Abstract
We present evidence that the performance of the traditional fully connected Hopfield model can be dramatically improved by carefully selecting an information-specific connectivity structure, while the synaptic weights of the selected connections are the same as in the Hopfield model. Starting from a completely disconnected network we let “genuine” Hebbian synaptic connections grow, one by one, until a desired degree of stability is achieved. Neural pathways are thus fixed notbefore, butduring the learning phase.
This is a preview of subscription content, log in via an institution to check access.
References
J. J. Hopfield,Proc. Natl. Acad. Sci. USA 79:2554–2558 (1982).
S. A. Janowsky,Phys. Rev. A 39:6600–6604 (1989).
P. Kuhlmann, R. Garcés, and H. Eissfeller,J. Phys. A 25:L593–598 (1992).
K. E. Kürten,J. Phys. France 51:1585–1594 (1990).
K. E. Kürten,J. Phys. 1 France 2:615–624 (1992).
Author information
Authors and Affiliations
Additional information
Communicated by D. Stauffer
Rights and permissions
About this article
Cite this article
Anderle, M., Schweng, H., Kürten, K.E. et al. Pattern-specific neural network design. J Stat Phys 81, 843–849 (1995). https://doi.org/10.1007/BF02179261
Received:
Issue Date:
DOI: https://doi.org/10.1007/BF02179261