Please use this identifier to cite or link to this item:
Title: Multi-phase fast learning algorithms for solving the local minimum problem in feed-forward neural networks
Authors: Cheung, CC 
Ng, SC
Lui, AKF
Keywords: Backpropagation
Local minimum
Multi-phase learning algorithms
Issue Date: 2012
Publisher: Springer
Source: Lecture notes in computer science (including subseries Lecture notes in artificial intelligence and lecture notes in bioinformatics), 2012, v. 7367 LNCS, no. PART 1, p. 580-589 How to cite?
Journal: Lecture notes in computer science (including subseries Lecture notes in artificial intelligence and lecture notes in bioinformatics) 
Abstract: Backpropagation (BP) learning algorithm is the most widely supervised learning technique which is extensively applied in the training of multi-layer feed-forward neural networks. Many modifications of BP have been proposed to speed up the learning of the original BP. However, they all have different drawbacks and they cannot perform very well in all kinds of applications. This paper proposes a new algorithm, which provides a systematic approach to make use of the characteristics of different fast learning algorithms so that the learning process can converge to the global minimum. During the training, different fast learning algorithms will be used in different phases to improve the global convergence capability. Our performance investigation shows that the proposed algorithm always converges in different benchmarking problems (applications) whereas other popular fast learning algorithms sometimes give very poor global convergence capabilities.
Description: 9th International Symposium on Neural Networks, ISNN 2012, Shenyang, 11-14 July 2012
ISBN: 9783642313455
ISSN: 0302-9743
EISSN: 1611-3349
DOI: 10.1007/978-3-642-31346-2_65
Appears in Collections:Conference Paper

View full-text via PolyU eLinks SFX Query
Show full item record


Last Week
Last month
Citations as of Nov 3, 2018

Page view(s)

Last Week
Last month
Citations as of Nov 11, 2018

Google ScholarTM



Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.