Title :
A pruning technique maximizing generalization
Author :
Burrascano, Pietro
Author_Institution :
INFOCOM Dept., Rome Univ., Italy
Abstract :
The paper proposes an estimator of the relevance of each weight connection on generalization which relies on a probabilistic description of the training phase: based on the above estimator, a pruning technique is derived.
Keywords :
generalisation (artificial intelligence); neural nets; probability; generalization maximization; neural net; probabilistic description; pruning technique; Constraint optimization; Cost function; Current measurement; Equations; Network topology; Neural networks; Phase estimation; Phase measurement; Probability; Stochastic processes;
Conference_Titel :
Neural Networks, 1993. IJCNN '93-Nagoya. Proceedings of 1993 International Joint Conference on
Print_ISBN :
0-7803-1421-2
DOI :
10.1109/IJCNN.1993.713928
