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

Guaranteed two-pass convergence for supervised and inferential learning.

M J Healy1, T P Caudell

  • 1Research and Technology, The Boeing Company, Seattle, WA 98124-2207, USA.

IEEE Transactions on Neural Networks
|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

Variation in the long-tailed field-mouse (Apodemus sylvaticus (L>)) in the Channel Islands.

Proceedings of the Royal Society of London. Series B, Biological sciences·2014
Same author

Molecular insights into the evolution of an enzyme; esterase6 in Drosophila.

Trends in ecology & evolution·2011
Same author

Reply from j.g. Oakeshott et Al.

Trends in ecology & evolution·2011
Same author

Photorefractive adaptive resonance neural network.

Applied optics·2010
Same author

Limitations of a class of binary phase-only filters.

Applied optics·2010
Same author

Hybrid optoelectronic adaptive resonance theory neural processor, ART1.

Applied optics·2010
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

We proved LAPART 2, an adaptive inferencing neural network, converges in just two passes. This fixed-pass convergence avoids template proliferation, unlike ARTMAP

Area of Science:

  • Machine Learning
  • Artificial Intelligence
  • Neural Networks

Background:

  • Adaptive inferencing neural networks are crucial for efficient data processing.
  • Existing architectures like ARTMAP have limitations in convergence speed and template proliferation.
  • Supervised learning typically uses label-affixing, limiting applicability to complex data.

Purpose of the Study:

  • To theoretically analyze a novel adaptive inferencing neural network architecture, LAPART 2.
  • To establish the convergence properties and efficiency of LAPART 2.
  • To demonstrate the advantages of LAPART 2 over existing models like ARTMAP.

Main Methods:

  • Theoretical analysis of the LAPART 2 neural network architecture.
  • Mathematical proof of convergence within a fixed number of passes.

Related Experiment Videos

  • Analysis of template proliferation to ensure model efficiency.

    • Main Results:

    • LAPART 2 demonstrates convergence in only two passes through a fixed training set.
    • The LAPART 2 architecture is proven to not suffer from template proliferation.
    • This fixed-pass convergence contrasts with ARTMAP's n-pass convergence (where n is input space size).

    Conclusions:

    • LAPART 2 offers a significant advancement in neural network convergence efficiency.
    • The architecture's fixed-pass convergence and lack of template proliferation are key benefits.
    • The findings extend to set-valued mappings, broadening applicability beyond traditional supervised learning.