Computation of Madalines' Sensitivity to Input and Weight Perturbations

Authors:
Yingfeng Wang;Xiaoqin Zeng;Daniel So Yeung;Zhihang Peng
Affiliations:
Department of Computer Science, Hohai University, Nanjing, China ywang802@vip.sina.com;Department of Computer Science, Hohai University, Nanjing, China xzeng@hhu.edu.cn;Department of Computing, Hong Kong Polytechnic University, Kowloon, Hong Kong csdaniel@comp.polyu.edu.hk;Department of Mathematics, Hohai University, Nanjing, China zhihangpeng@hotmail.com
Venue:
Neural Computation
Year:
2006

Citing 9
Cited 1

A quantified sensitivity measure for multilayer perceptron to input perturbation

Neural Computation
Hidden neuron pruning of multilayer perceptrons using a quantified sensitivity measure

Neurocomputing
Sensitivity analysis of neocognitron

IEEE Transactions on Systems, Man, and Cybernetics, Part C: Applications and Reviews
Sensitivity analysis of multilayer perceptron to input and weight perturbations

IEEE Transactions on Neural Networks
A new pruning heuristic based on variance analysis of sensitivity information

IEEE Transactions on Neural Networks
Using function approximation to analyze the sensitivity of MLP with antisymmetric squashing activation function

IEEE Transactions on Neural Networks
The selection of weight accuracies for Madalines

IEEE Transactions on Neural Networks
Sensitivity analysis of single hidden-layer neural networks with threshold functions

IEEE Transactions on Neural Networks
Sensitivity Analysis for Selective Learning by Feedforward Neural Networks

Fundamenta Informaticae

Computation of multilayer perceptron sensitivity to input perturbation

Neurocomputing

Quantified Score

Hi-index	0.01

Visualization

Abstract

The sensitivity of a neural network's output to its input and weight perturbations is an important measure for evaluating the network's performance. In this letter, we propose an approach to quantify the sensitivity of Madalines. The sensitivity is defined as the probability of output deviation due to input and weight perturbations with respect to overall input patterns. Based on the structural characteristics of Madalines, a bottomup strategy is followed, along which the sensitivity of single neurons, that is, Adalines, is considered first and then the sensitivity of the entire Madaline network. Bymeans of probability theory, an analytical formula is derived for the calculation of Adalines' sensitivity, and an algorithm is designed for the computation of Madalines' sensitivity. Computer simulations are run to verify the effectiveness of the formula and algorithm. The simulation results are in good agreement with the theoretical results.