引用本文: | 罗莉,胡守仁.有限精度的 BP 神经网络高速算法[J].国防科技大学学报,1998,20(1):88-91.[点击复制] |
Luo Li,Hu Shouren.A High Speed BP Algorithm for the Feed-forward Multilayer Neural Network Using Limited Precision Calculation[J].Journal of National University of Defense Technology,1998,20(1):88-91[点击复制] |
|
|
|
本文已被:浏览 6097次 下载 5441次 |
有限精度的 BP 神经网络高速算法 |
罗莉, 胡守仁 |
(国防科技大学 计算机学院 湖南 长沙 410073)
|
摘要: |
精度要求是神经网络硬件实现的一个关键问题。本文提出了一种有限精度的 BP 高速算法,通过实例验证,该算法适于低精度,且收敛速度快。 |
关键词: BP 学习算法,代价函数,活跃函数,有限精度计算,神经网络硬件 |
DOI: |
投稿日期:1997-03-28 |
基金项目: |
|
A High Speed BP Algorithm for the Feed-forward Multilayer Neural Network Using Limited Precision Calculation |
Luo Li, Hu Shouren |
(Department of Computer Science,NUDT,Changsha,410073)
|
Abstract: |
The precision required for neural network algorithm is an important question for hard-ware architects. This paper gives an excellent high speed BP algorithm for feed-forward muitilayer neural network using limited precision calculation. The experimental results show that the algorithm improves convergence speed with reduced precision bits. |
Keywords: back-propagation,feed-forward multilayer neural network,cost function,activation function,limited precision calculation,neural network hardware |
|
|