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

Toward optimizing a self-creating neural network.

J H Wang1, J D Rau, C Y Peng

  • 1Dept. of Electr. Eng., Nat. Taiwan Ocean Univ., Keelung.

IEEE Transactions on Systems, Man, and Cybernetics. Part B, Cybernetics : a Publication of the IEEE Systems, Man, and Cybernetics Society
|February 7, 2008
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

[Experimental study on the effects of 810-nm diode low-level laser on oxidative stress and wound healing].

Zhonghua kou qiang yi xue za zhi = Zhonghua kouqiang yixue zazhi = Chinese journal of stomatology·2026
Same author

[Neuroimmune mechanism of rosacea].

Zhonghua yu fang yi xue za zhi [Chinese journal of preventive medicine]·2024
Same author

[Analysis on epidemiological characteristics of allergens with allergic skin diseases in a hospital in Beijing City from 2017 to 2021].

Zhonghua yu fang yi xue za zhi [Chinese journal of preventive medicine]·2024
Same author

[Current status and prospects of non-surgical treatment for fecal incontinence].

Zhonghua wei chang wai ke za zhi = Chinese journal of gastrointestinal surgery·2023
Same author

[Validation study of the Chinese version of atopic dermatitis control tool].

Zhonghua yu fang yi xue za zhi [Chinese journal of preventive medicine]·2023
Same author

Clinical outcomes after interruption of entecavir therapy in HBeAg-negative chronic hepatitis B patients with compensated cirrhosis.

Alimentary pharmacology & therapeutics·2015
Same journal

Strategic Ability Updating in Concurrent Games by Coalitional Commitment.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2015
Same journal

Meta-Analysis of the First Facial Expression Recognition Challenge.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2012
Same journal

Adjustable model-based fusion method for multispectral and panchromatic images.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2012
Same journal

Face Feature Weighted Fusion Based on Fuzzy Membership Degree for Video Face Recognition.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2012
Same journal

A New Adaptive Fast Cellular Automaton Neighborhood Detection and Rule Identification Algorithm.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2012
Same journal

Human-arm-and-hand-dynamic model with variability analyses for a stylus-based haptic interface.

IEEE transactions on systems, man, and cybernetics. Part B, Cybernetics : a publication of the IEEE Systems, Man, and Cybernetics Society·2012
See all related articles

Optimizing Growing Cell Structures (GCS) involves a conservation principle in resource counters. Setting the decay factor alpha to zero maximizes information entropy for topology learning and minimizes resource use for vector quantization.

Area of Science:

  • Machine Learning
  • Computational Neuroscience
  • Artificial Intelligence

Background:

  • Growing Cell Structures (GCS) models are used for topology learning and vector quantization.
  • Each GCS node has a resource counter that increments for the best-matching node and decays over time.
  • The conservation of the sum of resource counters is a key characteristic of the GCS model.

Purpose of the Study:

  • To optimize the performance of the Growing Cell Structures (GCS) model.
  • To investigate the impact of resource counter dynamics on GCS learning.
  • To enhance topology learning and vector quantization using information entropy principles.

Main Methods:

  • Analyzing the conservation principle of resource counters in GCS.
  • Applying information entropy to evaluate GCS performance.

Related Experiment Videos

  • Implementing a threshold-free node-removal scheme with alpha=0.

    • Main Results:

    • The summation of all resource counters in GCS is conserved.
    • Optimizing GCS performance is achieved with alpha=0 and a threshold-free node-removal scheme.
    • This optimization is effective for both stationary and nonstationary input data.

    Conclusions:

    • The conservation principle offers insights into GCS optimization.
    • Setting alpha=0 maximizes information entropy for topology learning (equi-probable criterion).
    • Setting alpha=0 minimizes resource usage for vector quantization (equi-error criterion).