In this paper we introduce and exploit the real replica approach for a minimal generalization of the Hopfield model by assuming the learned patterns to be distributed according to a standard unit Gaussian. We consider the high storage case, when the number of patterns linearly diverges with the number of neurons. We study the infinite volume behavior of the normalized momenta of the partition function. We find a region in the parameter space where the free energy density in the infinite volume limit self-averages around its annealed approximation, as well as the entropy and the internal energy density. Moreover, we evaluate the corrections to their extensive counterparts with respect to their annealed expressions. The fluctuations of properly introduced overlaps, which act as order parameters, are also discussed.
Skip Nav Destination
,
Article navigation
December 2008
Research Article|
December 24 2008
About the ergodic regime in the analogical Hopfield neural networks: Moments of the partition function
Adriano Barra;
Adriano Barra
a)
1Dipartimento di Fisica,
Sapienza Università di Roma
, Piazzale Aldo Moro 2, 00185 Roma, Italy
2Dipartimento di Matematica,
Università di Bologna
, Piazza di Porta San Donato 5, 40126 Bologna, Italy
Search for other works by this author on:
Francesco Guerra
Francesco Guerra
b)
1Dipartimento di Fisica,
Sapienza Università di Roma
, Piazzale Aldo Moro 2, 00185 Roma, Italy
3
INFN
, Sezione Roma 1, Piazzale Aldo Moro 2, 00185 Roma, Italy
Search for other works by this author on:
Adriano Barra
1,2,a)
Francesco Guerra
1,3,b)
1Dipartimento di Fisica,
Sapienza Università di Roma
, Piazzale Aldo Moro 2, 00185 Roma, Italy
2Dipartimento di Matematica,
Università di Bologna
, Piazza di Porta San Donato 5, 40126 Bologna, Italy
3
INFN
, Sezione Roma 1, Piazzale Aldo Moro 2, 00185 Roma, Italy
a)
Electronic mail: [email protected].
b)
Electronic mail: [email protected].
J. Math. Phys. 49, 125217 (2008)
Article history
Received:
June 25 2008
Accepted:
November 10 2008
Citation
Adriano Barra, Francesco Guerra; About the ergodic regime in the analogical Hopfield neural networks: Moments of the partition function. J. Math. Phys. 1 December 2008; 49 (12): 125217. https://doi.org/10.1063/1.3039083
Download citation file:
Pay-Per-View Access
$40.00
Sign In
You could not be signed in. Please check your credentials and make sure you have an active account and try again.
Citing articles via
Well-posedness and decay structure of a quantum hydrodynamics system with Bohm potential and linear viscosity
Ramón G. Plaza, Delyan Zhelyazov
The BRST quantisation of chiral BMS-like field theories
José M. Figueroa-O’Farrill, Girish S. Vishwa
A sufficient criterion for divisibility of quantum channels
Frederik vom Ende
Related Content
The relativistic Hopfield model with correlated patterns
J. Math. Phys. (December 2020)
Complex dynamics in simple Hopfield neural networks
Chaos (August 2006)
The relativistic Hopfield network: Rigorous results
J. Math. Phys. (March 2019)
Linear kernel Hopfield neural network approach in horn clause programming
AIP Conf. Proc. (June 2018)
Flexibility of S-type random K satisfiability in hopfield neural network
AIP Conf. Proc. (August 2024)