Binary Factorization by Neural Autoassociators

0405182 - UIVT-O 20030150 RIV CH eng C - Konferenční příspěvek (zahraniční konf.)
Húsek, Dušan - Frolov, A. A. - Muraviev, I. - Řezanková, H. - Snášel, V. - Polyakov, P.Y.
Binary Factorization by Neural Autoassociators.
Artificial Intelligence and Applications. Zürich: ACTA Press, 2003 - (Hamza, M.), s. 649-653. ISBN 0-88986-390-3. ISSN 1482-7913.
[IASTED International Conference /3./. Benalmadena (ES), 08.09.2003-10.09.2003]
Grant CEP: GA MŠMT LN00B096
Klíčová slova: Boolean factorization * recurrent neural networks * single-step approximation
Kód oboru RIV: BD - Teorie informace

In this paper we demonstrate that Hebbian learning in Hopfield-like neural network is a natural procedure for binary factorization. Due to this learning, factors become the attractors of network dynamics. The neurodynamics is analyzed by Single-Step approximation, which is known to be rather accurate for sparsely encoded Hopfield-network. The accuracy of Single-Step approximation is confirmed by computer simulations.
Trvalý link: http://hdl.handle.net/11104/0125380