Solving a classification task with temporal input encoding by a spiking neural network with memristor-type plasticity in the form of hyperbolic tangent

The article shows that classification problems can be solved using a single-layer spike neural network with time coding and synapses with memristive plasticity in the form of a hyperbolic tangent. It has been shown that memrisitive plasticity can exhibit the effect of remembering a repetitive spike pattern. The model was tested on the Fisher and Wisconsin breast cancer datasets. The obtained accuracy is comparable with the canonical models of plasticity.