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

Optimal nonlinear training in the multi-class proximity problem

D Bollé1, G Jongen, G M Shim

  • 1Instituut voor Theoretische Fysica, K.U. Leuven, Belgium. Desire.Bolle@fys.kuleuven.ac.be

International Journal of Neural Systems
|November 1, 1996
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

Spectra of sparse regular graphs with loops.

Physical review. E, Statistical, nonlinear, and soft matter physics·2011
Same author

Localization transition in symmetric random matrices.

Physical review. E, Statistical, nonlinear, and soft matter physics·2011
Same author

Small-world hypergraphs on a bond-disordered Bethe lattice.

Physical review. E, Statistical, nonlinear, and soft matter physics·2008
Same author

Adaptive thresholds for neural networks with synaptic noise.

International journal of neural systems·2007
Same author

Thermodynamics of spin systems on small-world hypergraphs.

Physical review. E, Statistical, nonlinear, and soft matter physics·2007
Same author

Metastable configurations of small-world networks.

Physical review. E, Statistical, nonlinear, and soft matter physics·2006
Same journal

Latent Space Projections and Atlases, a Cautionary Tale in Deep Neuroimaging using Autoencoders.

International journal of neural systems·2026
Same journal

Transformer-Based Anomaly Detection for Neurodegenerative Screening in MRI Images.

International journal of neural systems·2026
Same journal

Discrete Wavelet Convolution for Learnable Time-Frequency Representation with Application to Seizure Prediction.

International journal of neural systems·2026
Same journal

Automatic Seizure Detection using Hierarchical Spectral-Temporal Feature Learning with an Imbalance-Aware Transformer.

International journal of neural systems·2026
Same journal

Pyramid Vision Transformer-Enhanced Conformer Network for Epileptic Seizure Recognition Using MultiChannel EEG Signals.

International journal of neural systems·2026
Same journal

A Time-Frequency Decoupled Contrastive Learning Framework for Electroencephalography-Based Parkinson's Disease Diagnosis.

International journal of neural systems·2026
See all related articles

This study explores nonlinear modulation of the Hebbian learning rule in multi-class problems. Optimal modulation strategies were identified for improved classification accuracy with binary inputs and multi-state outputs.

Area of Science:

  • Computational neuroscience
  • Machine learning theory
  • Statistical physics

Background:

  • The Hebbian learning rule is fundamental in neural network research.
  • Understanding nonlinear modulation is crucial for optimizing learning algorithms.
  • The multi-class proximity problem presents challenges in classification tasks.

Purpose of the Study:

  • To investigate the impact of nonlinear modulation on the Hebbian learning rule.
  • To analyze learning and generalization rates in multi-class proximity problems.
  • To determine optimal modulation functions for improved classification.

Main Methods:

  • Signal-to-noise analysis was employed.
  • Analytic expressions for learning and generalization rates were derived.

Related Experiment Videos

  • The study considered random, Gaussian, and binary teacher classifications.

    • Main Results:

    • Nonlinear modulation effects on the Hebbian rule were quantified.
    • Optimal modulation for binary inputs and Q'-state outputs was found to be a hybrid function.
    • Performance was illustrated for two-class and three-class problems.

    Conclusions:

    • Nonlinear modulation significantly impacts Hebbian learning in multi-class settings.
    • A combined hyperbolic tangent and linear function is optimal for specific proximity problems.
    • The findings offer insights into designing more effective learning rules.