计算机应用

• 人工智能与仿真 •    下一篇

基于非联合型学习机制的学习神经元

毕松,刁奇,柴小丰,韩存武   

  1. 北方工业大学
  • 收稿日期:2017-03-02 修回日期:2017-04-01 发布日期:2017-04-01 出版日期:2017-05-13
  • 通讯作者: 刁奇

Research on a Novel Neural Model Based on Non-associative Learning Mechanism

  • Received:2017-03-02 Revised:2017-04-01 Online:2017-04-01 Published:2017-05-13
  • Contact: Qi DIAO

摘要: 针对生物神经细胞所具有的非联合型学习机制,设计了具有非联合型学习机制的新型神经元模型——学习神经元。首先,研究了非联合型学习机制中习惯化学习机制和去习惯化学习机制的简化描述;其次,建立了习惯化和去 习惯化学习机制的数学模型;最后, 基于经典的M-P(McCulloch-Pitts)神经元模型,提出了具有习惯化和去习惯化学习能力的新型神经元模型——学习神经元。经仿真实验验证,学习神经元具有典型的习惯化和去习惯化学习能力,为构建新型神经网络提供良好的基础。

Abstract: Biological neurons had non-associative learning mechanisms, and a novel learning neuron with non-associative learning mechanisms was designed. Firstly, the simplified description of the habituation learning mechanism and the dishabituation learning mechanisms are researched in the non-associative learning mechanism; Secondly, the mathematical models of habituation and dishabituation learning mechanisms are established; Finally, based on the classical M-P (McCulloch - Pitts) neuron model, the learning neuron model with the ability of habituation and dishabituation learning was proposed. The simulation results verify that study neurons have typical habituation and dishabituation learning ability, and provides a good foundation for the construction of the new neural network.

中图分类号: