Published September 2004 | Version v1
Report Open

Spatial asymmetric retrieval states in symmetric Hebb network with uniform connectivity

  • 1. Abdus Salam International Centre for Theoretical Physics, Trieste (Italy)
  • 2. Depto. de Ingenieria Informatica and Instituto de Ingenieria del Conocimiento, Universidad Autonoma de Madrid, Madrid (Spain)
  • 3. Depto. de Fisica Fundamental, Universidad Nacional de Educacion a Distancia, Madrid (Spain) and Georgi Nadjakov Institute of Solid State Physics, Bulgarian Academy of Sciences, Sofia (Bulgaria)

Description

In this paper we show tat during the retrieval process in a binary Hebb recursive neural network, spatial localized states can be observed when the connectivity of the network is distance-dependent. We point out that the minimal condition that leads to this type of behaviour is the asymmetry between the retrieval and the learning states. (author)

Availability note (English)

Available from INIS in electronic form; Also available at: http://www.ictp.it

Files

36080713.pdf

Files (493.4 kB)

Name Size Download all
md5:2490ac91f15ed6d7cd58606b923f4dac
493.4 kB Preview Download

Additional details

Identifiers

Publishing Information

Imprint Pagination
14 p.
Report number
IC--2004/91

INIS

Country of Publication
International Atomic Energy Agency (IAEA)
Country of Input or Organization
International Atomic Energy Agency (IAEA)
INIS RN
36080713
Subject category
S71: CLASSICAL AND QUANTUM MECHANICS, GENERAL PHYSICS;
Descriptors DEI
ASYMMETRY; ATTRACTORS; EIGENVALUES; EIGENVECTORS; FREE ENERGY; GAUSS FUNCTION; INFORMATION RETRIEVAL; INTEGRAL CALCULUS; LEARNING; NETWORK ANALYSIS; NEURAL NETWORKS; RANDOMNESS; SIMULATION; SMOOTH MANIFOLDS; TOPOLOGY
Descriptors DEC
ENERGY; FUNCTIONS; MATHEMATICAL MANIFOLDS; MATHEMATICS; PHYSICAL PROPERTIES; THERMODYNAMIC PROPERTIES

Optional Information

Contract/Grant/Project number
Grant CICyT; TIC 01-572; DGI.M.CyT.BFM2001-291-CO2-01
Notes
6 refs, 5 figs