摘要
稳定点是决定人工神经网络联想记忆能力的最重要因素。本文在二进联想Hopfield网络的稳定点统计分布、计数和快速判别等方面给出了若干新结果。
The number of stable points is a key measurement of the capacity of neural network's associative memory. New results about the distribution of the stable points of binary associativ Hopfield neural networks are presented in this paper.
出处
《电子学报》
EI
CAS
CSCD
北大核心
1992年第1期1-8,共8页
Acta Electronica Sinica
关键词
稳定点
联想矩阵
纠错
神经网络
Hopfield neural network, Stable point, Associative matrix, Error-correcting encoder
