On Mon, Jun 1, 2015 at 11:33 PM, Piaget Modeler <[email protected]>
wrote:

> Wolfram Alpha



I am missing the point here, of course it could be tackled the narrow AI
way, but we are looking for something different, right? Are you trying to
outsource your analogies? Sell an analogy API? I think "in principle" the
analogy works when we can reuse a script, for example "compressing data is
like drying food, with a bit of time and technique you can use the original
while saving space and weight during transport and storage", and it would
take a bit of general intelligence to show all the different ways in which
the analogy does not work, just like so many of the analogies that dominate
our political debates.

As always, it would be easier to derive or solve analogies with some kind
of logical decomposition, it would be a pity to waste the toolkit of
"physical primitives" in TRIZ, or the tentative search for "irreducible
cognitive dimensions" at CYC or yours truly. Which is more or less the
"thought vectors" that recently appeared in some patents. I believe the
main difference between the search for primitives and the new vectors is
that the vectors are more ad hoc, there is neither the assumption nor the
intention to look for irreducible quantities, fundamental symmetries etc,
the ambition is simply to capture as many parameters of a situation or a
concept in a vector and then "reason" with familiar algebraic tools.

The discovery and application of anything that would look like "cognitive
DNA" would be the holy grail of AGI.

AT



-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657
Powered by Listbox: http://www.listbox.com

Reply via email to