By Topic

A neural network that uses a Hebbian/backpropagation hybrid learning rule

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$33 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

2 Author(s)
R. J. Wood ; Dept. of Comput. Sci., Worcester Polytech. Inst., MA, USA ; M. A. Gennert

A novel neural network architecture is proposed that provides a unified framework for Hebbian and backpropagation-based learning. The learning rule for this architecture, called the hybrid learning rule, combines the features of the Hebbian learning rule, which is a good feature extractor, and the backpropagation algorithm, which is an excellent classifier. By combining these two learning rules into the hybrid learning rule, the hybrid learning rule should have the strengths of both without any of the weaknesses. The hybrid learning rule was applied to the problem of isolated character recognition. While the hybrid learning rule failed to perform better than the backpropagation algorithm, it did generate receptive fields similar to those found by R. Linsker (1986), T.D. Sanger (1989), and D.H. Hubel and T.N. Wiesel (1962)

Published in:

Neural Networks, 1992. IJCNN., International Joint Conference on  (Volume:3 )

Date of Conference:

7-11 Jun 1992