how many neurons?

Rushi Bhatt rushi at cs.iastate.edu
Fri Nov 28 00:15:53 EST 1997


Zen Faulkes wrote:

> > If you don't mind answering my next question; what can C. elegans do
> > with regard to intelligence? Are those 300 neurons just enough for a
> > random movement, or can the organism do something more, such as simple
> > decisions?
> 
>    C. elegans can learn and remember quite well, thank you. I'd recommend

Hello,

I am pretty new to the topic, so please correct me if I am wrong. Isn't
it true that a reflex action could in principle be accomplished with
groups of relatively very few neurons (in biological systems), and, LTP
has been traced to single synaptic junctions? 

With the above in mind, wouldn't it be an overkill to train a neural
network of 300 neurons to do a single simple task? 

Also, my limited experience with Neural Networks would tell me that for
classification tasks you  don't need so many neurons.

Regards,
 
Rushi Bhatt
_______________________________________________________________________________
Rushi Bhatt (Comp Sci)			209, Campus Ave
226, Atanasoff Hall			Ames IA 50014
Iowa State University			Phone(R): (515)296-2343
Ames, IA 50011				http://www.cs.iastate.edu/~rushi	
_______________________________________________________________________________
*By having this crazy random design it was almost sure to work no matter
how
you built it.
		-Minsky on learning machine



More information about the Neur-sci mailing list