Title :
Back-propagation as the solution of differential-algebraic equations for artificial neural network training
Author :
Sanchez-Gasca, J.J. ; Klapper, D.B. ; Yoshizawa, J.
Author_Institution :
GE Industrial & Power Systems, Power Syst. Eng. Dept., Schenectady, NY, USA
Abstract :
The backpropagation algorithm for neural network training is formulated as the solution of a set of sparse differential algebraic equations (DAE). These equations are then solved as a function of time. The solution of the differential equations is performed using an implicit integrator with adjustable time step. The topology of the Jacobian matrix associated with the DAE´s is illustrated. A training example is included
Keywords :
backpropagation; differential equations; matrix algebra; neural nets; AI; Jacobian matrix; algorithm; artificial neural network training; integrator; learning; sparse differential algebraic equations; topology; Artificial neural networks; Differential algebraic equations; Differential equations; Industrial power systems; Industrial training; Jacobian matrices; Minimization methods; Network topology; Neural networks; Nonlinear equations;
Conference_Titel :
Neural Networks to Power Systems, 1991., Proceedings of the First International Forum on Applications of
Conference_Location :
Seattle, WA
Print_ISBN :
0-7803-0065-3
DOI :
10.1109/ANN.1991.213470
