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

Beyond Hebb: exclusive-OR and biological learning.

K Klemm1, S Bornholdt, H G Schuster

  • 1Institut für Theoretische Physik, Universität Kiel, Leibnizstrasse 15, D-24098 Kiel, Germany.

Physical Review Letters
|October 6, 2000
PubMed
Summary

This study introduces a biologically plausible learning algorithm for neural networks. The novel synaptic averaging mechanism enables learning complex problems like exclusive-OR (XOR) without error backpropagation and enhances noise robustness.

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

[Vascular surgical training concept for military surgeons in Germany].

Der Unfallchirurg·2018
Same author

Secondary interlocking nailing of previously infected nonunion of the tibia.

European journal of orthopaedic surgery & traumatology : orthopedie traumatologie·2013
Same author

[Improve your skills!: evaluation of a 2.5-day basic course in vascular surgery for surgical trainees].

Der Chirurg; Zeitschrift fur alle Gebiete der operativen Medizen·2013
Same author

Emergence and decline of scientific paradigms.

Physical review letters·2011
Same author

Intermittent claudication secondary to a traumatic arteriovenous fistula.

VASA. Zeitschrift fur Gefasskrankheiten·2008
Same author

[Clinical requirements of aortic imaging].

Der Radiologe·2007

Area of Science:

  • Computational neuroscience
  • Machine learning
  • Artificial intelligence

Background:

  • Current neural network learning often relies on error backpropagation, which lacks biological plausibility.
  • Experimental neurobiology suggests synaptic plasticity changes occur on a slower timescale than neural firing dynamics.

Purpose of the Study:

  • To develop and investigate a biologically plausible learning algorithm for multilayer neural networks.
  • To explore the role of synaptic averaging in learning.
  • To demonstrate learning capabilities without error backpropagation.

Main Methods:

  • Studied a learning algorithm for multilayer neural networks incorporating biologically plausible mechanisms.
  • Focused on synaptic averaging as a key mechanism for plasticity induction.

Related Experiment Videos

  • Evaluated the algorithm's performance on the exclusive-OR (XOR) problem and its robustness to noise.

    • Main Results:

    • The proposed synaptic averaging mechanism enables learning the exclusive-OR (XOR) problem.
    • The algorithm successfully learns without requiring error backpropagation.
    • Learning robustness is significantly increased in the presence of noise.

    Conclusions:

    • Synaptic averaging is a viable mechanism for biologically plausible neural network learning.
    • This approach offers an alternative to error backpropagation for certain learning tasks.
    • The algorithm demonstrates enhanced resilience to noisy data, suggesting potential for real-world applications.