APPLICATION AMD ANALYSIS OF BSB MODEL WITH DELAY

摘要：

In this paper we discuss the convergence property of a family of Brain-state-in-a-Box (BSB) models with delay. We propose a convergence theorem of the BSB with delay. We have performed a detailed convergence analysis of this network and found convergence theorem under proper assumptions of the weight matrices of this network: ones is symmetric and the other is row diagonal dominant. Meanwhile, theoretical analysis demonstrates that the BSB with delay performs much better than the original one in updating to an equilibrium point based on Hamming distance. In practical application, the delay items are considered as noise-items, which has many advantage. The advantage of the method is the ability to transmit equilibrium points to satisfactory Solution of application, which keep the evolution by the process of neuron selection from random variation.

关键词： Convergence Brain-state-in-a-Boz (BSB) model Delay

作者: JING-HUA GAO SHEN-SHAN QIU XUE-GANG LI

作者单位: School of Science, Dalian Jiaotong University ,Dalian 116028,China MiLeS Lab, Harbin Institute of Technology Shenzhen Graduate School, 518055,China Software College, Changzhou College Information of Technology 213164,China

会议类型: 国际会议

会议名称: 2008 International Conference on Machine Learning and Cybernetics(2008机器学习与控制论国际会议)

会议地点: 昆明

会议语种:英文

页码: 739-743

在线出版日期: 2008-07-12（万方平台首次上网日期，不代表论文的发表时间）