A spiking neuron classifier network with a deep architecture inspired by the olfactory system of the honeybee

We decompose the honeybee´s olfactory pathway into local circuits that represent successive processing stages and resemble a deep learning architecture. Using spiking neuronal network models, we infer the specific functional role of these microcircuits in odor discrimination, and measure their contribution to the performance of a spiking implementation of a probabilistic classifier, trained in a supervised manner. The entire network is based on a network of spiking neurons, suited for implementation on neuromorphic hardware.