离散指数型BAM网络的记忆容量

ON MEMORY CAPACITY OF THE DISCRETE EXPONENTIAL BIDIRECTIONAL ASSOCIATIVE MEMORY NETWORK

基于Kosko的双向联想记忆模型BAM(bidirectional associative memory)原理,本文提出了一个离散指数型双向联想记忆模型。通过理论分析证实,该模型的记忆容量远远大于BAM的记忆容量。本文给出了指数型BAM记忆容量的下界。

Based on Kosko's BAM (bidirectional associative memory) principle, a discrete exponential bidirectional associative memory network is presented in this paper. The theory analyse has shown that the proposed network has much higher capacity for pattern pair storage than the conventional BAM's. A lower bound of memory capacity is derived.