光学学报, 1996, 16 (11): 1591, 网络出版: 2006-12-04
高阶局域互联神经网络的关联存储
High Order Local Interconnected Neural Network for Associative Memory
摘要
给出了高阶局域互联神经网络的数学模型。在计算机上对其关联存储能力进行了模拟计算。结果表明,这种局域互联神经网络的互联权重数大大减少了,同时依然具有良好的关联存储能力;另一方面,如果限定了互联权重矩阵的大小,利用本文给出的高阶局域互联神经网络模型可构造较大的人工神经网络。
Abstract
The mathematical model of high order local interconnected neural network (LINN) is presented. Computer simulations on its associative memory are carried out. The results show that high order LINN has a much small interconnection weight matrix and the same associative memory ability, on the other hand, a large scale neural network can be constructed by using high order LINN if the size of interconnection weight matrix is limited.
参考文献
葛益, 张家军, 贺安之. 高阶局域互联神经网络的关联存储[J]. 光学学报, 1996, 16(11): 1591. 葛益, 张家军, 贺安之. High Order Local Interconnected Neural Network for Associative Memory[J]. Acta Optica Sinica, 1996, 16(11): 1591.