Title :
The quadratic property of the L-MBFGS methods for training neural networks
Author :
Lin Zhao ; Dali Wang ; Yueting Yang
Author_Institution :
Normal Sch., Beihua Univ., Jilin, China
Abstract :
In this paper, we introduce the use of limited memory modified BFGS method (L-MBFGS) to improve the efficiency of training algorithms for feedforward neural networks. The quadratic termination property of L-MBFGS algorithm is given which is an important quasi-Newton property.
Keywords :
feedforward neural nets; learning (artificial intelligence); L-MBFGS method; feedforward neural networks; limited memory modified BFGS method; quadratic termination property; quasi-Newton property; training algorithms; Algorithm design and analysis; Biological neural networks; Feedforward neural networks; Mathematical model; Optimization; Training; backpropagation (BP); limited memory technique; neural networks; quasi-Newton methods; training algorithm;
Conference_Titel :
Mechatronic Science, Electric Engineering and Computer (MEC), 2011 International Conference on
Conference_Location :
Jilin
Print_ISBN :
978-1-61284-719-1
DOI :
10.1109/MEC.2011.6025596
