Please use this identifier to cite or link to this item:
Title: Fisher information matrix of unipolar activation function-based multilayer perceptrons
Authors: Guo, Weili
Ong, Yew-Soon
Zhou, Yingjiang
Hervas, Jaime Rubio
Song, Aiguo
Wei, Haikun
Keywords: Engineering::Computer science and engineering
Issue Date: 2018
Source: Guo, W., Ong, Y.-S., Zhou, Y., Hervas, J. R., Song, A., & Wei, H. (2019). Fisher information matrix of unipolar activation function-based multilayer perceptrons. IEEE Transactions on Cybernetics, 49(8), 3088-3098. doi:10.1109/TCYB.2018.2838680
Journal: IEEE Transactions on Cybernetics
Abstract: The multilayer perceptrons (MLPs) are widely used in many fields, however, singularities in the parameter space may seriously influence the learning dynamics of MLPs and cause strange learning behaviors. Given that the singularities are the subspaces of the parameter space where the Fisher information matrix (FIM) degenerates, the FIM plays a key role in the study of the singular learning dynamics of the MLPs. In this paper, we obtain the analytical form of the FIM for unipolar activation function-based MLPs where the input subjects to the Gaussian distribution with general covariance matrix and the unipolar error function is chosen as the activation function. Then three simulation experiments are taken to verify the validity of the obtained results.
ISSN: 2168-2267
DOI: 10.1109/TCYB.2018.2838680
Rights: © 2018 IEEE. All rights reserved.
Fulltext Permission: none
Fulltext Availability: No Fulltext
Appears in Collections:SCSE Journal Articles

Citations 20

Updated on Mar 10, 2021

Citations 20

Updated on Mar 8, 2021

Page view(s)

Updated on May 17, 2022

Google ScholarTM




Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.