| View previous topic :: View next topic |
| Author |
Message |
Guest
|
 Posted: Mon Oct 06, 2008 9:23 pm Post subject: MindForth Google Code Update 2008 October 4 |
 |
|
The artificial Mind that thinks in English by virtue of spreading
activation (q.v.) has just been released on Google Code at
http://code.google.com/p/mindforth/wiki/ForthMindTextFile
and can simply be copied by drag-and-drop into an editor
(i.e., word-processor like MS Wordpad) and saved in the
C:\Win32Forth directory as Mind.F before running the AI.
Please note that MindForth is written expressly for the
W32FOR42_671.zip version of Win32Forth, easily
downloadable for auto-install by clicking on a link in
http://mind.sourceforge.net/mind4th.html or in
http://mentifex.virtualentity.com/mind4th.html --
both of which are older versions of the AI Mind.
http://mind.sourceforge.net/m4thuser.html and
http://mentifex.virtualentity.com/m4thuser.html
are webpages of the MindForth User Manual.
http://www.scn.org/~mentifex/mindforth.txt
is another source of the MindForth text file.
The current AI Mind code has been run through
a JavaScript program that stripped out all its
verbose slash-comments and reduced the
artificial intelligence in size by half to 87K.
A. T. Murray |
|
| |
|
 Back to top |
Joachim Pimiskern
Guest
|
| Posted: Tue Oct 07, 2008 8:39 am Post subject: Re: MindForth Google Code Update 2008 October 4 |
|
|
mentifex@myuw.net schrieb:
| Quote: | The artificial Mind that thinks in English by virtue of spreading
activation (q.v.) [...]
|
Sounds interesting. What is spreading: single symbols
or propositions?
Regards,
Joachim |
|
| |
|
| Back to top |
Guest
|
| Posted: Tue Oct 07, 2008 8:14 pm Post subject: Re: MindForth Google Code Update 2008 October 4 |
|
|
On Oct 7, 12:08 am, Joachim Pimiskern <JoachimPimisk...@web.de> wrote:
| Quote: | menti...@myuw.net schrieb:
> The artificial Mind that thinks in English by virtue of spreading
> activation (q.v.) [...]
Sounds interesting. What is spreading: single symbols
or propositions?
Regards,
Joachim
|
The activation spreads from symbol to symbol
during the formation of propositions.
Regards,
Arthur |
|
| |
|
| Back to top |
Guest
|
| Posted: Wed Oct 08, 2008 8:31 pm Post subject: Re: MindForth Google Code Update 2008 October 4 |
|
|
On Oct 8, 10:09 am, Joachim Pimiskern <JoachimPimisk...@web.de> wrote:
| Quote: | menti...@myuw.net schrieb:
The activation spreads from symbol to symbol
during the formation of propositions.
Which representation do you use for propositions?
I suppose objects with a name property for concepts/frames
like verbs, and a list of property / value pairs for the
slots, right?
|
No, nothing so formal, and nothing so traditionally AI-esque.
Because MindForth is the software implementation of a
http://mind.sourceforge.net/theory5.html -- theory of
brain-mind function, MindForth starts out fresh with
spreading activation from the analog of neuron-to-neuron
(actually to ganged neuron bundles, redundantly robust).
For the first ten years or so of MindForth (see ACM SIGPLAN
http://doi.acm.org/10.1145/307824.307853 from 1998),
the AI simply let the spikes of activation spread from
noun to verb to object noun in a simple S-V-O format.
Of course, the AI was not debugged enought to think
properly until January of 2008. Therefore, the propositions
were of a factual, subject-verb-object nature. Now
MindForth has broken free from the SVO straightjacket
and thinks by selecting a subject first, and then either
a transitive verb or an intransitive form of "to be". (See
http://mentifex.virtualentity.com/beverb.html and the
updated beVerb module in the Google Code-base.)
| Quote: |
Another question: most symbolic AI systems have their
own language to describe relations between things
and to define rules for deriving new from old facts.
Does your system use an own language?
|
No, the Forth AI system has associative tags for
"relations between things." See section 2 of
http://mentifex.virtualentity.com/fp080925.html
on "Machine Reasoning" for the latest plans
on "deriving new from old facts."
By the way, Prof/Dr/Mr/Herr Pimiskern, I wish
to thank you for having a link to Mentifex AI
on your own website starting many years ago.
Regards,
Arthur |
|
| |
|
| Back to top |
Joachim Pimiskern
Guest
|
| Posted: Wed Oct 08, 2008 10:09 pm Post subject: Re: MindForth Google Code Update 2008 October 4 |
|
|
mentifex@myuw.net schrieb:
| Quote: | The activation spreads from symbol to symbol
during the formation of propositions.
|
Which representation do you use for propositions?
I suppose objects with a name property for concepts/frames
like verbs, and a list of property / value pairs for the
slots, right?
Another question: most symbolic AI systems have their
own language to describe relations between things
and to define rules for deriving new from old facts.
Does your system use an own language?
Regards,
Joachim |
|
| |
|
| Back to top |
bwgqpxxdy
Guest
|
| Posted: Sat Nov 22, 2008 12:07 pm Post subject: UHBEWsFucrn |
|
|
U6hYS1 <a href="http://yapgnweejzwa.com/">yapgnweejzwa</a>,
Linking to external sites is not allowed,
[link=http://ebqhncwtqcqp.com/]ebqhncwtqcqp[/link],
http://tvtyrqkvbnfs.com/ |
|
| |
|
| Back to top |
|
