The storage capacity of the Hopfield model and moderate deviations
This note relates the storage capacity of the Hopfield model of neural networks with possibly correlated patterns to a moderate deviation upper bound for the empirical correlation of the patterns. Examples are, among others, independent patterns with spins that are correlated as in an Ising model or as in a Curie-Weiss model.
Year of publication: |
2005
|
---|---|
Authors: | Löwe, Matthias ; Vermet, Franck |
Published in: |
Statistics & Probability Letters. - Elsevier, ISSN 0167-7152. - Vol. 75.2005, 4, p. 237-248
|
Publisher: |
Elsevier |
Keywords: | Hopfield model Neural networks Storage capacity Moderate deviations Large deviations Spin glasses |
Saved in:
Saved in favorites
Similar items by person
-
The swapping algorithm for the Hopfield model with two patterns
Löwe, Matthias, (2009)
-
The capacity of q-state Potts neural networks with parallel retrieval dynamics
Löwe, Matthias, (2007)
-
Gaussian fluctuations for sample covariance matrices with dependent data
Friesen, Olga, (2013)
- More ...