Sensitivity analysis for feedforward artificial neural networks with differentiable activation functions
- 2 January 2003
- conference paper
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
Abstract
A method for computing the network output sensitivities with respect to variations in the inputsfor multilayer feedforward artificial neural networks with differentiable activation functionsis presented. It is applied to obtain expressions for the first and second order sensitivities. Anexample is introduced along with a discussion to illustrate how the sensitivities are calculatedand to show how they compare to the actual derivatives of the function being modeled by theneural network.1...Keywords
This publication has 10 references indexed in Scilit:
- Use of influence diagrams and neural networks in modeling LPCVDPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- Learning a nonlinear model of a manufacturing process using multilayer connectionist networksPublished by Institute of Electrical and Electronics Engineers (IEEE) ,2002
- A sensitivity analysis of a back-propagation neural network for manufacturing process parametersJournal of Intelligent Manufacturing, 1991
- Modeling chemical process systems via neural computationIEEE Control Systems Magazine, 1990
- Universal approximation of an unknown mapping and its derivatives using multilayer feedforward networksNeural Networks, 1990
- Multilayer feedforward networks are universal approximatorsNeural Networks, 1989
- Approximation by superpositions of a sigmoidal functionMathematics of Control, Signals, and Systems, 1989
- Theory of the backpropagation neural networkPublished by Institute of Electrical and Electronics Engineers (IEEE) ,1989
- On the approximate realization of continuous mappings by neural networksNeural Networks, 1989
- Parallel Distributed ProcessingPublished by MIT Press ,1986