全连接回归神经网络的稳定性分析

Stability of Completely Connected Recursive Neural Networks

讨论全连接回归神经网络的稳定性 ,具体来说 ,就是分析网络输入和网络权值的扰动对网络输出的影响 ,得到了在上述扰动下 ,网络输出误差不累积的条件 ,即整个网络稳定的条件 .

This paper discusses the stability of completely connected recursive neural networks, that is, analyzes the affection of output of the neural networks under which the errors don't accumulate, brought by disturbances of inputs and weights. The corresponding conditions are obtained, which are also the conditions of structural stability of the neural networks.\;