Parameter quantization method of spiking neural network (SNN)

Abstract

The invention relates to the field of neural network technology, and particularly to a parameter quantization method of a spiking neural network (SNN). According to the method of the invention, the original spiking neural network of which training is completed is obtained through offline mapping or online training, parameters of weights, thresholds, leakage constants, set voltage, refractory periods, synaptic delay and the like of the spiking neural network of which training is completed are quantified, and all layers of the neural network can share the same set of quantified parameters or each layer respectively has one set of quantified parameters. The spiking neural network after parameter quantization requires only a small number of parameters for realizing high-precision spiking-neural-network functions. According to the method, parameter storage space of the spiking neural network is effectively saved while high precision is maintained, operation speed is improved, and operationpower consumption is lowered.

Description

technical field [0001] The invention relates to the technical field of neural networks, in particular to a parameter quantification method of a pulse neural network. Background technique [0002] Spiking neural network (SNN for short) is called the third-generation neural network, which is closer to the way the human brain processes information and is the future development direction of neural network technology. SNNs receive information based on pulse sequences. There are many encoding methods that can interpret the pulse sequence as an actual number. Commonly used encoding methods include pulse encoding and frequency encoding. The communication between neurons is also carried out through pulses. When the membrane potential of a neuron is greater than its threshold, it will generate a pulse signal and transmit it to other neurons to increase or decrease its membrane potential. The hardware platform of SNN is called a neuromorphic chip or a brain-like chip, which completely...

AI Technical Summary

Problems solved by technology

There are two main ways to train SNN. One is to map the trained parameters to SNN by training the corresponding artificial neural network (ANN for short) under specific conditions, but it is often necessary to pass a large number of parameters during the mapping process. ; The other is to directly carry out online learning of SNN, which is also accompanied by a large number of parameters

If traditional memory (such as SRAM, DRAM, etc.) is used to store parameters, a huge storage space is required. If new devices such as memristors are used to store parameters, it is difficult to accurately and stably realize many parameters; at the same time, the huge amount of parameters will reduce the calculation speed , increase computing power consumption

There is currently no method that can compress a large number of parameters in SNN

Images