A new training algorithm is presented as a fast alternative to the backpropagation method. The new approach is based on the solution of a linear system at each step of the learning phase. The squared error at the output of each layer before the nonlinearity is minimised on the entire set of the learning patterns by a block LS algorithm. The optimal weights for each layer are then computed by using the SVD technique. The simulation results have shown considerable improvements from the point of view of both the accuracy and the speed of convergence.
LS-backpropagation algorithm for training multilayer perceptrons / DI CLAUDIO, Elio; Parisi, Raffaele; Orlandi, Gianni. - STAMPA. - (1993), pp. 768-771. (Intervento presentato al convegno 1993 International Conference on Artificial Neural Networks (ICANN’93) tenutosi a Amsterdam nel Sept. 13-16, 1993).
LS-backpropagation algorithm for training multilayer perceptrons
DI CLAUDIO, Elio;PARISI, Raffaele;ORLANDI, Gianni
1993
Abstract
A new training algorithm is presented as a fast alternative to the backpropagation method. The new approach is based on the solution of a linear system at each step of the learning phase. The squared error at the output of each layer before the nonlinearity is minimised on the entire set of the learning patterns by a block LS algorithm. The optimal weights for each layer are then computed by using the SVD technique. The simulation results have shown considerable improvements from the point of view of both the accuracy and the speed of convergence.I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.