are given in the paper.
出处
《系统工程与电子技术》
CSCD
1996年第5期23-30,共8页
Systems Engineering and Electronics
基金
国家自然科学基金
国防预研基金
关键词
神经元
随机过程
神经网络
建模
Neural network, Neuron, Shunting model, Learning and memory, Noise,Stochastic process.
