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 nets for optimization.

Michele Milano1, Petros Koumoutsakos, Jürgen Schmidhuber

  • 1Graduate Aeronautics Laboratories, California Institute of Technology, Pasadena, CA 91125, USA. milano@galcit.caltech.edu

IEEE Transactions on Neural Networks
|September 24, 2004
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

Optimized mechano-fluidic metamaterials inspired by deep-sea sponges.

Nature communications·2026
Same author

Ising energy model for the stochastic prediction of tumor islets.

ArXiv·2025
Same author

Individualizing glioma radiotherapy planning by optimization of a data and physics-informed discrete loss.

Nature communications·2025
Same author

Quantitative 3D histochemistry reveals region-specific amyloid-β reduction by the antidiabetic drug netoglitazone.

PloS one·2025
Same author

Accelerating the inference of string generation-based chemical reaction models for industrial applications.

Journal of cheminformatics·2025
Same author

emg2vec: Self-supervised Pretraining in Electromyography-based Silent Speech Interfaces.

Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference·2025

This study introduces adaptive neural networks for efficient optimization, improving candidate selection in complex search spaces. The novel approach enhances performance over existing methods and achieves significant drag reduction in fluid dynamics.

Area of Science:

  • Computational intelligence
  • Optimization algorithms
  • Machine learning

Background:

  • Efficiently selecting solution candidates in expensive optimization problems is challenging.
  • Existing methods may struggle with local optima and reliability.
  • Need for adaptive strategies to guide search space exploration.

Purpose of the Study:

  • To develop novel adaptive algorithms for efficient candidate selection in optimization.
  • To leverage learning algorithms inspired by self-organizing maps and neural gas.
  • To improve the exploitation of promising search space regions.

Main Methods:

  • Embedding optimization strategies into learning algorithms (adaptive nets/grids).
  • Utilizing node attraction to promising candidates for biased data selection.

Related Experiment Videos

  • Employing techniques inspired by Kohonen's self-organizing maps and neural gas networks.

    • Main Results:

    • Adaptive nets effectively identify and exploit regions with higher probability of finding optima.
    • The proposed method demonstrates more reliable performance than covariance matrix adaptation evolution strategy on benchmark functions.
    • Achieved unprecedented drag reduction in actively controlled flow past a circular cylinder.

    Conclusions:

    • The adaptive learning approach offers a robust and efficient method for complex optimization tasks.
    • This technique provides a powerful tool for both theoretical optimization and practical engineering problems like drag reduction.