All-digital bionic circuit and system applied to neurons

Abstract

The embodiment of the invention discloses an all-digital bionic circuit and system applied to neurons. The circuit comprises a neuron input module, a clock selection module, a counting module, a zerosetting delay module and a neuron output module, wherein the counting module is provided with a pulse signal input end, a clock signal input end, a zero clearing end and a counting signal output end;the neuron input module is connected with the pulse signal input end; the clock selection module is connected with the clock signal input end; the zero setting delay module is connected with the zeroclearing end; and the neuron output module is connected with the counting signal output end. Based on the embodiment of the invention, upward or downward counting of neuron signals can be dynamicallycarried out in a neural network, zero setting signals and counting signals can be broadened through the zero setting delay module, and the refractory period of the neurons is simulated, so that data in the counting module is cleared. The bionic circuit adopts an all-digital design, so that the complexity of the circuit can be simplified, the power consumption of the circuit can be reduced, and large-scale circuit integration can be conveniently realized.

Description

technical field [0001] The invention relates to the field of artificial intelligence, in particular to an all-digital bionic circuit and system applied to neurons. Background technique [0002] Since the 1980s, the spiking neuron network has become a new hotspot in artificial intelligence research due to its powerful computing power and characteristics close to biological neurons. [0003] At present, the modeling models of neuron circuits mainly include: Hodgkin-Huxley circuit, Leaky Integrate-and-Fire circuit and Izhikevich circuit. like figure 1 Shown is a simplified schematic diagram of the Hodgkin-Huxley circuit. Although the Hodgkin-Huxley modeling circuit has good dynamic characteristics and is suitable for a variety of pulse modes, it can be seen that the Hodgkin-Huxley modeling circuit includes a large number of field effect transistors, a large number of The use of field effect transistors directly increases the complexity of the circuit, which seriously restrict...

AI Technical Summary

Problems solved by technology

But it is only a mathematical fitting, and lacks the information inside the neuron, making it difficult to realize the neural network coding, dynamic characteristics and hardware.

Images