论文部分内容阅读
一、概述自从Bumelhart提出了训练多层感知器网络的反向传播学习算法以来，人们改变了对多层感知器网络难以训练学习的悲观认识，从而使多层感知器网络的处境大为改观。近年来，人们已将BP算法成功地应用于语音识别、图象处理等领域。然而，在实际应用中，BP算法训练多层网络存
I. Overview Since Bumelhart proposed a back propagation learning algorithm to train multi-layer perceptron networks, people have changed their pessimistic understanding about the hard-to-train learning of multi-layer perceptron networks and have greatly changed the situation of multi-layer perceptron networks. In recent years, BP algorithm has been successfully used in speech recognition, image processing and other fields. However, in practical applications, BP algorithm training multi-layer network storage