![](/images/graphics-bg.png)
A Simplified Natural Gradient Learning Algorithm
Joint Authors
Bastian, Michael R.
Moon, Todd K.
Gunther, Jacob H.
Source
Advances in Artificial Neural Systems
Issue
Vol. 2011, Issue 2011 (31 Dec. 2011), pp.1-9, 9 p.
Publisher
Hindawi Publishing Corporation
Publication Date
2011-07-24
Country of Publication
Egypt
No. of Pages
9
Main Subjects
Information Technology and Computer Science
Abstract EN
Adaptive natural gradient learning avoids singularities in the parameter space of multilayer perceptrons.
However, it requires a larger number of additional parameters than ordinary backpropagation in the form of the Fisher information matrix.
This paper describes a new approach to natural gradient learning that uses a smaller Fisher information matrix.
It also uses a prior distribution on the neural network parameters and an annealed learning rate.
While this new approach is computationally simpler, its performance is comparable to that of adaptive natural gradient learning.
American Psychological Association (APA)
Bastian, Michael R.& Gunther, Jacob H.& Moon, Todd K.. 2011. A Simplified Natural Gradient Learning Algorithm. Advances in Artificial Neural Systems،Vol. 2011, no. 2011, pp.1-9.
https://search.emarefa.net/detail/BIM-469689
Modern Language Association (MLA)
Bastian, Michael R.…[et al.]. A Simplified Natural Gradient Learning Algorithm. Advances in Artificial Neural Systems No. 2011 (2011), pp.1-9.
https://search.emarefa.net/detail/BIM-469689
American Medical Association (AMA)
Bastian, Michael R.& Gunther, Jacob H.& Moon, Todd K.. A Simplified Natural Gradient Learning Algorithm. Advances in Artificial Neural Systems. 2011. Vol. 2011, no. 2011, pp.1-9.
https://search.emarefa.net/detail/BIM-469689
Data Type
Journal Articles
Language
English
Notes
Includes bibliographical references
Record ID
BIM-469689