A Learning Algorithm of the Neural Network Based on Kalman Filtering
Summary :
A novel algorithm based on Kalman filtering is developed for the learning of a layered neural network. The problem of adjusting the weight can be regarded as that of estimating a signal state vector of a linear process. The proposed algorithm, though computationally complex, has an adaptively varying learning rate, while the back-propagation algorithm has constant learning rate. Some experiments conducted for XOR and auto-associative image compression problems have shown that the proposed learning algorithm usually converges in a few iterations and the error is comparable to that of the well-known back-propagation algorithm.
- Publication
- IEICE TRANSACTIONS on Fundamentals Vol.E74-A No.5 pp.1059-1065
- Publication Date
- 1991/05/25
- Publicized
- Online ISSN
- DOI
- Type of Manuscript
- PAPER
- Category
- Nonlinear Problems
Authors
Keyword
Latest Issue
Copyrights notice of machine-translated contents
The copyright of the original papers published on this site belongs to IEICE. Unauthorized use of the original or translated papers is prohibited. See IEICE Provisions on Copyright for details.
Cite this
Copy
Tong HUANG, Masaharu TSUYUKI, Makoto YASUHARA, "A Learning Algorithm of the Neural Network Based on Kalman Filtering" in IEICE TRANSACTIONS on Fundamentals,
vol. E74-A, no. 5, pp. 1059-1065, May 1991, doi: .
Abstract: A novel algorithm based on Kalman filtering is developed for the learning of a layered neural network. The problem of adjusting the weight can be regarded as that of estimating a signal state vector of a linear process. The proposed algorithm, though computationally complex, has an adaptively varying learning rate, while the back-propagation algorithm has constant learning rate. Some experiments conducted for XOR and auto-associative image compression problems have shown that the proposed learning algorithm usually converges in a few iterations and the error is comparable to that of the well-known back-propagation algorithm.
URL: https://global.ieice.org/en_transactions/fundamentals/10.1587/e74-a_5_1059/_p
Copy
@ARTICLE{e74-a_5_1059,
author={Tong HUANG, Masaharu TSUYUKI, Makoto YASUHARA, },
journal={IEICE TRANSACTIONS on Fundamentals},
title={A Learning Algorithm of the Neural Network Based on Kalman Filtering},
year={1991},
volume={E74-A},
number={5},
pages={1059-1065},
abstract={A novel algorithm based on Kalman filtering is developed for the learning of a layered neural network. The problem of adjusting the weight can be regarded as that of estimating a signal state vector of a linear process. The proposed algorithm, though computationally complex, has an adaptively varying learning rate, while the back-propagation algorithm has constant learning rate. Some experiments conducted for XOR and auto-associative image compression problems have shown that the proposed learning algorithm usually converges in a few iterations and the error is comparable to that of the well-known back-propagation algorithm.},
keywords={},
doi={},
ISSN={},
month={May},}
Copy
TY - JOUR
TI - A Learning Algorithm of the Neural Network Based on Kalman Filtering
T2 - IEICE TRANSACTIONS on Fundamentals
SP - 1059
EP - 1065
AU - Tong HUANG
AU - Masaharu TSUYUKI
AU - Makoto YASUHARA
PY - 1991
DO -
JO - IEICE TRANSACTIONS on Fundamentals
SN -
VL - E74-A
IS - 5
JA - IEICE TRANSACTIONS on Fundamentals
Y1 - May 1991
AB - A novel algorithm based on Kalman filtering is developed for the learning of a layered neural network. The problem of adjusting the weight can be regarded as that of estimating a signal state vector of a linear process. The proposed algorithm, though computationally complex, has an adaptively varying learning rate, while the back-propagation algorithm has constant learning rate. Some experiments conducted for XOR and auto-associative image compression problems have shown that the proposed learning algorithm usually converges in a few iterations and the error is comparable to that of the well-known back-propagation algorithm.
ER -
English
