[comp.arch] In reply to Paul W. on Neural Networks

dwong@zgov03.dec.com (Daniel Wong Su Chun PDE/Test) (08/18/87)

	I think that having discussions on neural networks on this conference
is a great idea and look forward to seeing some.  One suggestion is cross-
posting it to comp.ai.  Perhaps the reason why current artifical intelligence
work is not getting very intelligent is because AI researchers are trying to
emulate human intelligence on a very inhuman machine.  Perhaps the best way
is to go back to the fundamentals of human intelligence ie. the nervous
system, try to copy what is there and then build over the concepts.  I am
sure there are many researchers engaging in this field out there in netland
and would like to hear about developments.  If I may be so bold as to only
state the obvious, if the answer to intelligence is having super-high speed
logic circuits, then we would all have nerves made up of obscure
semiconductors.

DISCLAIMER:	My opinions are strictly my own and do not in any way
		reflect the opinions of my employer, the sane world,
		Spot or rubber ducky.