时变容错域的感知联想记忆模型及其实现算法  

A Time-Varying Fault-Tolerant Field Perceptive Associative Memory Model and Its Realization Algorithms

作　　者：杨国为[1] 涂序彦[2] 王守觉[3] 

机构地区：[1]青岛大学模式识别与智能系统研究所,青岛266071 [2]北京科技大学计算机与系统科学研究所,北京100083 [3]中国科学院半导体研究所神经网络实验室,北京100083

出　　处：《计算机学报》2006年第3期431-440,共10页Chinese Journal of Computers

基　　金：国家自然科学基金(60375014);山东省自然科学基金(Y2002G16);中国博士后科学基金资助.

摘　　要：提出并用软件实现了一种时变容错域的感知联想记忆模型.该模型有以下特点:(1)模拟了大脑联想记忆的容错域随时间变化的特点.模型可以根据记忆样本的重要性,为记忆样本在不同时间设计适当的容错域;(2)实现了n维到m维空间的无穷值模式非线性联想,而且该模型的样本容错域充满整个实空间Rn,模型没有伪吸引子;(3)联想记忆速度快.A time varying fault-tolerant field perceptive associative memory model is proposed and realized by software. One of the characters of this model is that it simulates the time-dependent changeability character of the fault-tolerant field of human brain＇s associative memory. This model can design proper fault-tolerant field for memory sample at different time according to the essentiality of memory sample. The second character is that it realizes the nonlinear association of infinite value pattern from n dimension space to m dimension space. And the fault-tolerant field of the memory samples is full of the whole real space R^n and the associative memory model has no spurious attractors. The third character is that the speed of associative memory about the model is faster.

关 键 词：联想记忆 神经网络 时变容错域 人工生命 人工脑 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]