Title :
Model trust region technique in parallel Newton´s method for training feedforward neural networks
Author :
Zhao, M.D. ; Wang, X.
Author_Institution :
Harbin Inst. of Technol., China
Abstract :
The double dogleg trust region approach of unconstrained minimization is introduced in the parallel Newton´s (PN) algorithm proposed by M. D. Zhao (1993). The PN algorithm uses a recursive procedure for computing both the Hessian matrix and the Newton direction. The input weights of each neuron in the network are updated after each presentation of the training data with a global strategy. Experimental results indicate that the double dogleg trust region approach is superior to the line search technique in the PN algorithm, and that the PN algorithm with both global strategies exhibits better convergence performance than the well-known backpropagation algorithm
Keywords :
Hessian matrices; Newton method; convergence of numerical methods; feedforward neural nets; learning (artificial intelligence); Hessian matrix; Newton direction; convergence performance; double dogleg trust region approach; feedforward neural networks; global strategy; input weights; parallel Newton´s method; recursive procedure; training data; Backpropagation algorithms; Convergence; Feedforward neural networks; Intelligent networks; Neural networks; Neurons; Newton method; Signal processing algorithms; Supervised learning; Training data;
Conference_Titel :
Circuits and Systems, 1993., ISCAS '93, 1993 IEEE International Symposium on
Conference_Location :
Chicago, IL
Print_ISBN :
0-7803-1281-3
DOI :
10.1109/ISCAS.1993.394247