Display options
Share it on

IEEE Trans Cybern. 2019 Aug;49(8):3088-3098. doi: 10.1109/TCYB.2018.2838680. Epub 2018 Jun 08.

Fisher Information Matrix of Unipolar Activation Function-Based Multilayer Perceptrons.

IEEE transactions on cybernetics

Weili Guo, Yew-Soon Ong, Yingjiang Zhou, Jaime Rubio Hervas, Aiguo Song, Haikun Wei

PMID: 29994240 DOI: 10.1109/TCYB.2018.2838680

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.

Publication Types