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

Learning and generalization in radial basis function networks

J A Freeman1, D Saad

  • 1Department of Physics, University of Edinburgh, United Kingdom.

Neural Computation
|September 1, 1995
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

Artificial Intelligence Clustering Approach for Force Mapping Analysis of Polyacrylic Acid (PAA)/Polyethylene Oxide (PEO) Polymer Brushes for Biosensor Applications.

Journal of molecular recognition : JMR·2026
Same author

Identifying individuals at risk for surgical supravalvar aortic stenosis by polygenic risk score with graded phenotyping.

medRxiv : the preprint server for health sciences·2024
Same author

Balance Right in Multiple Sclerosis (BRiMS): a feasibility randomised controlled trial of a falls prevention programme.

Pilot and feasibility studies·2021
Same author

Regulation of Tissue Growth in Crustacean Larvae by Feeding Regime.

The Biological bulletin·2018
Same author

Balance Right in Multiple Sclerosis (BRiMS): a guided self-management programme to reduce falls and improve quality of life, balance and mobility in people with secondary progressive multiple sclerosis: a protocol for a feasibility randomised controlled trial.

Pilot and feasibility studies·2017
Same author

Interleukin (IL) 31 induces in cynomolgus monkeys a rapid and intense itch response that can be inhibited by an IL-31 neutralizing antibody.

Journal of the European Academy of Dermatology and Venereology : JEADV·2016
Same journal

A Model-Free Reinforcement Learning Implementation of Decision Making Under Uncertainty by Sequential Sampling.

Neural computation·2026
Same journal

DROP: Distributional and Regular Optimism and Pessimism for Reinforcement Learning.

Neural computation·2026
Same journal

Hierarchical Active Inference Using Successor Representations.

Neural computation·2026
Same journal

W-Kernel and Its Principal Space for Frequentist Evaluation of Bayesian Estimators.

Neural computation·2026
Same journal

A Hidden Markov Model-Inspired Sequence Classification Method for Hyperdimensional Computing.

Neural computation·2026
Same journal

Sparse Graphical Modeling for Electrophysiological Phase-Based Connectivity Using Circular Statistics.

Neural computation·2026
See all related articles

This study analyzes radial basis function networks during stochastic training, revealing that generalization error is linked to prediction error and free energy. Training optimization and faster learning were observed with joint hidden-layer unit activations.

Area of Science:

  • Machine Learning
  • Artificial Neural Networks
  • Computational Neuroscience

Background:

  • Radial basis function (RBF) networks are a type of artificial neural network.
  • Understanding generalization error is crucial for effective model training and performance.
  • Stochastic training introduces randomness into the learning process.

Purpose of the Study:

  • To analyze the generalization properties of two-layer RBF networks under stochastic training.
  • To derive generic learning curves and explore generalization error definitions.
  • To investigate methods for optimizing training and improving learning speed.

Main Methods:

  • Analysis of a two-layer RBF network with fixed basis function centers.
  • Application of stochastic training paradigms.

Related Experiment Videos

  • Derivation of learning curves and generalization properties using analytical methods.
  • Examination of weight decay effects and optimization via generalization error minimization.

    • Main Results:

    • Generalization error is analytically related to evidence, prediction error, and free energy.
    • Generic learning curves are inversely proportional to the number of training pairs.
    • Joint activations between hidden-layer units were shown to accelerate training.
    • Optimization strategies were explored by minimizing generalization error.

    Conclusions:

    • The study provides analytical insights into the generalization behavior of RBF networks.
    • The findings offer a basis for optimizing training algorithms and improving network performance.
    • Joint hidden-layer unit activations present a viable method for speeding up the training process.