摘要
基于LMS的标准BP算法收敛速度极慢,而共轭梯度法要求精确的线性搜索,这在神经网络的高维权空间中是难以实现的。本文提出了一种新的BP学习算法,它采用一种对线性搜索要求不高的改进的共轭梯度法与一种简单的不精确线性搜索相结合,极大地提高了BP学习速度。经多次测试表明,与标准BP算法相比,该算法的效率提高了二个数量极。
The LMS- based conventional backpropagation(BP)algoritim is very slow inconvergence speed in the high-dimensional weight space of neural network, the exact linesearches required by conjugate gradient algorithm is very difficult in implementation. This paperpresents a novel BP algorithm, which greatly improves leaming speed by using a modifiedconjugate gradient algorithm incorporating a simple inexact line sedch algorithm. Given manytesting results show that the efficiency of the algorithm relative to conventional BP is of morethan two order magnitude.
出处
《电子科技大学学报》
EI
CAS
CSCD
北大核心
1994年第5期462-467,共6页
Journal of University of Electronic Science and Technology of China
基金
国家"八五"军事电子预研基金
关键词
神经网络
BP学习
共轭梯度法
无约束最优化
neural network
BP learning
conjugate gradient algorithm
unconstrained optimization