Neurons
- There are a lot of models of neurons. I'm interested in neural
models not MLPs, and I'm talking about reasonably plausible
neurophysiological models.
- Given that, I'm happy with point models, (e.g. integrate
and fire, LIF, and Boltzmann machines) and compartmental
models (e.g. Hodgkin-Huxley).
- I've worked with fatiguing leaky integrate and fire neurons
with discrete cycles.
- I like spiking models, but this stuff could work
with continous value output neurons.
- Also, I've really only worked with discrete cycle neurons.
- Again, this stuff will work with continuous cycle neurons, but
we need to figure out more details (which will eventually
need to be understood).
