The application relates to the field of
artificial intelligence, in particular to a state updating method of a pulse
neuron model and a pulse neural network, which are used to improve the stability of the pulse emission behavior of the pulse
neuron model. The method comprises the following steps: according to the
potassium state quantity of the pulse
neuron model at the last moment, performing shunt adjustment on the input
signal at the current moment to obtain a shunt
signal, then updating the reset
membrane potential at the last moment to obtain the initial
membrane potential at the current moment; determining the emission pulse at the current moment according to the comparison result between the initial
membrane potential at the current moment and a preset emission threshold; determining the
potassium state quantity at the current moment according to the
potassium state quantity at the last moment, the initial membrane potential at the current moment and the emission pulse at the current moment; and performing reset adjustment on the initial membrane potential at the current moment according to the emission pulse at the current moment, the potassium state quantity at the current moment and the emission threshold to obtain the reset membrane potential of the pulse neuron model at the current moment.