Self organising maps with a point neuron model
Christian Huyck, Ian G. Mitchell · 2013
This abstract describes simulations using a reasonably biological accurate point neuron model, a fatiguing leaky integrate and fire model. These model neurons use a novel compensatory Hebbian learning rule to categorise data items, a standard machine learning task. The resulting system is a kind of self organising map, which compares favourably with a Kohonen map on one machine learning task.