IUBio

Mind.forth State of the Art: 6 JAN 1999

Mentifex mentifex at scn.org
Wed Jan 6 15:43:37 EST 1999


The most recent coding in Mind.forth Robot AI was done yesterday:

Screen # 36    ram:homunculus
 --------------------------------
 0  \ DEEP-STRUCTURE finds a nascent idea in psi{ }       990105atm
 1  : DEEP-STRUCTURE  0 prevmax !  \ Init previous maximum at zero.
 2   ( down to t = ) 1  t @  DO  ( cycle backwards through time )
 3   \ 5jan1999:  Should we look for most active verb and/or noun?
 4    I  1  psi{ }get dyn !  \  Fetch the dynamism at t = Index
 5    dyn @  prevmax @  > IF I 0 psi{ }get f !  I 3 psi{ }get ukx !
 6   \ Abandon intermediate values and keep final value of "f".
 7    dyn @ prevmax !  f @ motjuste !  0 f !  THEN  -1  +LOOP
 8   CR ." D-S Scr #36 has found the most active psi{ } fiber "
 9  motjuste @ . CR ." leading to " 1 t @  DO I 0 uk{ }get datum !
10  datum @ ukx @ = IF ." uk{ } #" ukx @ . ." flooding ear{ } with "
11   \ The rv from uk{ }get serves immediately as t for ear{ }look!
12   I 5 uk{ }get 0 ear{ }look I 5 uk{ }get 1+ 0 ear{ }look ." ...."
13   \ No iterative ear{ } search is necessary with a recall-vector.
14   THEN  -1  +LOOP  \  Ends the iteration down through uk{ }.
15  ;  \  DEEP-STRUCTURE returns to ...
 --------------------------------

http://www.geocities.com/Athens/Agora/7256/m-forth.html Mind.forth
is now at one of those stages where two opposites try to meet in the
middle.  We are trying to code how a mindcore thinks a "deep thought"
and sends it percolating up into the Chomskyan superstructure of
syntax for expression as a communicable idea, and we are coding how
a message of human input will be received by the AI in such a way as
to influence, to determine, to predispose thoughts in the mindcore:

  /^^^^^^^^^^^\ Mind-grid Arrays{ } in Robot PDAI /^^^^^^^^^^^\
 /visual memory\                   _________     /  auditory   \
|      /--------|---------\       / LANG-UK \   |   memory      |
|      |  recog-|nition   |       \_________/---|-------------\ |
|   ___|___     |         | flush-vector|       |   ________  | |
|  /image  \    |     ____V_        ____V__     |  /        \ | |
| / percept \   |    /psi{ }\------/ uk{ } \----|-/ ear{ }   \| |
| \ engrams /---|---/concepts\----/ lexicon \---|-\ phonemes /  |
|  \_______/    |   \________/    \_________/   |  \________/   |

As we try here and now to set up a "bootstrap sequence" of some
initial words and concepts to be pre-coded into the mindcore psi{ }
we discover that a Digitalker "Master Word List" of 137 words is at

   http://www.jameco.com/cfm/viewdesc.cfm?part_num=16491

Jameco Electronics along with their Digitalker chip set.  For non-
commercial research puposes only, could somebody please either post
or e-mail (mentifex at scn.org) or snow-mail Mentifex POB 98103-1326
such a word list for consideration of matching software to hardware?



More information about the Neur-sci mailing list

Send comments to us at biosci-help [At] net.bio.net