J. Phys. I France
Volume 4, Numéro 2, February 1994
Page(s) 253 - 260
DOI: 10.1051/jp1:1994105
J. Phys. I France 4 (1994) 253-260

Neural networks : iterative unlearning algorithm converging to the projector rule matrix

A. Yu. Plakhov and S. A. Semenov

Institute of Physics and Technology, Prechistenka Str. 13/7, Moscow 119034, Russia

(Received 31 March 1993, received in final form 4 October 1993, accepted 12 October 1993)

The iterative unlearning algorithm for connectivity self-correction is proposed. No presentation of patterns during the iteration process is required. Starting from the Hebbian connectivity, the convergence of the (rescaled) iterated connection matrix to the projector rule one is proven, for arbitrary set of p < N binary patterns.

© Les Editions de Physique 1994