用分维神经网络实现短期记忆
FRACTAL NEURAL NETWORKS FOR SHORT TERM MEMORY
-
摘要: 神经网络中的突触连接矩阵储存有限的信息。当神经网络在线使用时,基于Hebb规则的网络模型将不再适用。本文提出用分维神经网络实现短期记忆,要求旧的模式随时间而遗忘,最近存储的模式可以可靠记忆。实验结果表明,该网络模型适用于在线工作,且优于对比方法。
-
关键词:
- 分维神经网络; 短期记忆; 突触连接矩阵
Abstract: Synaptic connection matrix encodes limited information. It is well known that neural network memory with storage prescriptions based on Hebb s rule will collapse as more patterns are stored. By requiring that old patterns are automatically forgotten and the memory recalls only the most recent ones, a new short-term neural network memory based on Y.Baram s fractal neural network is proposed. Comparison is made with Morris and Wong's method and the experimental results are shown to be rather satisfactory and encouraging. -
Hopfield J J. Neural networks and physical systems with emergent computational abilities, Proc[J].Nat. Acad. Sci.1982, 79:2554-2558[2]Ho C Y, Sasase I, Mori S. On the capacity of the Hopfield associative memory, Proc. of IEEE/INNS IJCNN92, Baltimore: 1993, 196-201.[3]Morris Robert J T, Wong Wing Shing. A short-term neural network memory, SIAM J. Comput., 1988, 17(6): 1103-1118.[4]Baram Y. Associative memory in fractal neural networks, IEEE Trans. on SMC, 1989, SMC-19(5): 1133-1141.[5]Mezard M, Nadal J P, Toulouse G. Solvable models of working memories, J. Physique, 1996, 47(7): 1457-1462.[6]Li Xia, Wong Wing Shing. Fractal neural networks for short term memory, Proc. of 1994 ICNN of the IEEE WCCI, Orlando: 1995, 2454-2459.
计量
- 文章访问数: 1967
- HTML全文浏览量: 108
- PDF下载量: 603
- 被引次数: 0