Title :
A stochastic learning algorithm for generalization problems
Author :
Ramamoorthy, C.V. ; Shekhar, Shashi
Author_Institution :
Div. of Comput. Sci., California Univ., Berkeley, CA, USA
Abstract :
A discussion is presented of the requirements of learning for generalization, which is NP-complete and cannot be addressed by traditional methods based on gradient descent. The authors present a stochastic learning algorithm based on simulated annealing in weight space and discuss stopping criteria for the algorithm, to avoid overfitting of learning examples
Keywords :
learning systems; neural nets; simulated annealing; stochastic processes; NP-complete; generalization; generalization problems; simulated annealing; stochastic back propagation; stochastic backpropagation; stochastic learning algorithm; stopping criteria; weight space; Backpropagation algorithms; Computer science; Neural networks; Noise shaping; Predictive models; Shape; Simulated annealing; Speech recognition; Stochastic processes; Testing;
Conference_Titel :
TENCON '89. Fourth IEEE Region 10 International Conference
Conference_Location :
Bombay
DOI :
10.1109/TENCON.1989.176913
