ITEM METADATA RECORD
Title: Topology-preserving map formation achieved with a purely local unsupervised competitive learning rule
Authors: Van Hulle, Marc # ×
Issue Date: 1997
Publisher: PERGAMON-ELSEVIER SCIENCE LTD
Series Title: Neural networks vol:10 issue:3 pages:431-446
Abstract: An unsupervised competitive learning rule is introduced for topology-preserving map formation. The rule, called vectorial boundary adaptation rule (VBAR), achieves a ''maximally''-ordered map by performing local weight updates only: hence, contrary to Kohonen's self-organizing map (SOM) algorithm, no neighborhood function is needed. We formalize VBAR's learning dynamics and show by simulations that it converges in a completely different way from the SOM algorithm. Our rule is aimed at achieving an equiprobable quantization of the input space (unconditional information-theoretic entropy maximization). Due to this property, we argue that VBAR is able to generate a ''sparse-distributed'' representation. Finally, we argue that entropy maximization is a plausible computational principle for topology-preserving map formation and that VBAR is a minimal assumption rule in this respect. (C) 1997 Elsevier Science Ltd.
ISSN: 0893-6080
Publication status: published
KU Leuven publication type: IT
Appears in Collections:Research Group Neurophysiology
× corresponding author
# (joint) last author

Files in This Item:

There are no files associated with this item.

Request a copy

 




All items in Lirias are protected by copyright, with all rights reserved.

© Web of science