Academic Journals Database
Disseminating quality controlled scientific knowledge

Empirical Study of Least Sensitive FFANN for Weight- Stuck-at Zero Fault

ADD TO MY LIST
 
Author(s): Amit Prakash Singh | Pravin Chandra | Chandra Shekhar Rai

Journal: International Journal of Computer Applications
ISSN 0975-8887

Volume: 2;
Issue: 2;
Start page: 47;
Date: 2010;
VIEW PDF   PDF DOWNLOAD PDF   Download PDF Original page

Keywords: Artificial Neural Network | Fault models | Sensitivity analysis.

ABSTRACT
An important consideration for neural hardware is its sensitivityto input and weight errors. In this paper, an empirical study isperformed to analyze the sensitivity of feedforward neuralnetworks for Gaussian noise to input and weight. 30 numbers ofFFANN is taken for four different classification tasks. Leastsensitive network for input and weight error is chosen for furtherstudy of fault tolerant behavior of FFANN. Weight stuck-at zerofault is selected to study error metrics of fault tolerance.Empirical results for a WSZ fault is demonstrated in this paper.
Affiliate Program      Why do you need a reservation system?