DocumentCode :
467844
Title :
Comparison Study of Sensitivity Definitions of Neural Networks
Author :
Li, Chun-guo ; Li, Hai-feng ; Yao, Ai-Ke ; Xu, Ning
Author_Institution :
Hebei Univ., Baoding
Volume :
6
fYear :
2007
fDate :
19-22 Aug. 2007
Firstpage :
3472
Lastpage :
3477
Abstract :
This paper compares the sensitivity definitions of neural networks´ output to input and weight perturbations. Based on the essence of the sensitivity definitions, the authors classify these sensitivity definitions into 3 categories: Noise-to-Signal Ratio, Geometrical property of derivative, Angle perturbation in Geometry space. The characteristics of these 3 categories of sensitivity definition are discussed respectively. It is sensible to classify these sensitivity definitions based on the essence of them for researchers can find other new sensitivity definitions of neural networks.
Keywords :
neural nets; angle perturbation; geometry space; neural networks; sensitivity definitions; Computer science; Cybernetics; Geometry; Machine learning; Mathematics; Neural networks; Random variables; Sensitivity analysis; Signal to noise ratio; Working environment noise; Categories; Comparison; Neural networks; Sensitivity definition;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Machine Learning and Cybernetics, 2007 International Conference on
Conference_Location :
Hong Kong
Print_ISBN :
978-1-4244-0973-0
Electronic_ISBN :
978-1-4244-0973-0
Type :
conf
DOI :
10.1109/ICMLC.2007.4370748
Filename :
4370748
Link To Document :
بازگشت