The spiking neural network is a new artificial neural network that expresses and transmits biological information through spiking sequences based on precise time coding. Compared with the traditional frequency-based neural network, spiking neural network has stronger bionics and computing power, which can more accurately simulate human brain and brain activity. This paper introduces the typical unsupervised spiking neural network learning method-STDP (Spike-Timing-Dependent Plasticity) and the classic supervised spiking neural network learning method-Tempotron and ReSuMe (Remote Supervised Method), combines the spatio- temporal transmission characteristics of biological neural information to analyze the characteristics and performance of various existing spiking sequence learning methods. This review contributes to the applicability research of various spiking neural network models.
Paper
Full text
Research on Learning Algorithm of Spiking Neural Network
Semantic Scholar · Computer Science · 2019
Abstract
The spiking neural network is a new artificial neural network that expresses and transmits biological information through spiking sequences based on precise time coding. Compared with the traditional frequency-based neural network, spiking neural network has stronger bionics and computing power, which can more accurately simulate human brain and brain activity. This paper introduces the typical unsupervised spiking neural network learning method-STDP (Spike-Timing-Dependent Plasticity) and the classic supervised spiking neural network learning method-Tempotron and ReSuMe (Remote Supervised Method), combines the spatio- temporal transmission characteristics of biological neural information to analyze the characteristics and performance of various existing spiking sequence learning methods. This review contributes to the applicability research of various spiking neural network models.