Standards in Artificial Intelligence

Eugene Miya eugene at cse.ucsc.edu
Fri Nov 3 14:24:56 EST 2000


I have been loking for this quote, now I have found it:
	In no part of the field
	have the discoveries produced the major impact promised.
        Sir James Lighthill, 1972


In article <8tp7o1$rtf$1 at nnrp1.deja.com>, Mentifex  <mentifex at scn.org> wrote:
>http://www.scn.org/~mentifex/standard.html -- Standards in AI:

Well, Arthur, an amusing set of pointers.
Followups reduced.  I suggest others do so as well.

Not one pointer to a standards organization, not CBEMA (or whatever it's
new name is), not NIST, not the IEEE.
Quite amusing.

>THEORY (Implement and anticipate a Standard Model of the Mind):
>http://act.psy.cmu.edu -- the ACT-R Unified Theory of Cognition;
>http://www.scn.org/~mentifex/theory5.html -- basis of Mind.Forth;
>http://ai.eecs.umich.edu/cogarch0/soar/ State, Operator & Result.
>
>CONCEPTS:  Bootstrap the mindcore with 64, 128 or 256 concepts.
>HARDWARE:  Aim for a 64-bit implementation (128-bit exception).

I think that we don't know enough to know what difference word size makes.

>SENSORIUM:  Use chaotic attractors in software and manufacturing.
>
>MOTORIUM:  Emulate the cerebellum in software for robot hardware.
>
>MASS-PAR:  Gradually shift from von Neumann to || architectures.

A set/collection/group of von Neumann architectures is still a
von Neumann archiecture, Arthur.

>SOFTWARE:  Start then deviate from procedural or OOP Mind loops.
>
>VARIABLES:  Adopt pre-existing standards and naming conventions.
>
>HCI:  Create familiar features in the Human-Computer Interface.
>
>SOCIETY:  Smooth the pathways to Technological Singularity
>http://www.ugcs.caltech.edu/~phoenix/vinge/vinge-sing.html

I have to email Vernor on this.  I've not started the last book he sent me.

>REALITY:  Co-exist with AI in virtual and non-virtual reality.

"What a concept.."  Robin Williams (the comedian, not the one at Almaden).






More information about the Neur-sci mailing list