Title :
A new algorithm for training multilayer perceptrons
Author :
Palmieri, Francesco ; Shah, Samir A.
Author_Institution :
Dept. of Electr. & Syst. Eng., Connecticut Univ., Storrs, CT, USA
Abstract :
A fast version of the packpropagation algorithm based on the recursive least squares (RLS) technique is introduced. The added storage requirement and computation associated with RLS can be easily incorporated into the network architecture, and consequently the algorithm is still local. The enhanced algorithm performs consistently better than the backpropagation algorithm in a set of simulations involving two benchmark problems
Keywords :
artificial intelligence; neural nets; multilayer perceptrons; network architecture; neural nets; packpropagation algorithm; recursive least squares; storage requirement; Computer architecture; Computer networks; Convergence; Couplings; Iterative algorithms; Least squares methods; Multilayer perceptrons; Nonlinear filters; Resonance light scattering; Systems engineering and theory;
Conference_Titel :
Systems, Man and Cybernetics, 1989. Conference Proceedings., IEEE International Conference on
Conference_Location :
Cambridge, MA
DOI :
10.1109/ICSMC.1989.71330
