Search research articles

Search research articles

Related Experiment Videos

Selective information enhancement learning for creating interpretable representations in competitive learning.

Ryotaro Kamimura¹

¹IT Education Center, Tokai University, 1117 Kitakaname, Hiratsuka, Kanagawa 259-1292, Japan. ryo@keyaki.cc.u-tokai.ac.jp

Neural Networks : the Official Journal of the International Neural Network Society

|February 9, 2011

Summary

This summary is machine-generated.

Related Concept Videos

Woodward–Hoffmann Selection Rules and Microscopic Reversibility

Woodward–Hoffmann Selection Rules and Microscopic Reversibility

Electrocyclic reactions, cycloadditions, and sigmatropic rearrangements are concerted pericyclic reactions that proceed via a cyclic transition state. These reactions are stereospecific and regioselective. The stereochemistry of the products depends on the symmetry characteristics of the interacting orbitals and the reaction conditions. Accordingly, pericyclic reactions are classified as either symmetry-allowed or symmetry-forbidden. Woodward and Hoffmann presented the selection criteria for...

You might also read

Related Articles

Articles linked to this work by shared authors, journal, and citation graph.

Sort by

Same author

Collective mutual information maximization to unify passive and positive approaches for improving interpretation and generalization.

Neural networks : the official journal of the International Neural Network Society·2017

Same author

Cooperation-controlled learning for explicit class structure in self-organizing maps.

TheScientificWorldJournal·2014

Same author

Self-enhancement learning: target-creating learning and its application to self-organizing maps.

Biological cybernetics·2011

Same author

Cooperative information maximization with Gaussian activation functions for self-organizing maps.

IEEE transactions on neural networks·2006

Same author

Unifying cost and information in information-theoretic competitive learning.

Neural networks : the official journal of the International Neural Network Society·2005

Same author

Multi-layered greedy network-growing algorithm: extension of greedy network-growing algorithm to multi-layered networks.

International journal of neural systems·2004

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

We developed selective information enhancement learning to interpret AI representations by identifying key variables. This method clarifies class boundaries using fewer, more informative inputs, improving AI model understanding.

Area of Science:

Artificial Intelligence
Machine Learning
Information Theory

Background:

Interpreting complex representations in machine learning models remains a challenge.
Identifying crucial variables for accurate classification is essential for model transparency.

Purpose of the Study:

To propose a novel information-theoretic method, selective information enhancement learning, for explicit interpretation of learned representations.
To enhance the clarity of class boundaries by focusing on a minimal set of important variables.

Main Methods:

Variable selection via information enhancement, measuring mutual information between input patterns and competitive units.
Retraining neural networks using free energy minimization with selected important variables.
Evaluating the method on artificial data, the Senate problem, and a voting attitude problem.

Related Experiment Videos

Main Results:

Achieved clear class boundaries using a reduced number of variables across all tested problems.
Demonstrated that a smaller subset of input variables often contains the majority of information.
Observed that this tendency is more pronounced in larger network architectures.

Conclusions:

Selective information enhancement learning effectively interprets AI representations by highlighting critical variables.
The method enhances model transparency and efficiency by reducing feature dimensionality.
This approach offers a pathway to more understandable and robust machine learning models.