truck

no profile record

comments

RE: #1, even when we can simulate the entire brain, how will we know what topology the network should have and what the node nonlinearities should be? 100 billion nodes can be connected a few different ways. Parameter estimation of a dynamic nonlinear network of this size is the real limitation.

truck

18 anni fadiscuss

I wonder how they plan to do parameter estimation? Millions of nonlinear dynamic models running, each with at least 3 free parameters in the simplest of neuron models, with an almost infinite number of possible topologies of the networks. That is quite the search space and neuroscience can provide very little prior knowledge.