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

Regularization algorithms for learning that are equivalent to multilayer networks.

T Poggio, F Girosi

    Science (New York, N.Y.)
    |February 23, 1990
    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

    Correction: Excess of NPM-ALK oncogenic signaling promotes cellular apoptosis and drug dependency.

    Oncogene·2025
    Same author

    An analysis of training and generalization errors in shallow and deep networks.

    Neural networks : the official journal of the International Neural Network Society·2019
    Same author

    Response from sinha and poggio.

    Trends in cognitive sciences·2011
    Same author

    Multiclass cancer diagnosis using tumor gene expression signatures.

    Proceedings of the National Academy of Sciences of the United States of America·2001
    Same author

    Learning in brains and machines.

    Spatial vision·2001
    Same author

    Categorical representation of visual stimuli in the primate prefrontal cortex.

    Science (New York, N.Y.)·2001
    Same journal

    Erratum for the Research Article "Detecting supramolecular organic nanoparticles during heat wave".

    Science (New York, N.Y.)·2026
    Same journal

    Local signals, systemic decline.

    Science (New York, N.Y.)·2026
    Same journal

    The mechanics of liver regeneration.

    Science (New York, N.Y.)·2026
    Same journal

    Computing in a memory with physics.

    Science (New York, N.Y.)·2026
    Same journal

    Retraction.

    Science (New York, N.Y.)·2026
    Same journal

    Making time.

    Science (New York, N.Y.)·2026
    See all related articles

    This study demonstrates the equivalence between regularization networks and classical approximation techniques like splines. These regularization networks offer a new perspective on function approximation and pattern recognition.

    Area of Science:

    • Machine Learning
    • Artificial Intelligence
    • Computational Mathematics

    Background:

    • Neural networks learn input-output mappings, akin to approximating multidimensional functions.
    • This learning process is related to hypersurface reconstruction and classical approximation methods.

    Purpose of the Study:

    • To establish a theoretical equivalence between regularization and a specific class of neural networks.
    • To explore the relationship between these networks, generalized splines, and radial basis functions.

    Main Methods:

    • Developing a theoretical framework to link regularization theory with three-layer neural networks.
    • Analyzing the properties of regularization networks and their connection to existing approximation techniques.

    Main Results:

    Related Experiment Videos

    • Demonstrated equivalence between regularization and regularization networks (hyper basis functions).
    • Established that these networks are equivalent to generalized splines and related to radial basis functions.
    • Provided an interpretation of these networks in terms of synthesized and optimally combined prototypes.

    Conclusions:

    • Regularization networks provide a unified framework connecting neural network learning with classical approximation theory.
    • These networks offer novel insights into function approximation, interpolation, and pattern recognition tasks.
    • The prototype-based interpretation enhances understanding of the learning process in these networks.