A. Kartashov et al., QUALITY AND EFFICIENCY OF RETRIEVAL FOR WILLSHAW-LIKE AUTOASSOCIATIVENETWORKS .3. WILLSHAW-POTTS MODEL, Network, 8(1), 1997, pp. 71-86
We study the recognition and correction properties of a structured ass
ociative memory with a floating threshold which can be considered a mo
dification of the Willshaw network and bears some resemblance to a net
work of Potts neurons. An analytical study is performed for the asympt
otic case of large networks and single-step retrieval. The main inform
ational characteristics of the network are obtained; the number of spu
rious stable states and the size of their attraction basins are calcul
ated. A comparison is made with the Willshaw network considered earlie
r in this series of papers.