O. Yu. Sinyavskiy, A. I. Kobrin
Methods of learning with the teacher of spiking neurons are investigated in this work. Defining key features of stochastic spiking neurons generalized spiking neuron model is constricted. Introduction of the generalized spiking neuron model allows to conveniently formalize problems of its learning with teacher. Information theory language is used in lsearning problems description. Specifically, learning of spiking neuron with the teacher problem is solved using information entropy minimization algorithm. Examples of entropy minimization algorithm are provided for concrete model of spiking neuron extended from Spike Response Model. The task of time delay maintenance between input and output spikes and task of detecting of spiking pattern in a noisy stream of impulse signals are considered.