Binary Factorization in Hopfield-Like Neural Networks: Single-Step Approximation and Computer Simulations
Result description
The unsupervised learning of feature extraction in high-dimensional patterns space is a central problem for the neural network approach. Feature extraction is a procedure which maps original patterns into the feature (or factor) space of reduced dimension. In this paper we demonstrate that Hebbian learning in Hopfield-like neural network is a natural procedure for unsupervised learning of feature extraction. Due to this learning, factors become the attractors of network dynamics, hence they can be revealed by the random search. The neurodynamics is analysed by Single-Step approximation which is known citeFROHUM97 to be rather accurate for sparsely encoded Hopfield-network. Thus, the analysis is restricted by the case of sparsely encoded factors. The accuracy of Single-Step approximation is confirmed by computer simulations.
Keywords
nonlinear binary factor analysisfeature extractionrecurrent neural networkSingle-Step approximationneurodynamics simulationattraction basinsHebbian learningunsupervised learningneurosciencebrain function modeling
The result's identifiers
Result code in IS VaVaI
Result on the web
—
DOI - Digital Object Identifier
—
Alternative languages
Result language
angličtina
Original language name
Binary Factorization in Hopfield-Like Neural Networks: Single-Step Approximation and Computer Simulations
Original language description
The unsupervised learning of feature extraction in high-dimensional patterns space is a central problem for the neural network approach. Feature extraction is a procedure which maps original patterns into the feature (or factor) space of reduced dimension. In this paper we demonstrate that Hebbian learning in Hopfield-like neural network is a natural procedure for unsupervised learning of feature extraction. Due to this learning, factors become the attractors of network dynamics, hence they can be revealed by the random search. The neurodynamics is analysed by Single-Step approximation which is known citeFROHUM97 to be rather accurate for sparsely encoded Hopfield-network. Thus, the analysis is restricted by the case of sparsely encoded factors. The accuracy of Single-Step approximation is confirmed by computer simulations.
Czech name
Binární faktorová analýza pomocí Hopfieldovy neuronové sítě: Jednokroková aproximace a počítačová simulace
Czech description
Učení extrakce příznaků bez učitele, ve vysocedimenziálních vzorech, je centrálním problémem Neuronového přístupu. Extrakce příznaků je procedura, která zobrazuje originální vzory do prostoru příznaků (faktorů) menší dimenze. V této práci dokazujeme, žeHebbovo učení v síti Hopfieldova typu je přirozenou procedurou pro učení extrakce příznaků bez učitele. Makroskopické chování této sítě je modelováno pomocí aproximace prvního kroku (single step aproximation).
Classification
Type
Jx - Unclassified - Peer-reviewed scientific article (Jimp, Jsc and Jost)
CEP classification
BA - General mathematics
OECD FORD branch
—
Result continuities
Project
GA201/01/1192: Research of neural networks capability to provide nonlinear Boolean factor analysis
Continuities
Z - Vyzkumny zamer (s odkazem do CEZ)
Others
Publication year
2004
Confidentiality
S - Úplné a pravdivé údaje o projektu nepodléhají ochraně podle zvláštních právních předpisů
Data specific for result type
Name of the periodical
Neural Network World
ISSN
1210-0552
e-ISSN
—
Volume of the periodical
14
Issue of the periodical within the volume
2
Country of publishing house
CZ - CZECH REPUBLIC
Number of pages
14
Pages from-to
139-152
UT code for WoS article
—
EID of the result in the Scopus database
—
Result type
Jx - Unclassified - Peer-reviewed scientific article (Jimp, Jsc and Jost)
CEP
BA - General mathematics
Year of implementation
2004