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 second-order learning algorithm for multilayer networks based on block Hessian matrix.

Yi Jen Wang1, Chin Teng Lin

  • 1Department of Electrical and Control Engineering, National Chiao-Tung University, Hsinchu, Taiwan

Neural Networks : the Official Journal of the International Neural Network Society
|March 29, 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

$n$-Cylindrical Symbolic Response, a Standalone and Synergistic Biomarker for Epilepsy Diagnosis on EEG Modality.

IEEE journal of biomedical and health informatics·2025
Same journal

Exploiting audio-visual modalities in videos: Object detection via multi-stage bilateral coupling network.

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

Reliability-aware modality completion with cross-modal distillation for federated learning with missing modalities.

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

IGFD-Net: Illumination-guided frequency decoupling for polarization image fusion.

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

Multiple-Strategies dung beetle optimizer and its applications in engineering optimization and bankruptcy prediction.

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

Aggregating global-scale pixel-wise forgery cues within a graph.

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

Finite-Time intermittent control for secure synchronization of Neutral-Type stochastic delayed neural networks under aperiodic DoS attacks.

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

This study introduces a novel second-order learning algorithm for multilayer perceptron (MLP) networks, improving convergence and overcoming limitations of existing methods like backpropagation. The revised Newton's method offers enhanced efficiency and accuracy in training neural networks.

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Neural Networks

Background:

  • Multilayer perceptron (MLP) networks are widely used but face training challenges.
  • Standard backpropagation algorithm suffers from slow convergence and local minima issues.
  • Existing Newton's methods for MLPs have implementation and computational drawbacks.

Purpose of the Study:

  • To propose a novel second-order learning algorithm for MLP training.
  • To address the limitations of standard backpropagation and conventional Newton's methods.
  • To enhance convergence rate, accuracy, and robustness in MLP training.

Main Methods:

  • A revised Newton's method is proposed.
  • A forward-backward propagation scheme computes the Hessian matrix (H).

Related Experiment Videos

  • A block Hessian matrix (H(b)) approximates H, with efficient recursive inverse computation. Least squares estimation minimizes local minima.

    • Main Results:

    • The block Hessian (H(b)) effectively approximates the true Hessian (H), preserving key properties.
    • The algorithm demonstrates efficient recursive computation of the inverse Hessian.
    • The proposed method overcomes drawbacks of standard backpropagation and normal Newton's methods.
    • Example problems show the algorithm's efficiency and superior performance.

    Conclusions:

    • The developed second-order learning algorithm offers significant improvements over existing methods for MLP training.
    • It provides faster convergence, better accuracy, and mitigation of local minima problems.
    • The algorithm is computationally efficient and practical for MLP network training.