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

Designing asymmetric Hopfield-type associative memory with higher order hamming stability.

Donq-Liang Lee1, Thomas C Chuang

  • 1Department of Computer Science and Information Engineering, Vanung University, Chung-Li 32056, Taiwan, ROC. dllee@csie.vnu.edu.tw

IEEE Transactions on Neural Networks
|December 14, 2005
PubMed
Summary

This study introduces a novel approach to designing Hopfield-type associative memory (HAM) by expanding attraction basins using Hamming distance. The method enhances recall capability and memory capacity in associative memory systems.

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

Pure exciplex-based white organic light-emitting diodes with imitation daylight emissions.

RSC advances·2022
Same author

Pattern sequence recognition using a time-varying Hopfield network.

IEEE transactions on neural networks·2008
Same author

Improvements of complex-valued Hopfield associative memory by using generalized projection rules.

IEEE transactions on neural networks·2006
Same author

Design of general bidirectional associative memories with improved recall capability.

International journal of neural systems·2004

Area of Science:

  • Artificial Intelligence
  • Machine Learning
  • Computational Neuroscience

Background:

  • Traditional Hopfield-type associative memory (HAM) design often relies on hyperplane optimization or constrained optimization for weight matrix calculation.
  • Existing methods face limitations in effectively expanding the basins of attraction for prototype vectors in the bipolar state space.

Purpose of the Study:

  • To propose a new method for designing asymmetric Hopfield-type associative memory (HAM) that improves recall capability and memory capacity.
  • To introduce the concept of 'higher order Hamming stability' for systematic expansion of attraction basins.
  • To investigate the impact of non-zero diagonal elements in the weight matrix on memory performance.

Main Methods:

  • The study proposes systematically increasing the training set size to achieve desired basin of attraction sizes, termed 'higher order Hamming stability'.

Related Experiment Videos

  • The conventional minimum-overlap algorithm is modified to incorporate this higher order Hamming stability concept.
  • The research explores allowing non-zero diagonal elements (self-connections) in the weight matrix during learning.

    • Main Results:

    • Experimental results demonstrate improved recall capability and a reduced number of spurious memories using the proposed method.
    • Allowing non-zero diagonal elements in the weight matrix effectively enlarges the basin width around prototype vectors.
    • Relaxing the constraint of zero self-connections can increase basin width with only a minor increase in spurious memories.

    Conclusions:

    • The proposed 'higher order Hamming stability' method offers a more effective approach to designing Hopfield-type associative memory.
    • Enlarging attraction basins via Hamming distance and strategic self-connection learning enhances memory recall and capacity.
    • This research provides a valuable contribution to the field of artificial neural networks and associative memory design.