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

A self-organising network that grows when required.

Stephen Marsland1, Jonathan Shapiro, Ulrich Nehmzow

  • 1Division of Imaging Science and Biomedical Engineering, University of Manchester, UK. smarsland@cs.man.ac.uk

Neural Networks : the Official Journal of the International Neural Network Society
|November 6, 2002
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

Tokens enable cooperation without identification or memory.

Proceedings. Biological sciences·2026
Same author

Darier disease-A review highlighting new insights from the Darier Disease International Task Force.

Journal of the European Academy of Dermatology and Venereology : JEADV·2026
Same author

Publication trends in vitiligo research from 1991 to 2024: a comprehensive bibliometric analysis.

Dermatology reports·2026
Same author

Neonatal Stabilization Team: A Critical Team Providing Neonatal Care for Military Families in Guam.

Military medicine·2026
Same author

Beyond black-box AI: Comparing ChatGPT-4 interpretability and accuracy to CNNs in melanocytic lesions diagnosis.

Journal der Deutschen Dermatologischen Gesellschaft = Journal of the German Society of Dermatology : JDDG·2026
Same author

The Role of AI-Generated Clinical Image Descriptions in Enhancing Teledermatology Diagnosis: A Cross-Sectional Exploratory Study.

Diagnostics (Basel, Switzerland)·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
Same journal

Dynamic analysis and reliable mechanical optimization application of ring HNN effected with a memristive neuron.

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

DAFF-Net: A detection and search method for small-scale low surface brightness galaxies.

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

Quasi-synchronization for complex networks with hybrid pinning intermittent control.

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

This study introduces a novel growing neural network that dynamically adds nodes to better map input data. This self-organizing approach improves accuracy and adapts to changing data distributions, outperforming existing methods in novelty detection tasks.

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Computational Neuroscience

Background:

  • Traditional neural networks often have fixed structures, limiting their ability to adapt to complex or changing data.
  • Existing growing neural networks typically add nodes based on error thresholds or fixed iteration intervals.
  • Dynamic input distributions pose a challenge for static network architectures.

Purpose of the Study:

  • To develop a self-organizing neural network capable of dynamically growing its structure.
  • To enable the network to adapt more accurately and parsimoniously to input spaces.
  • To address the limitations of fixed-size networks in handling dynamic data distributions.

Main Methods:

  • A novel learning algorithm that adds nodes when the network's current state insufficiently matches the input.

Related Experiment Videos

  • Demonstration of neighborhood relations preservation within the network's map space.
  • Comparative analysis against the Growing Neural Gas (GNG) on artificial datasets with changing input distributions.

    • Main Results:

    • The proposed network exhibits rapid growth when presented with new data, halting growth once data is sufficiently matched.
    • Preservation of neighborhood relations in the data is demonstrated.
    • Effective handling of changes in input distribution was shown in comparative studies.

    Conclusions:

    • The developed growing neural network offers a more adaptive and accurate approach to approximating input spaces compared to fixed-size networks.
    • The network's ability to dynamically adjust its structure is crucial for handling evolving data patterns.
    • The proposed method shows promise in novelty detection tasks, outperforming established benchmarks.