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

Neocognitron capable of incremental learning.

Kunihiko Fukushima1

  • 1School of Media Science, Tokyo University of Technology, 1404-1 Katakura Hachioji, Tokyo 192-0982, Japan. fukushima@media.teu.ac.jp

Neural Networks : the Official Journal of the International Neural Network Society
|December 24, 2003
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

Efficient IntVec: High recognition rate with reduced computational cost.

Neural networks : the official journal of the International Neural Network Society·2019
Same author

Introduction.

International journal of neural systems·2018
Same author

Margined winner-take-all: New learning rule for pattern recognition.

Neural networks : the official journal of the International Neural Network Society·2017
Same author

Training multi-layered neural network neocognitron.

Neural networks : the official journal of the International Neural Network Society·2013
Same author

Artificial vision by multi-layered neural networks: neocognitron and its advances.

Neural networks : the official journal of the International Neural Network Society·2012
Same author

Increasing robustness against background noise: visual pattern recognition by a neocognitron.

Neural networks : the official journal of the International Neural Network Society·2011
Same journal

TraNce: Type-aware hypergraph neural network with biological mediators for drug repositioning.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Decentralized ADMM for factorization-based Low-rank matrix estimation.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Memristive neuromorphic circuit design inspired by the neural mechanisms of conditioned fear.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Q-learning based asynchronous Boolean control networks stabilization with data loss.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

New results on prescribed-time synchronization of complex networks via intermittent control.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Variance-constrained multi-view ensemble broad network for imbalanced data.

Neural networks : the official journal of the International Neural Network Society·2026
See all related articles

This study introduces a novel neocognitron capable of incremental learning without compromising existing memories or learning speed. Its competitive learning approach enables simultaneous layer progression, preventing data loss and enhancing efficiency.

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Computational Neuroscience

Background:

  • Conventional neocognitrons often sacrifice incremental learning ability for faster training speeds.
  • Sequential layer construction in traditional models can lead to inefficiencies and data loss.
  • High learning speeds in conventional neocognitrons can result in 'garbage cells' that hinder performance.

Purpose of the Study:

  • To develop a new neocognitron model that supports incremental learning.
  • To maintain learning speed without damaging previously acquired knowledge.
  • To prevent the formation of non-functional 'garbage cells' during rapid training.

Main Methods:

  • Implementation of a novel neocognitron architecture.
  • Utilizing a competitive learning mechanism for network training.

Related Experiment Videos

  • Simultaneous progression of learning across all hierarchical network stages.

    • Main Results:

    • The proposed neocognitron successfully integrates incremental learning capabilities.
    • Existing memory integrity is preserved during the learning process.
    • The model avoids the generation of 'garbage cells' even at high learning speeds.
    • Simultaneous layer learning progresses efficiently without performance degradation.

    Conclusions:

    • The new neocognitron model offers a significant advancement in artificial neural network design.
    • It provides a robust solution for incremental learning in complex hierarchical networks.
    • This approach enhances both learning efficiency and memory retention in artificial systems.