摘要
多层感知器神经网络是典型的人工神经网络模型。本文在分析二值感知器神经网络实现二值数字逻辑运算后,将其推广到多值逻辑。根据感知器神经元的分类特点,采用三层前向稳健感知器神经网络实现了三值格代数系统中的基本运算,从而也实现了任意三值逻辑函数。
Multi layer perceptron(MLP) is a typical artificial neural network. In this paper, it is extended to multi-valued logic after analyzing implementation of binary logical operations based on MLP. Also, basic operations on ternary logic in post algebra are realized using three-layer feed-forward robust neural networks based on its classification property, accordingly arbitrary ternary logical functions can be realized.
出处
《电路与系统学报》
CSCD
2003年第6期91-94,共4页
Journal of Circuits and Systems
关键词
神经网络
感知器
多值逻辑
格代数系统
neural network
perceptron
multi-valued logic
post algebra
