[ExI] Bio/Tech: Embodying cultured Living Neurson via robotics

[Bryan Bishop]

On Wednesday 03 September 2008, nvitamore at austin.rr.com wrote:
> Has anyone read "Removing some 'A' from AI: Embodied Cultured
> Networks"?  

No, but I do remember an SL4 post by a fellow realizing that he was 
already an 'artificial intelligence' and then vanishing or something. I 
can't cite the email any more -- it was long before I rabidly 
bookmarked things of interest, sadly.

> Link:
> http://www.neuro.gatech.edu/groups/potter/papers/DagstuhlAIBakkumprep
>rint.pdf

Yummy. Thank you.

> If neurons are taken out of a brain and continue to create signals
> that instruct the robot to perform tasks; what is driving the neurons
> to give off signals?  Are these signals of any value?  If so, what?

These neurons are operating on biophysics like at any other time. This 
is what 'drives' them to give off signals. The same action potentials 
and electrical discharges accumulate in roughly the same fashion 
outside of a brain as within a brain. 

As for any implicit value of a single neuron's activity or silence, 
that's a little bit of a harder question to talk about. An interesting 
thing to note is that the Wikipedia article about artificial neural 
networks says that these ANNs are for "learing" quote "in an opitmal 
sense". But what exactly is the mathematical background of that search 
for optimality? That sounds intriguing since optimality is something we 
can generally try to prove in computer science. Anyway, from this 
perspective, what's this optimization about in relation to value?

- Bryan
________________________________________
http://heybryan.org/
Engineers: http://heybryan.org/exp.html
irc.freenode.net #hplusroadmap