Jove
Visualize
Contact Us
JoVE
x logofacebook logolinkedin logoyoutube logo
ABOUT JoVE
OverviewLeadershipBlogJoVE Help Center
AUTHORS
Publishing ProcessEditorial BoardScope & PoliciesPeer ReviewFAQSubmit
LIBRARIANS
TestimonialsSubscriptionsAccessResourcesLibrary Advisory BoardFAQ
RESEARCH
JoVE JournalMethods CollectionsJoVE Encyclopedia of ExperimentsArchive
EDUCATION
JoVE CoreJoVE BusinessJoVE Science EducationJoVE Lab ManualFaculty Resource CenterFaculty Site
Terms & Conditions of Use
Privacy Policy
Policies

Related Experiment Videos

Self-organizing maps: ordering, convergence properties and energy functions.

E Erwin1, K Obermayer, K Schulten

  • 1Beckman Institute, University of Illinois, Urbana-Champaign 61801.

Biological Cybernetics
|January 1, 1992
PubMed
Summary
This summary is machine-generated.

Related Concept Videos

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by
Same author

Dynamic Behavior of Trigger Factor on the Ribosome.

Journal of molecular biology·2016
Same author

Neural systems for choice and valuation with counterfactual learning signals.

NeuroImage·2013
Same author

The effect of novelty on reinforcement learning.

Progress in brain research·2013
Same author

How Quantum Coherence Assists Photosynthetic Light Harvesting.

The journal of physical chemistry letters·2012
Same author

Squeezing protein shells: how continuum elastic models, molecular dynamics simulations, and experiments coalesce at the nanoscale.

Biophysical journal·2010
Same author

Neurotransmitter systems in alcohol dependence.

Pharmacopsychiatry·2009

This study analyzes the convergence of self-organizing feature maps (SOFMs) for topographic mapping. Researchers extend convergence proofs and introduce a new potential function model for SOFM learning dynamics.

Area of Science:

  • Artificial Intelligence
  • Computational Neuroscience
  • Machine Learning

Background:

  • Self-organizing feature maps (SOFMs) are unsupervised neural networks used for dimensionality reduction and topographic mapping.
  • Convergence properties of SOFMs are crucial for understanding their stability and effectiveness.

Purpose of the Study:

  • To investigate the convergence properties of the SOFM algorithm for topographic representation of the unit interval.
  • To extend existing convergence proofs and provide a new theoretical framework for SOFM learning dynamics.

Main Methods:

  • Extending convergence proofs by Kohonen and Cottrell and Fort.
  • Developing a new model describing learning dynamics using a set of potential functions.
  • Deriving potential functions for one- and multi-dimensional cases.

Related Experiment Videos

Main Results:

  • Convergence proofs are extended to cases with monotonically decreasing neighborhood functions.
  • SOFM learning dynamics are shown to be a stochastic gradient descent on a set of potential functions, not a single energy function.
  • The derived potential functions are shown to be more accurate than previous energy function approximations, especially for disordered maps.

Conclusions:

  • The study provides a more accurate theoretical understanding of SOFM convergence and learning dynamics.
  • The findings have implications for designing and analyzing SOFMs for various applications in AI and neuroscience.
  • The new potential function framework offers a more robust approach to studying SOFM behavior.