FYI: Array "foo{ n }" in Mind.forth Robot AI

Arthur T. Murray uj797 at victoria.tc.ca
Sun Jun 14 07:16:53 EST 1998


In the public-domain artificial-intelligence project for robots
http://victoria.tc.ca/~uj797/aisource.html "Mind.forth," if we
have an array simulating concepts in the human or robot brain
and if we index into the array with time "t" and with whichever
column-element that we are trying to retrieve, it may look like
       t @   1 ( first column)  foo{  n  }recall engram !
where we use time "t" to identify a row containing concept "n" and
its associated flags such as valence "v" and recall vector "r" --
a pointer to where a word is stored in auditory memory, because
the conceptual array "lexicon{" will not actually *contain* words.

Other operations upon concepts in the robot brain-mind may include
       }augment -- to increase the valence "v" of a concept;
       }flush -- to find the concept with the highest valence;
       }damp -- to decrease the valence of one or all concept(s);
       }spread -- to distribute valence to associated concepts;
       }browse -- to let the programmer inspect any concept;
and so forth.

The present FYI post to all interested parties serves not only to
describe the current 14.jun.1998 state of Mind.forth Robot AI but
also to let experts intervene if the Forth coding is going astray.

Because Mind.forth is adopting a "quick and dirty" brute-force
approach to the implementation of verbal concepts in a robot AI,
exemplified by searching for all instances of concept "n" rather
than maintaining its own special array for concept "n", it may not
be possible for early releases of Mind.forth AI to conform to the
http://forth.org/fsl/docs/fsl_arrays.html "Forth Scientific Library
Arrays" document detailing how "foo{ n }" arrays are to be created.

Specifically, the left side of the array format in Mind.forth will
variously need to cycle through a few, or recent, or all points "t"
in search of varying combinations of the quasi-synaptic "flags"
standing in for what real neurons do in a real wetware brain-mind.
Therefore anyone able to shed further light here, please do so.
<PRE>
  /^^^^^^^^^^^\ Syntax Strings Together a Thought /^^^^^^^^^^^\
 /visual memory\           ________   semantic   /  auditory   \
|      /--------|-------\ / syntax \  memory    |episodic memory|
|      |  recog-|nition | \________/------------|-------------\ |
|   ___|___     |       |     |flush-vector     |    _______  | |
|  /image  \    |     __|__  / \  _______       |   /stored \ | |
| / percept \   |    /     \/   \/ Verbs \------|--/ phonemes\| |
| \ engrams /---|---/ Nouns \    \_______/      |  \ of words/  |
|  \_______/    |   \_______/-------------------|---\_______/   |
</PRE>



More information about the Neur-sci mailing list