扩展双向联想记忆神经网络及其应用

提出了一种采用δ规则作为学习算法的扩展双向联想记忆神经网络模型，并从理论上证明了其稳定性。该模型克服了现有采用Ｈｅｂｂ规则作为学习算法的联想记忆神经网络对记忆模式有正交性要求和所模式吸引域小的不足。实验结果表明，其联想记忆能力优于目前现有的联想记忆网络。

Extended Bidirectional Associative Menory Neural Network and Its Applications

Anew model,an extended bidirectional associative memory neural network is proposed and thetheoretical proof for this new model's stability is given. This new model utilizes Delta learning rule as thelearning algorithm so that the weaknesses,i.e. small attractive domain and the requirement of theprototypes'orthogonality,of the AM neural networks using the Hebb learning rule are overcome.Experiments are also conducted and their results show that the associative memory performance of this newmodel is better than the existing associative memory models.