Quasi Biologically Plausible Category Learning

We, humans and animals, do most of our thinking with neurons.
All those neurons work by spiking, and there are good computational models of actual spiking neurons.
Those neurons are connected by synapses, and those synapses learn by Hebbian learning.
So, if you want to do biologically plausible learning, you need to learn with (unsupervised) Hebbian learning, and compute with spiking neurons.
I, however, have taken advantage of the work on balanced networks, so the topologies that are used are biologically unlikely.
So, the simulations are just Quasi Biologically Plausible.

by Chris Huyck