Growing neural gas

From Wikipedia, the free encyclopedia

Growing Neural Gas is a self organization neural network first proposed by Bernd Fritzke. Unlike the earlier Neural Gas, Growing Neural Gas (GNG) can add and delete nodes during algorithm execution. The growth mechanism is based on Growing Cell Structures and competitive hebbian learning


Compared to Neural Gas, GNG has the following distinctions:

1. Ability to add and delete nodes.

2. Local Error measurements are noted at each step helping it to locally insert/delete nodes.

3. Edges are connected between nodes, so a sufficiently old edge is deleted. Such edges are intended placeholders for localized data distribution.

4. Such edges also help to locate distinct clusters (those clusters are not connected by edges).

[edit] See also

[edit] References and sources

  • Fritzke's site [1]
  • Paper: B. Fritzke.Fast learning with incremental RBF networks.Neural Processing Letters, 1(1):-5, 1994b.

[edit] External links

  • Growing Neural Gas videos Three videos that show how neural gas grow inside a 3d structure.
  • DemoGNG An interactive applet showing how growing neural gas compares to other similar algorithms.
  • Java Competitve Learning Application A suite of Unsupervised Neural Networks (including Growing Neural Gas) written in Java. Complete with all source code.