Marek Drogosz, Marcin Jaworski
PRIMECAT:A NEURAL NETWORK MODEL OF CATEGORIAL PRIMING CONSEQUENCES

Abstract. We propose theoretical and simulation models of the alternative category inhibition in categorial priming task. Experiments (Jaworski & Drogosz, 1998) show that relative facilitation of category activated by priming is mostly based on the mechanism of inhibition of alternative category. These results can be easily explained in neural network assuming that there are categorial units in the brain, which inhibit each other. Our neural network simulations prove that even in simpler neural systems – without “gnositc” units – inhibition of alternative categories can be the basic and problaby the only one answer on priming.
Autor: Ewelina Soszyńska
Ostatnia aktualizacja: 13.06.2007, godz. 11:08 - Ewelina Soszyńska