05962105

a

05962105 Kryzhanovskiy, Vladimir Binary patterns identification by vector neural network with measure of proximity between neuron states. Honkela, Timo (ed.) et al., Artificial neural networks and machine learning -- ICANN 2011. 21st international conference on artificial neural networks, Espoo, Finland, June 14--17, 2011. Proceedings, Part II. Berlin: Springer (ISBN 978-3-642-21737-1/pbk). Lecture Notes in Computer Science 6792, 119-126 (2011). 2011 Berlin: Springer EN binary patterns identification vector neural networks

doi:10.1007/978-3-642-21738-8_16

Summary: I describe a new vector neural network, in which a priori information about the distribution of noise is easily and naturally embedded. Taking into account the noise distribution allows to essentially increase the system noise immunity. A measure of proximity between neuron states is embedded for the first time. It makes possible to use the prior information. On binary identification problem the one order increase of storage capacity is shown.