1
$\begingroup$

Neurons can be simulated using different models that vary in the degree of biophysical realism. When designing an artificial neuronal network, I am interested in the consequences of choosing a degree of neuronal realism.

In terms of computational performance, the FLOPS vary from integrate-and-fire to the Hodgkin–Huxley model (Izhikevich, 2004). However, properties, such as refraction, also vary with the choice of neuron.

  1. When selecting a neuronal model, what are consequences for the ANN other than performance? For example, would there be trade-offs in terms of stability/plasticity?

  2. Izhikevich investigated the performance question in 2004. What are the current benchmarks (other measures, new models)?

  3. How does selecting a neuron have consequences for scalability in terms of hardware for a deep learning network?

  4. When is the McCulloch-Pitts neuron inappropriate?


References

Izhikevich, E. M. (2004). Which model to use for cortical spiking neurons? IEEE Transactions on Neural Networks, 15(5). https://www.izhikevich.org/publications/whichmod.pdf

$\endgroup$
  • $\begingroup$ ANN has nothing to do with neurons for now..Did you get confused by the name? $\endgroup$ – DuttaA Sep 13 '18 at 15:47
  • $\begingroup$ @DuttaA I am new to the field, but from my brief literature review this seems to be a possibility: scholarpedia.org/article/… Perhaps I should have asked more explicitly about the current state of the art of using spiking neurons for ANNs. $\endgroup$ – noumenal Sep 13 '18 at 15:51
  • 1
    $\begingroup$ What you are looking for is neuromorphic computing, mainstream ANN currently do not possess any similarity to neurons...Since you must be knowing it is still impossible to know how neurons communicate by spikes..But some work is being done by neuromorphic​ hardware although they are very far from exact biophysical models...Memristors though have shown some promise in this field by having a memory of past currents $\endgroup$ – DuttaA Sep 13 '18 at 15:55
  • 1
    $\begingroup$ ai.stackexchange.com/questions/5955/…. Might be of help. ai.stackexchange.com/questions/5239/… $\endgroup$ – DuttaA Sep 13 '18 at 15:59

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.