摘要
通常的联想记忆模型的联想性能由于受到输入模式间交叉相关项的影响而有所下降,并且在输入与输出之间缺乏非线性映射能力。本文介绍一种高性能联想记忆模型,它将低维输入向量映射到一个高维的中间向量,从而提高了系统的联想能力,又使系统具有非线性映射能力,最后给出了几种推广。
The associative performance of ordinary associative memories depend on cross correlation term between input patterns. Hence their performance is degraded. In addition, the mapping between inputs and outputs is also only linear. The author propose an associative memory model with high performance, its associative performance is improved via mapping input vectors with low dimension to those with high dimension, it has also the nonlinear mapping ability. Finally several extensions are given.
出处
《数据采集与处理》
CSCD
1996年第4期257-259,共3页
Journal of Data Acquisition and Processing
基金
江苏省自然科学基金
关键词
神经网络
非线性映射
联想记忆模型
neural networks
associative memories
nonlinear mapping