Analogue imprecision in mlp training, progress in neural processing, vol 4
Auteur :
Edwards, Peter J.
Éditeur :
Edwards, Peter J.Murray, A.F.,
ISBN :
9789810227395
Date de publication :
8 janv. 1996
Langue :
Anglais
Pays d'origine :
Singapour
Hardware inaccuracy and imprecision are important factors when implementing neural algorithms. This text presents a study of synaptic weight noise as a typical fault model for analogue VLSI realizations of MLP neural networks, and examines the implications for learning and network performance.