Analysis of orthogonal memory patterns

It is well known that the storage capacity may be large if all memory patterns are orthogonal to each other. In this paper, a clear description is given about the relation between the dimension N and the maximal number of orthogonal vectors with components +/-1, and also the conception of attractive index is proposed to estimate the basin of attraction. Theoretic analysis and computer simulation show that each memory pattern's basin of attraction contains at least one Hamming ball when the storage capacity is less than 0.33N which is better than usual 0.15 N.

It is well known that the storage capacity may be large if all memory patterns are orthogonal to each other. In this paper, a clear description is given about the relation between the dimension N and the maximal number of orthogonal vectors with components +/-1, and also the conception of attractive index is proposed to estimate the basin of attraction. Theoretic analysis and computer simulation show that each memory pattern's basin of attraction contains at least one Hamming ball when the storage capacity is less than 0.33N which is better than usual 0.15 N.

于2010-10-29批量导入

Made available in DSpace on 2010-10-29T06:37:07Z (GMT). No. of bitstreams: 0 Previous issue date: 1998

Japanese Neural Network Soc.; Asian Pacific Neural Network Assembly.

Acad Sinica, Inst Semicond, Artificial Neural Network Grp, Beijing 100083, Peoples R China

Japanese Neural Network Soc.; Asian Pacific Neural Network Assembly.