Stabilities in optimal cluster separation networks

Two clusters of normalized vectors are optimally separated in a neural network for which thresholds and weights are trained to give maximum pattern stability. The performance of local, iterative algorithms that treat threshold and weights all in one and converge to the optimally stable network is in...

Full description

Saved in:
Bibliographic Details
Published in:Neural networks 1995-01, Vol.8 (3), p.387-390
Main Author: Wendemuth, Andreas
Format: Article
Language:English
Subjects:
Applied sciences
Artificial intelligence
Bias
Comparison with replica calculation
Computer science
control theory
systems
Connectionism. Neural networks
Exact sciences and technology
Finite size effects
Learning algorithms
Learning and adaptive systems
Optimal cluster separation
Optimally set threshold
Pattern stability
Perceptron
Citations: Items that this one cites
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Two clusters of normalized vectors are optimally separated in a neural network for which thresholds and weights are trained to give maximum pattern stability. The performance of local, iterative algorithms that treat threshold and weights all in one and converge to the optimally stable network is investigated. It is shown that the separation/stability obtained matches theoretical predictions and is superior to existing algorithms, even at small system sizes.
ISSN:0893-6080
1879-2782
DOI:10.1016/0893-6080(94)00075-W