基于自组织原理的同构神经元簇绑定机制

Homology Neural Cluster Binding Mechanism on the Basis of Self-organization

以同构神经元簇和异构神经元构成的系统为研究对象 ,基于生物学特征构建系统的序参量———谐振频率 ,并采用福克 -普朗克方程对序参量的演化进行描述 ,然后利用系统的序参量对同构神经元簇的最优响应机制进行拓展 ,即对于相似的外界激励 ,同构神经元簇采取绑定机制对周围异构神经元进行支配 ,优化系统的网络体系结构 ,实现对相似外界激励的最优响应 ,增强系统对外界信息的表征能力。

The system composed of homology neural cluster and different neuron is regarded as the object in the paper.On the basis of biologic characteristic,a order parameter is founded and Fuke-Plank equation is used to describe the system order parameter,and then the system order parameter is used to extend the optimized response of homology neuron cluster,namely for resemble external stimulus,the homology neuron cluster predominates the different neurons by binding mechanism,optimizes the net architecture and realizes the optimized response to resemble external stimulus.The result can enhance the ability of representation to the external information.