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

Probabilistic sequential independent components analysis.

Max Welling1, Richard S Zemel, Geoffrey E Hinton

  • 1Department of Computer Science, University of Toronto, Toronto M5S 3G4 ON, Canada. welling@ics.uci.edu

IEEE Transactions on Neural Networks
|October 6, 2004
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 Spatiotemporal Perspective on Dynamical Computation in Neural Information Processing Systems.

ArXiv·2026
Same author

Unsupervised Representation Learning From Sparse Transformation Analysis.

IEEE transactions on pattern analysis and machine intelligence·2025
Same author

Learned free-energy functionals from pair-correlation matching for dynamical density functional theory.

Physical review. E·2025
Same author

A foundation model for the Earth system.

Nature·2025
Same author

Learning Neural Free-Energy Functionals with Pair-Correlation Matching.

Physical review letters·2025
Same author

Image segmentation with traveling waves in an exactly solvable recurrent neural network.

Proceedings of the National Academy of Sciences of the United States of America·2025
Same journal

Universal perceptron and DNA-like learning algorithm for binary neural networks: LSBF and PBF implementations.

IEEE transactions on neural networks·2013
Same journal

Guest editorial: special section on white box nonlinear prediction models.

IEEE transactions on neural networks·2011
Same journal

Data-based fault-tolerant control of high-speed trains with traction/braking notch nonlinearities and actuator failures.

IEEE transactions on neural networks·2011
Same journal

Guest editorial: special section on data-based control, modeling, and optimization.

IEEE transactions on neural networks·2011
Same journal

Neural network-based multiple robot simultaneous localization and mapping.

IEEE transactions on neural networks·2011
Same journal

Data-driven model-free adaptive control for a class of MIMO nonlinear discrete-time systems.

IEEE transactions on neural networks·2011
See all related articles

This study introduces the under-complete product of experts (UPoE) model for dimensionality reduction. It offers an exact learning algorithm for under-complete independent components, effective for high-dimensional data.

Area of Science:

  • Machine Learning
  • Data Science
  • Signal Processing

Background:

  • Under-complete models are crucial for reducing high-dimensional data, especially temporal image sequences.
  • Existing methods for under-complete independent component analysis (UICA) often rely on approximate learning rules.

Purpose of the Study:

  • To introduce the under-complete product of experts (UPoE) model.
  • To develop tractable and exact maximum-likelihood learning rules for under-complete independent components.
  • To derive an efficient sequential learning algorithm and explore its connections to related methods.

Main Methods:

  • Developed the under-complete product of experts (UPoE) model, where experts model one-dimensional data projections.
  • Derived maximum-likelihood learning rules for exact under-complete independent component learning.

Related Experiment Videos

  • Formulated an efficient sequential learning algorithm related to sequential independent component analysis (ICA).

    • Main Results:

    • The derived learning rules for UPoE are exact and tractable, coinciding with prior approximate UICA rules.
    • An efficient sequential learning algorithm was successfully derived from the UPoE model.
    • The novel algorithms demonstrated efficacy on high-dimensional continuous datasets.

    Conclusions:

    • The under-complete product of experts (UPoE) provides an effective framework for learning under-complete independent components.
    • The developed exact and sequential learning algorithms offer significant advancements for high-dimensional data analysis.
    • This work bridges UPoE with existing techniques like sequential ICA and projection pursuit density estimation.