Convergence properties of backpropagation for neural nets via theory of stochastic gradient methods. Part 1
作者:
Alexei A. Gaivoronski,
期刊:
Optimization Methods and Software
(Taylor Available online 1994)
卷期:
Volume 4,
issue 2
页码: 117-134
ISSN:1055-6788
年代: 1994
DOI:10.1080/10556789408805582
出版商: Gordon and Breach Science Publishers
关键词: Neural network,Backpropagation, Stochastic gradient
数据来源: Taylor
摘要:
We study here convergence properties of serial and parallel backpropagation algorithm for training of neural nets, as well as its modification with momentum term. It is shown that these algorithms can be put into the general framework of the stochastic gradient methods. This permits to consider from the same positions both stochastic and deterministic rules for the selection of components (training examples) of the error function to minimize at each iteration. We obtained weaker conditions on the stepsize for deterministic case and provide quite general synchronization rule for parallel version.
点击下载:
PDF (556KB)
返 回