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

Equivariant nonstationary source separation.

Seungjin Choi1, Andrzej Cichocki, Shunichi Amari

  • 1Department of Computer Science and Engineering, Pohang University of Science and Technology, Nam-gu, South Korea. seungjin@postech.ac.kr

Neural Networks : the Official Journal of the International Neural Network Society
|April 18, 2002
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

EDSF-Net : An enhanced dynamic spatiotemporal-frequency attention network for robust EEG decoding in motor imagery.

Neural networks : the official journal of the International Neural Network Society·2026
Same author

Breaking the Depth Barrier in Motor Imagery Classification via a Residual Depthwise-Separable Network.

IEEE transactions on cybernetics·2026
Same author

Enhancing Target Recognition Performance in SSVEP-Based Brain-Computer Interfaces via Deep Neural Networks With Pyramid Squeeze Attention.

IEEE transactions on image processing : a publication of the IEEE Signal Processing Society·2026
Same author

BR-SFDA: A Source-Target Bidirectional Refined SFDA for Privacy Preserving EEG-based BCIs.

IEEE journal of biomedical and health informatics·2026
Same author

Enhancing the Capability and Accuracy of Motor Imagery Classification: A Deep Neural Network-Powered Multifaceted Strategy Model.

IEEE transactions on cybernetics·2026
Same author

TBMSCCN: Two-Branch Multi-Scale Convolutional Correlation Network for Steady-State Visual Evoked Potential Classification.

IEEE transactions on bio-medical engineering·2026
Same journal

Dynamic analysis and reliable mechanical optimization application of ring HNN effected with a memristive neuron.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

DAFF-Net: A detection and search method for small-scale low surface brightness galaxies.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Quasi-synchronization for complex networks with hybrid pinning intermittent control.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Physics-encoded convolutional neural operators for parametric PDEs: A convergence-guaranteed framework via pre-computed kernel fields.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Exploiting audio-visual modalities in videos: Object detection via multi-stage bilateral coupling network.

Neural networks : the official journal of the International Neural Network Society·2026
Same journal

Reliability-aware modality completion with cross-modal distillation for federated learning with missing modalities.

Neural networks : the official journal of the International Neural Network Society·2026
See all related articles

This study introduces natural gradient learning algorithms for source separation of nonstationary signals using second-order decorrelation. These methods offer improved stability and steepest descent compared to previous approaches.

Area of Science:

  • Signal Processing
  • Machine Learning
  • Neural Networks

Background:

  • Traditional source separation methods often require higher-order statistics for stationary sources.
  • Nonstationary source separation can be achieved using second-order decorrelation techniques.
  • Prior work by Matsuoka et al. proposed a cost function for this task.

Purpose of the Study:

  • To derive natural gradient learning algorithms for source separation of nonstationary sources.
  • To adapt the Matsuoka et al. cost function for both recurrent and feedforward neural networks.
  • To enhance the stability and convergence properties of source separation algorithms.

Main Methods:

  • Utilized the natural gradient method for algorithm derivation.
  • Applied the method to fully connected recurrent neural networks and feedforward networks.

Related Experiment Videos

  • Analyzed the equivariant property and steepest descent direction of the proposed algorithms.

    • Main Results:

    • Developed novel natural gradient learning algorithms for nonstationary source separation.
    • Demonstrated that the proposed algorithms possess the equivariant property.
    • Showed that the algorithms find a steepest descent direction, unlike prior methods.
    • Confirmed local stability of the algorithms irrespective of source probability distributions.

    Conclusions:

    • The derived natural gradient algorithms provide a stable and efficient approach for nonstationary source separation.
    • These methods offer advantages over existing algorithms by ensuring steepest descent and local stability.
    • The findings contribute to advancing signal processing techniques for complex, time-varying signals.