Title :
Correct implementation of digital neural networks
Author :
Basaglia, A. ; Fornaciari, W. ; Salice, F.
Author_Institution :
ITALTEL-SIT, Castelletto di Settimo, Italy
Abstract :
This paper presents a theoretical analysis and some experimental results concerning the effects of both weights and activation function actual implementation, for feedforward multi-layer digital neural networks. The analysis allows us to derive some practical boundaries on the number of bits for the weight representation and on the number of levels for a stepwise approximation of the activation function (however, the analysis can be applied to different implementation technologies). By following such a rules, the equivalence between the abstract neural model and an actual cost-effective VLSI implementation is a priori guaranteed
Keywords :
VLSI; digital integrated circuits; feedforward neural nets; neural chips; abstract neural model; activation function; cost-effective VLSI implementation; digital neural networks; feedforward multilayer neural networks; stepwise approximation; weights; Costs; Electronic mail; Feedforward neural networks; Feeds; Multi-layer neural network; Neural networks; Neurons; Piecewise linear approximation; Proposals; Table lookup; Very large scale integration;
Conference_Titel :
Circuits and Systems, 1995., Proceedings., Proceedings of the 38th Midwest Symposium on
Print_ISBN :
0-7803-2972-4
DOI :
10.1109/MWSCAS.1995.504383
