摘要: 在Izhikevich提出的脉冲神经元模型中,引入随机变化的输入电流,使神经元的脉冲发放具有随机性,不同数量的神经元采用连接权值组成网络的脉冲发放。实验结果表明,选择适当的连接权值可以得到环路的持续振荡发放。通过脉冲发放,可以在网络中选择神经环路,完成环路记忆联想过程,并给出研究脉冲神经智能的新思路。
关键词:
脉冲神经元,
神经环路,
振荡发放,
联想记忆
Abstract: This paper brings forward a stochastic variable into a typical spiking neuron model that is proposed by Izhikevich in 2003. Improvement makes neuron firing stochastically. Status of networks’ firing consisting of some these neurons is tested. The experimental results indicate that, through setting appropriate connection weights, it can obtain oscillating firing. The loop can be selected from networks by the status of special oscillating firing status. The experiment accomplishes loop memory associative, then puts forward a new train of thought to study spiking neuron’s intelligence.
Key words:
spiking neuron,
neurons’ loop,
oscillating firing,
associative memory
中图分类号:
陈贤富;姚海东;金燕晖;路 烽. 脉冲神经元环路振荡发放与环路选择[J]. 计算机工程, 2008, 34(6): 219-220.
CHEN Xian-fu; YAO Hai-dong; JIN Yan-hui; LU Feng. Oscillating Firing and Loop Selection of Spiking Neurons’ Loop[J]. Computer Engineering, 2008, 34(6): 219-220.