Jove
Visualize
Contact Us

Related Experiment Videos

Stochastic competitive learning.

B Kosko1

  • 1Dept. of Electr. Eng., Univ. of Southern California, Los Angeles, CA.

IEEE Transactions on Neural Networks
|January 1, 1991
PubMed
Summary
This summary is machine-generated.

This study analyzes competitive learning systems as stochastic dynamical systems. These systems efficiently estimate probability density functions and converge to pattern class centroids, simplifying competitive learning to stochastic gradient descent.

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

Adaptive bidirectional associative memories.

Applied optics·2010
Same author

Robust stochastic resonance: signal detection and adaptation in impulsive noise.

Physical review. E, Statistical, nonlinear, and soft matter physics·2001
Same author

Fuzzy function approximation with ellipsoidal rules.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·1996
Same author

Adaptive fuzzy systems for backing up a truck-and-trailer.

IEEE transactions on neural networks·1992
Same author

Differential competitive learning for centroid estimation and phoneme recognition.

IEEE transactions on neural networks·1991
Same author

Unsupervised learning in noise.

IEEE transactions on neural networks·1990
Same journal

Universal perceptron and DNA-like learning algorithm for binary neural networks: LSBF and PBF implementations.

IEEE transactions on neural networks·2013
Same journal

Guest editorial: special section on white box nonlinear prediction models.

IEEE transactions on neural networks·2011
Same journal

Data-based fault-tolerant control of high-speed trains with traction/braking notch nonlinearities and actuator failures.

IEEE transactions on neural networks·2011
Same journal

Guest editorial: special section on data-based control, modeling, and optimization.

IEEE transactions on neural networks·2011
Same journal

Neural network-based multiple robot simultaneous localization and mapping.

IEEE transactions on neural networks·2011
Same journal

Data-driven model-free adaptive control for a class of MIMO nonlinear discrete-time systems.

IEEE transactions on neural networks·2011
See all related articles
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

Area of Science:

  • Computational neuroscience
  • Machine learning theory

Background:

  • Competitive learning systems are crucial for pattern recognition and density estimation.
  • Understanding their dynamical behavior is key to improving their efficiency and convergence properties.

Purpose of the Study:

  • To analyze competitive learning systems as stochastic dynamical systems.
  • To investigate the convergence properties of synaptic vectors in feedforward competitive neural networks.
  • To establish a link between competitive learning and stochastic gradient descent.

Main Methods:

  • Formulation of continuous and discrete competitive learning systems (unsupervised, supervised, differential).
  • Analysis using stochastic dynamical systems theory.
  • Application of stochastic Lyapunov arguments to prove convergence.

Related Experiment Videos

Main Results:

  • Competitive learning systems act as adaptive vector quantizers, estimating probability density functions.
  • Synaptic vectors in feedforward networks converge exponentially quickly to pattern class centroids.
  • Competitive learning is shown to be equivalent to stochastic gradient descent, independent of specific neuronal activation dynamics.

Conclusions:

  • The convergence of competitive learning systems is robust and efficient.
  • These findings provide a theoretical foundation for competitive learning algorithms.
  • The framework extends to estimating higher-order statistics and covariances.