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

Pruning using parameter and neuronal metrics.

P van de Laar1, T Heskes

  • 1Theoretical Foundation, Foundation for Neural Networks, Department of Medical and Bio-Physics, University of Nijmegen, CPK1 213, Geert Grooteplein 21, 6525 EZ Nijmegen, The Netherlands.

Neural Computation
|May 5, 1999
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

Role of conduct problems in the relation between Attention-Deficit Hyperactivity disorder, substance use, and gaming.

European neuropsychopharmacology : the journal of the European College of Neuropsychopharmacology·2018
Same author

Self-organizing maps, vector quantization, and mixture modeling.

IEEE transactions on neural networks·2008
Same author

On 'natural' learning and pruning in multi-layered perceptrons.

Neural computation·2000
Same author

Prediction of bladder outlet obstruction in men with lower urinary tract symptoms using artificial neural networks.

The Journal of urology·1999
Same author

Partial retraining: a new approach to input relevance determination.

International journal of neural systems·1999
Same author

How dependencies between successive examples affect on-line learning.

Neural computation·1996

We introduce a novel optimality measure for neural network architecture selection. This framework extends existing methods like GOBS and offers computational improvements for pruning algorithms.

Area of Science:

  • Machine Learning
  • Artificial Intelligence
  • Computer Science

Background:

  • Neural network architecture selection is crucial for performance.
  • Existing methods like GOBS have limitations in their theoretical range of validity.
  • Quantifying the optimality of architecture changes is an open challenge.

Purpose of the Study:

  • Introduce a new measure of optimality for neural network architecture selection algorithms.
  • Develop pruning algorithms based on novel metrics in parameter and neuron space.
  • Extend the theoretical foundation and explain experimental results of existing algorithms like GOBS.

Main Methods:

  • Define optimality based on the distance between network probability distributions.
  • Derive two pruning algorithms: one using parameter space metric, another using neuron space metric.

Related Experiment Videos

  • Analyze the relationship of new algorithms to existing ones, such as GOBS.

    • Main Results:

    • The proposed framework extends the theoretical validity of GOBS.
    • The new algorithms can explain previously observed experimental outcomes.
    • Computational improvements for the derived pruning algorithms are presented.

    Conclusions:

    • The novel optimality measure provides a robust framework for architecture selection.
    • The derived pruning algorithms offer theoretical and practical advantages.
    • This work enhances understanding and application of neural network optimization techniques.