Regularization and complexity control in feed-forward networks


In this paper we consider four alternative approaches to complexity control in feed-forward networks based respectively on architecture selection, regularization, early stopping, and training with noise. We show that there are close similarities between these approaches and we argue that, for most practical applications, the technique of regularization should be the method of choice.

Divisions: Aston University (General)
Additional Information: International Conference on Artificial Neural Networks ICANN'95.
Uncontrolled Keywords: NCRG complexity control feed-forward networks architecture selection regularization early stopping training with noise
ISBN: 2-910085-19-8
Last Modified: 09 Feb 2024 08:00
Date Deposited: 17 Oct 2011 13:22
PURE Output Type: Chapter
Published Date: 1995
Authors: Bishop, Christopher M.


Export / Share Citation


Additional statistics for this record