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

Efficient learning algorithms for three-layer regular feedforward fuzzy neural networks.

Puyin Liu1, Hongxing Li

  • 1Department of Mathematics, Beijing Normal University, Beijing 100875, China. liupuyin@public.cs.hn.cn

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

Endogenous peptide derived from c-Cbl-associated protein counteracts its inhibitory effect on enteric neural crest cell colonization in Hirschsprung disease.

The Journal of neuroscience : the official journal of the Society for Neuroscience·2026
Same author

Factors associated with <i>de novo</i> aneuploidy across different preimplantation genetic testing cycles: a retrospective cohort study.

Frontiers in endocrinology·2026
Same author

Reliability and sensitivity of mechanical properties of phase-specific during the 505-test using a radar device in soccer players.

Frontiers in physiology·2026
Same author

A LASSO-based nomogram for predicting acute bilirubin encephalopathy in newborns with severe hyperbilirubinemia.

Frontiers in pediatrics·2026
Same author

Delocalized Electronic States Induced by Phosphorus Doping Suppress Charge Recombination in Cu<sub>2</sub>ZnSnS<sub>4</sub> Photocathodes.

Small (Weinheim an der Bergstrasse, Germany)·2026
Same author

Targeting WWP1 ameliorates osteoarthritis by suppressing KLF15 ubiquitination to restore mtDNA and mitochondrial function.

Molecular genetics and genomics : MGG·2026

This study develops efficient learning algorithms for fuzzy neural networks (FNNs) by addressing the differentiation of max-min functions. Novel fuzzy conjugate gradient and back-propagation algorithms are presented for improved convergence and application to fuzzy inference.

Area of Science:

  • Artificial Intelligence
  • Computational Intelligence
  • Fuzzy Systems

Background:

  • Gradient descent methods are crucial for training feedforward fuzzy neural networks (FNNs).
  • A key challenge lies in differentiating max-min functions inherent in FNN operations.
  • Existing methods often lack efficiency or generalizability for complex fuzzy number representations.

Purpose of the Study:

  • To investigate the differentiation of max-min functions in the context of FNNs.
  • To define a regular FNN using general fuzzy numbers and their level sets.
  • To develop and analyze efficient fuzzy learning algorithms for FNNs.

Main Methods:

  • The study employs general fuzzy numbers, approximated by finite level sets, to define a three-layer FNN.

Related Experiment Videos

  • A fuzzy back-propagation algorithm is presented for network training.
  • A fuzzy conjugate gradient algorithm is developed to accelerate convergence, with systematic analysis of its convergence properties.

    • Main Results:

    • The input-output relationship of the FNN is approximated using functions of finite level set endpoints.
    • The proposed fuzzy conjugate gradient algorithm demonstrates faster convergence compared to standard methods.
    • Simulations confirm the efficiency of the developed learning algorithms for FNNs.

    Conclusions:

    • The research provides effective methods for differentiating max-min functions, enabling robust FNN training.
    • The developed fuzzy learning algorithms, particularly the conjugate gradient approach, enhance training efficiency and convergence.
    • The FNNs are successfully applied to approximate fuzzy inference rules and functions on compact sets.