From: Anders Sandberg (nv91-asa@nada.kth.se)
Date: Thu Jan 09 1997 - 07:07:55 MST
It seems to me that Eugene and John are talking about different things.
Eugene gave an estimate of the computing requirements of a somewhat
realistic uploading scenario that was essentially a bound on what we need
to simulate neurons well enough to have identical activities. John argues
that memory information is distributed between synapses (a fairly common
view among neuroscientists). But that doesn't change the bound on
uploading requirements!
It might be possible to compress the brain into a more efficient
simulation (replacing 100 synapses with one meta-synapse etc), but my
intuition tells me that is practically impossible or at least NP complete;
it is just a restatement of how information is stored distributed in the
weights of a large neural net. It might also mean that the inevitable
errors in uploading are manageable, since there is a large redundancy, but
we should not rely on this (it just requires a few erroneous neural
circuits to create an epileptic locus, and some neural circuits seem to
be very finely tuned).
Far too often people think the brain works as a computer; it don`t. A
synapse isn't a bit, but a cluster of synapses isn't a bit either!
On Tue, 7 Jan 1997, John K Clark wrote:
> >Ask Joe Strout, he burns any number of MFlops for hours to
> >simulate just _one_ biologically realistic neuron, and far
> >from running in realtime.
>
> The complexity of an individual neuron is irrelevant, the Madison and Schuman
> findings are about redundancy. The point is, it may not take any more computer
> power to simulate many neurons than to simulate one.
This is the cruicial question. My guess is that it is possible to capture
much of the dynamics in easier models, but it is the complex behaviors
that escape the basic model that are relevant! There is a lot of
subtelity in how neurons fire, and how this subtelity is used varies from
locus to locus...
-----------------------------------------------------------------------
Anders Sandberg Towards Ascension!
nv91-asa@nada.kth.se http://www.nada.kth.se/~nv91-asa/main.html
GCS/M/S/O d++ -p+ c++++ !l u+ e++ m++ s+/+ n--- h+/* f+ g+ w++ t+ r+ !y
This archive was generated by hypermail 2.1.5 : Fri Nov 01 2002 - 14:43:59 MST