...
首页> 外文期刊>電子情報通信学会技術研究報告. ニュ-ロコンピュ-ティング. Neurocomputing >Hardware implementation of pulsed neuron with GHA learning rule based on △∑ modulation, Pulsed neuron, Hardware implementation
【24h】

Hardware implementation of pulsed neuron with GHA learning rule based on △∑ modulation, Pulsed neuron, Hardware implementation

机译:基于△Σ调制,脉冲神经元,硬件实现,具有GHA学习规则的脉冲神经元的硬件实现

获取原文
获取原文并翻译 | 示例
   

获取外文期刊封面封底 >>

       

摘要

Pulsed Neuron (PN) can transfer information with only 1-line connection between two neurons. Therefore, PN is effective for hardware implementation from the viewpoint of circuit scale. Then it's important to keep the circuit scale as same as to operate precisely in consideration of parallel computation. Formerly, we proposed a novel Pulsed Neuron model based on ΔΣ modulation (DSM-PN). DSM-PN consists of simple circuit, and operate multi-input summantion and weight multiplications precisely. In this article, we discuss 1 bit bipolar type DSM-PN. In addition, we implemented GHA which is learning rule of PCA to demonstrate usefulness of the proposed method. And simulation results for GHA is comparable to one by floating-point unit.
机译:脉冲神经元(PN)可以在两个神经元之间仅具有1线连接的信息。 因此,从电路规模的角度来看,PN对硬件实现是有效的。 然后,将电路比例与以考虑并行计算精确运行是很重要的。 以前,我们提出了一种基于ΔΣ调制(DSM-PN)的新型脉冲神经元模型。 DSM-PN由简单的电路组成,并精确地操作多输入求和和重量乘法。 在本文中,我们讨论了1位双极型DSM-PN。 此外,我们实施了GHA,它是PCA的学习规则,以证明所提出的方法的有用性。 并且GHA的仿真结果与浮点单元相当。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号