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

Generative models for discovering sparse distributed representations

G E Hinton1, Z Ghahramani

  • 1Department of Computer Science, University of Toronto, Ontario, Canada.

Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences
|August 29, 1997
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

A study on subcutaneous infection model of colibacillosis in broilers: Selected immunological and hematological aspects.

Veterinary and animal science·2025
Same author

Outcome Determinants of Patients with Traumatic Pelvic Fractures: A Cohort Study in a Level I Trauma Center in Southern Iran.

Malaysian orthopaedic journal·2018
Same author

Glove-TalkII--a neural-network interface which maps gestures to parallel formant speech synthesizer controls.

IEEE transactions on neural networks·2008
Same author

Analysis of the synergies underlying complex hand manipulation.

Conference proceedings : ... Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual Conference·2007
Same author

Reducing the dimensionality of data with neural networks.

Science (New York, N.Y.)·2006
Same author

Bilinear dynamical systems.

Philosophical transactions of the Royal Society of London. Series B, Biological sciences·2005
Same journal

The microlandscapes of tree trunks: the effect of lichen and tree-level characteristics on arthropod communities.

Philosophical transactions of the Royal Society of London. Series B, Biological sciences·2026
Same journal

Centimetre-scale landscapes to assess the motion behaviour and cognition of gastropods and bivalves.

Philosophical transactions of the Royal Society of London. Series B, Biological sciences·2026
Same journal

Intertidal microcosms of wave-swept rocky shores: ecological and physiological insights from a uniquely stressful environment.

Philosophical transactions of the Royal Society of London. Series B, Biological sciences·2026
Same journal

Temporal and spatial variation in temperature and oxygen at the microscale: key niche axes for aquatic life.

Philosophical transactions of the Royal Society of London. Series B, Biological sciences·2026
Same journal

Natural microcosms in ecology: fulfilling the promise of model systems?

Philosophical transactions of the Royal Society of London. Series B, Biological sciences·2026
Same journal

Microbe-induced galls and plant defence: metabolite crosstalk in a co-evolutionary battle.

Philosophical transactions of the Royal Society of London. Series B, Biological sciences·2026
See all related articles

This study introduces a novel hierarchical generative model, a nonlinear factor analysis, implemented in a neural network. The model efficiently learns sparse, distributed, and hierarchical representations through Bayesian perceptual inference.

Area of Science:

  • Computational neuroscience
  • Machine learning
  • Artificial intelligence

Background:

  • Factor analysis is a statistical method for explaining variance among observed variables in terms of fewer unobserved variables.
  • Hierarchical generative models offer a framework for understanding complex data structures.
  • Neural networks provide a powerful platform for implementing complex computational models.

Purpose of the Study:

  • To introduce a hierarchical, generative model as a nonlinear generalization of factor analysis.
  • To implement this model within a neural network architecture.
  • To demonstrate its capability for accurate Bayesian perceptual inference and efficient learning.

Main Methods:

  • Development of a hierarchical, generative neural network model.

Related Experiment Videos

  • Incorporation of bottom-up, top-down, and lateral connections for Bayesian perceptual inference.
  • Application of a simple, local learning rule for updating connection strengths.

    • Main Results:

    • The model successfully performs Bayesian perceptual inference.
    • The network learns to extract representations that are sparse, distributed, and hierarchical.
    • The learning rule is simple and requires only locally available information.

    Conclusions:

    • The proposed model is a powerful nonlinear generalization of factor analysis.
    • The neural network implementation enables efficient learning and representation extraction.
    • This approach advances the understanding of hierarchical processing and Bayesian inference in artificial systems.