Back to search

Hosted by Utah State University Libraries

Neural Networks and the Natural Gradient

Abstract

dc:description.abstract

Neural network training algorithms have always suffered from the problem of local minima. The advent of natural gradient algorithms promised to overcome this shortcoming by finding better local minima. However, they require additional training parameters and computational overhead. By using a new formulation for the natural gradient, an algorithm is described that uses less memory and processing time than previous algorithms with comparable performance.

Degree

thesis:*
Name thesis:degree_name
Doctor of Philosophy (PhD)
Level thesis:degree_level
Dissertation
Discipline thesis:degree_discipline
Electrical and Computer Engineering
Grantor dc:publisher
Hosted by Utah State University Libraries

Author and committee

dc:creator, dc:contributor.*
Author dc:creator
  • Bastian, Michael R.
Contributors dc:contributor
  • Jacob H. Gunther
  • Todd K. Moon
  • YangQuan Chen

Subjects

dc:subject × 7

Identifiers

dc:identifier.*

Chain of custody

source
Harvested from
Utah State University
Base URL
digitalcommons.usu.edu/do/oai/
Last updated
2026-11-19
Source record
OAI-PMH GetRecord
citation

Bastian, Michael R.. Neural Networks and the Natural Gradient. Dissertation thesis, Hosted by Utah State University Libraries, https://digitalcommons.usu.edu/etd/539