Search research articles

Search research articles

Related Experiment Videos

Simultaneous pattern classification and multidomain association using self-structuring kernel memory networks.

Tetsuya Hoya¹, Yoshikazu Washizawa

¹Department of Mathematics, College of Science and Technology, Nihon University, Tokyo 101-8308, Japan. hoya@brain.riken.jp

IEEE Transactions on Neural Networks

|May 29, 2007

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

Steady-State Visual Evoked Potential Classification Using Complex Valued Convolutional Neural Networks.

Sensors (Basel, Switzerland)·2021

Same author

Neural decoding of code modulated visual evoked potentials by spatio-temporal inverse filtering for brain computer interfaces.

Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference·2017

Same author

Reliability-based automatic repeat request for short code modulation visual evoked potentials in brain computer interfaces.

Annual International Conference of the IEEE Engineering in Medicine and Biology Society. IEEE Engineering in Medicine and Biology Society. Annual International Conference·2016

Same author

Mahalanobis distance on extended Grassmann manifolds for variational pattern analysis.

IEEE transactions on neural networks and learning systems·2014

Same author

A novel EEG-based spelling system using N100 and P300.

Studies in health technology and informatics·2014

Same author

Adaptive subset kernel principal component analysis for time-varying patterns.

IEEE transactions on neural networks and learning systems·2014

ABOUT JoVE

Overview Leadership Blog JoVE Help Center

AUTHORS

Publishing Process Editorial Board Scope & Policies Peer Review FAQ Submit

LIBRARIANS

Testimonials Subscriptions Access Resources Library Advisory Board FAQ

RESEARCH

JoVE Journal Methods Collections JoVE Encyclopedia of Experiments Archive

EDUCATION

JoVE Core JoVE Business JoVE Science Education JoVE Lab Manual Faculty Resource Center Faculty Site

Terms & Conditions of Use

This study introduces a novel self-structuring kernel memory (SSKM) for pattern classification and association. SSKM offers a stable, efficient method for creating compact classifiers with high generalization capabilities.

Area of Science:

Machine Learning
Artificial Intelligence
Computational Neuroscience

Background:

Traditional pattern classification and association methods often involve complex parameter tuning and can suffer from numerical instability.
Hebbian learning principles offer a biologically inspired foundation for neural network construction.
Kernel-based methods are powerful but can be computationally intensive.

Purpose of the Study:

To propose a novel exemplar-based constructive approach using kernels for simultaneous pattern classification and multidomain pattern association.
To develop a self-structuring kernel memory (SSKM) that avoids iterative parameter tuning and numerical instability.
To extend the SSKM for cross-domain pattern association tasks.

Main Methods:

A modular, one-shot self-structuring algorithm based on the Hebbian principle is used to construct kernel networks.

Related Experiment Videos

The constructed networks function as flexible memory units capable of generalization.

The approach is extended to handle multidomain pattern association via cross-domain connections between kernel units.

Main Results:

The SSKM approach constructs compact pattern classifiers with high generalization capabilities.
It avoids arduous and iterative network parameter tuning, mitigating numerical instability.
The SSKM demonstrates effectiveness in both pattern classification and multidomain pattern association tasks.

Conclusions:

The proposed SSKM is an efficient and stable method for simultaneous pattern classification and association.
It offers a viable alternative to conventional approaches like Support Vector Machines (SVMs), particularly in terms of classifier compactness and stability.
The SSKM's ability to handle multidomain associations opens possibilities for more integrated AI systems.