TY - GEN

T1 - A hierarchical method for training embedded sigmoidal neural networks

AU - Hu, Jinglu

AU - Hirasawa, Kotaro

PY - 2001

Y1 - 2001

N2 - This paper discusses the problem of applying sigmoidal neural networks to identification of nonlinear dynamical systems. When using sigmoidal neural networks directly as nonlinear models, one often meets problems such as model parameters lack of physical meaning, sensitivity to noise in model training. In this paper, we introduce an embedded sigmoidal neural network model, in which the neural network is not used directly as a model, but is embedded in a shield such that part of the model parameters become meaningful. Corresponding to the meaningful part and the meaningless part of model parameters, a hierarchical training algorithm consisting of two learning loops is then introduced to train the model. Simulation results show that such a dual loop learning algorithm can solve the noise sensitivity and local minimum problems to some extent.

AB - This paper discusses the problem of applying sigmoidal neural networks to identification of nonlinear dynamical systems. When using sigmoidal neural networks directly as nonlinear models, one often meets problems such as model parameters lack of physical meaning, sensitivity to noise in model training. In this paper, we introduce an embedded sigmoidal neural network model, in which the neural network is not used directly as a model, but is embedded in a shield such that part of the model parameters become meaningful. Corresponding to the meaningful part and the meaningless part of model parameters, a hierarchical training algorithm consisting of two learning loops is then introduced to train the model. Simulation results show that such a dual loop learning algorithm can solve the noise sensitivity and local minimum problems to some extent.

UR - http://www.scopus.com/inward/record.url?scp=33645276097&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=33645276097&partnerID=8YFLogxK

U2 - 10.1007/3-540-44668-0_129

DO - 10.1007/3-540-44668-0_129

M3 - Conference contribution

AN - SCOPUS:33645276097

SN - 3540424865

SN - 9783540446682

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 937

EP - 942

BT - Artificial Neural Networks - ICANN 2001 - International Conference, Proceedings

A2 - Hornik, Kurt

A2 - Dorffner, Georg

A2 - Bischof, Horst

PB - Springer Verlag

T2 - International Conference on Artificial Neural Networks, ICANN 2001

Y2 - 21 August 2001 through 25 August 2001

ER -