摘要
设计了一种可控容错域的自联想记忆模型。该自联想记忆模型的优点是 :可以根据记忆样本的重要性 ,为记忆样本设计适当的容错域 ,特别 ,可把记忆样本设计为容错域中点 ,让记忆样本容错域半径达到最大。这与人类记忆联想记忆实际情况类似 ,重要事情或多次见过的事情容错域较大 ,只需用少量信息就可以联想起来。而且给出了模型的分块并行实现软件。试验表明该模型有很好的抗噪音性。
A kind of self associative memory model about controllable tolerant fault field is designed. In the model we can design a proper tolerant fault field for every memory sample according to the importance of the memory sample. Especially, (the memory sample be designed at the center of the tolerant fault field in order to maximize the radius of the memory samples’ tolerant fault field.The model is similar with the human’s real associative memory. More important or noticed frequently things has bigger tolerant fault field and can be associated with fewer information. Moreover, the block & parallel software of the model is given. The simulation shows that the model has a excellent anti-noise ability.
出处
《中国医学影像技术》
CSCD
2003年第z1期81-82,共2页
Chinese Journal of Medical Imaging Technology
基金
国家自然科学基金项目 (No .60 0 750 1 2 )
山东省自然科学基金项目 (No.Y2 0 0 2G1 6)
关键词
人工生命
人工脑
联想记忆
神经网络
Artificial life
Artificial brain
Aassociative memory
Nneural network
