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

Algebraic geometrical methods for hierarchical learning machines.

S Watanabe1

  • 1Tokyo Institute of Technology, Precision & Intelligence Laboratory, Yokohama, Japan. swatanab@pi.titech.ac.jp

Neural Networks : the Official Journal of the International Neural Network Society
|October 30, 2001
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

Ischemia-induced impairment of 2-deoxyglucose uptake and CA1 field potentials in rat hippocampal slices: protection by 5-HT1A receptor agonists and 5-HT2 receptor antagonists.

European journal of pharmacology·1992
Same author

Effect of substance P on circadian rhythms of firing activity and the 2-deoxyglucose uptake in the rat suprachiasmatic nucleus in vitro.

Brain research·1992
Same author

Cancer mortality trends in Japan 1960-1990: three-dimensional graphical presentation.

Japanese journal of clinical oncology·1992
Same author

Effect of time of day on adaptive response to a 4-week aerobic exercise program.

The Journal of sports medicine and physical fitness·1992
Same author

[A case of hepatic hydrothorax treated by pleuro-venous shunt].

Kyobu geka. The Japanese journal of thoracic surgery·1992
Same author

Type 3 GM1 gangliosidosis: characteristic MRI findings correlated with dystonia.

Acta neurologica Scandinavica·1992

Hierarchical learning machines face challenges with traditional statistical theory due to non-identifiability. This study introduces algebraic geometry methods to analyze their learning curves and generalization error.

Area of Science:

  • Machine Learning
  • Computational Statistics
  • Algebraic Geometry

Background:

  • Hierarchical learning machines (e.g., layered perceptrons) are non-identifiable, violating assumptions of conventional statistical asymptotic theory.
  • This non-identifiability prevents standard application of Bayesian methods and leads to issues with generalization error analysis in neural network learning theory.

Purpose of the Study:

  • To overcome limitations of applying statistical asymptotic theory to hierarchical learning machines.
  • To elucidate the relationship between learning curves and the algebraic geometrical structure of parameter spaces.
  • To develop methods for calculating Bayesian stochastic complexity and generalization error.

Main Methods:

  • Utilizing blowing-up technology from algebraic geometry to develop an algorithm.

Related Experiment Videos

  • Calculating Bayesian stochastic complexity for hierarchical learning machines.
  • Analyzing the algebraic geometrical structure of the parameter space.

    • Main Results:

    • An algorithm for calculating Bayesian stochastic complexity was established using algebraic geometry techniques.
    • It was proven that the Bayesian generalization error of hierarchical learning machines is lower than that of regular statistical models.
    • This holds true even when the true distribution is outside the specified parametric model.

    Conclusions:

    • Algebraic geometry provides a novel framework for understanding the learning dynamics of hierarchical machines.
    • The proposed methods offer a way to accurately assess generalization error, outperforming traditional statistical models.
    • This research bridges machine learning theory and algebraic geometry, offering new insights into model complexity and learning performance.