GAF:一个自适应通用神经网络框架

GAF: A GENERAL ADAPTIVE NEURAL NETWORK FRAMEWORK

本文首先讨论了通用PDP框架的局限性,然后介绍了我们提出的一个自适应通用神经网络框架.它使得PDP框架成为它的一个特例.由于在提出本框架的过程中充分考虑了人脑的退化、自适应等机制,从而使得利用本框架来动态模拟神经系统的退化及自适应等功能更为简便.而且,本框架对更合理地解释人脑的工作机制及思维模拟有许多优越性.

The PDP Framework is evaluted first. Then, a general adaptive neural network framework, GAF, is presented. The PDP Framework is a special case of GAF. Since the mechanisms of the degradation, adaptation of human brain have been taken into consideration during the studying of GAF, it is much easier, using GAF, to do dynamic simulation of the degradation, generation and disappearance of connections.