SO I had to beef up the parser to be able to ignore any NON NVP, no
pre and no seq to them but yet connect the SVO to eachother, pre and
seq: here is the result, you will see how the SVO are connected
despite the non NVP located between them.
Psi mindcore concepts
time: psi act jux pre pos seq enx
1727: 74 0 0 0 5 60 74 to -> MAN
1728: 60 0 74 74 8 199 60 to -> HAS
1729: 137 0 60 0 12 0 137 to -> 3
1730: 199 10 137 60 5 0 199 to ->
CHILDREN
1731: 199 10 199 0 5 97 199 to ->
CHILDREN
1732: 97 4 199 199 8 227 97 to -> TALK
1733: 95 8 97 0 6 0 95 to -> TO
1734: 227 3 95 97 5 0 227 to ->
MOTHER
1735: 227 0 227 0 5 97 227 to ->
MOTHER
1736: 97 0 227 227 8 199 97 to -> TALK
1737: 199 10 97 97 5 0 199 to ->
CHILDREN
1738: 196 12 199 0 5 198 196 to ->
FATHER
1739: 198 14 196 196 8 112 198 to -> LOVE
1740: 112 13 198 198 7 199 112 to -> HIS
1741: 199 10 112 112 5 0 199 to ->
CHILDREN
1742: 199 0 199 0 5 198 199 to ->
CHILDREN
1743: 198 14 199 199 8 196 198 to -> LOVE
1744: 196 0 198 198 5 0 196 to ->
FATHER
1745: 112 2 196 0 7 198 112 to -> HIS
1746: 198 2 112 112 8 198 198 to -> LOVE
1747: 54 2 198 0 2 0 54 to -> WHAT
ok
As I work thru my errors in coding, I noticed what is either a
change or an error in the REIFY of yours. You state load the highest
activation level found for a concept, and your previous code did that
but that is not what I see happening in the revision. I have a copy
of the code from 27 Apr. No where is there a comparison of the
activation in holding to the one found. AM I missing something?
Frank
Frank,
After you posted the above on about 28.APR.2007, I did
not see it until today 9.MAY.2007. So I will respond now
belatedly.
http://mind.sourceforge.net/mind4th.html includes the
http://mind.sourceforge.net/reify.html "Reify" module.
As general background not only for yourself but also
for any interested party, I would like to state that the
Reify module exists to transfer conceptual activation
from the deep-structure "Psi" array of concepts up
to the English lexical "En" array, so that the
Chomskyan linguistic superstructure (e.g., the
http://mind.sourceforge.net/svo.html SVO module)
may generate a thought in English or in any
other language available to the AI Mind.
Until 4 April 2007 the Reify module had a serious
bug in it which was fixed and documented with
"4apr07A.F uses a nested BEGIN (flag) UNTIL
loop in the Reify module" in the changelog at
the start of the Mind.Forth free AI source code.
The bug was causing only the most recent
activation-level for a psi concept to be passed
from the Psi array up to the En(glish) array.
As of 4 April 2007 I thought that the bug
was fixed. If you find otherwise, please
let me know.
Back in 1998 I was working heavily on Mind.Forth
and posting heavily here in comp.lang.forth about
the AI. Then I recoded Mind.Forth in JavaScript as
http://mind.sourceforge.net/Mind.html for MSIE.
Debugging Mind.Forth in 2005 and 2006, I
found bugs that I had never even suspected.
Now, having debugged Mind.Forth and the
JSAI to the point that they think albeit crudely,
I need to switch back and forth, coding both.
My goal is to build up not only the sideways
http://mentifex.virtualentity.com/actrules.html
spread of activation horizontally among the
elements of an incipient thought, but also
vertically from the top consciousness tier
down through the subconscious tier and below.
At that point in the Mentifex AI development,
it should be a lot easier for neophytes to see
what is going on at the core of the AI Mind,
and we see emrging AI Forthminds beyond
http://AIMind-I.com by you and Mind.Forth by
very truly yours.