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

A methodology to explain neural network classification.

Raphael Féraud1, Fabrice Clérot

  • 1France Télécom R&D, Lannion.

Neural Networks : the Official Journal of the International Neural Network Society
|May 23, 2002
PubMed
Summary

This study introduces a new method to interpret neural network models. It explains classifications by identifying key variables and clustering data, making complex models understandable.

Related Concept Videos

You might also read

Related Articles

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

Sort by
Same journal

Injecting reasoning into vision-language models via weight-decomposed merging.

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

A Non-equilibrium thermodynamic framework for neural networks: A principled correspondence and parameter dynamics.

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

Dual-pathway mask ranking guided selective fine-tuning for backdoor purification.

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

Discriminative transfer feature learning for unsupervised domain adaptation.

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

ARetinex-Net: Low-light image enhancement with adaptive retinex model and global illumination representation.

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

A Multi-Task Learning Framework with Physics Embedded for Signal Reconstruction and State Prediction in Underground Multi-Robot Systems.

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

Area of Science:

  • Computer Science
  • Artificial Intelligence
  • Machine Learning

Background:

  • Neural networks offer high performance in data mining but lack transparency.
  • Understanding the internal structure of neural network models remains a significant challenge.

Purpose of the Study:

  • To develop a methodology for explaining classifications made by multilayer perceptron neural networks.
  • To enhance the interpretability of complex neural network models in data mining.

Main Methods:

  • Introduced 'causal importance' and a saliency measurement for selecting relevant variables.
  • Developed a data clustering approach based on the hidden layer representation of trained neural networks.
  • Combined saliency and causal importance for a cluster-by-cluster interpretation of classifiers.

Main Results:

  • Successfully identified relevant variables for neural network training.
  • Enabled the interpretation of neural network classifiers through a novel methodology.
  • Demonstrated the effectiveness of the approach on three benchmark datasets.

Conclusions:

  • The proposed methodology significantly improves the interpretability of multilayer perceptron models.
  • This approach bridges the gap between high performance and understanding in neural networks.
  • Offers a pathway to more transparent and trustworthy AI in data mining applications.

Related Experiment Videos