条件联想神经网络研究

ON THE CONDITIONAL ASSOCIATIVE NEURAL NETWORK

在人脑的某些功能和神经系统中的突前抑制机制启发下,本文提出一个新型的神经网络模型——条件联想神经网络.模型是一个有突触前抑制的联想记忆神经网络.通过初步分析和计算机模拟,证明本模型具有一般联想记忆模型所未有的一些新的特性,如可以在不同条件下,对同一输入有不同的反应.对同一输入,在不同的条件下,又可以有相同的反应.这些特点将有助于人们对神经系统中信息处理过程的了解.此外,文中也指出可能实现本模型的神经结构.

A new neural network model, the conditional associative neural network, was proposed in this paper.The model consists of a conventional associative memory network and a presynapstic inhibitation network. The preliminary analysis and computer simulation show that the model has some capabilities that conventional associative memory model do not have, such as it can response to same input with different outputs under different conditions, and response to different inputs with same output under different conditions. The model will help us to understand some information processing in the brain. Furthermore, a possible neural realization of the model was Proposed.