Ben Goertzel wrote:

> Congratulations!
>
> Next, how will you handle "Maybe"? ;-)
>
> ... ben
>
> On Thu, Jun 23, 2011 at 2:35 AM, A. T. Murray <[email protected]> wrote:
> Hello Ben,
[...]

[Somehow I do not yet have the BenG response in my
e-mail inbox, so I am replying to a switch from JAR.]

Yesterday when I was "clumsily" coding the new 
retroactive KB-adjustment feature for AGI handling 
of terse Yes/No responses from human users, I kept
typing into the AGI "boys play games". Meanwhile, 
I had installed a routine to make the AGI turn
all plural-verb input into an AGI output question
with the subject switched automatically to robots:
"DO ROBOTS PLAY GAMES".

Then each time I entered "no" and pressed [RETURN].

The AGI Mind went back to the assertion of the
question -- whether robots play games -- and
filled a "jux(taposition)" tag with the identifier
for the "NOT" concept. Then I made sure that the 
AGI made a statement of what it now knew:
"ROBOTS DO NOT PLAY GAMES".

Since I was running out of time on 22jun2011
to code the "yes" train of operation, I was
glad to see that entering "yes" as an answer
let the assertion "robots play games" remain
in KB memory by default as a piece of knowledge -- 
although properly such a procedure is actually
in error, because the "yes" answer should be
making a mere assertion change into an 
affirmed fact. Assuming, therefore, that I
did finish the code for the yes-answer, 
an answer of "maybe" would simply get 
recorded in the AGI memory as an adverbial
episode (Robots maybe play games) and not
as either a (yes) affirmation of the idea
or a (no) negation of the idea.

Does the above maybe answer your question?

TIA,

ATM
-- 
http://www.scn.org/~mentifex/AiMind.html


-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/8660244-d750797a
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=8660244&id_secret=8660244-6e7fb59c
Powered by Listbox: http://www.listbox.com

Reply via email to