Summon a Robot AI Mind into your presence with MSIE.

The verbPhrase Mind-Module


1. Overview and Brain-Mind Diagram

   /^^^^^^^^^\ Verb-Phrase = Verb +/- Noun-Phrase  /^^^^^^^^^\
  /   EYE     \   Generation of "cats eat fish"   /   EAR     \
 /             \    | | |    ___________         /             \
|   _______     |   | | |   / Sentence  \       |               |
|  /"cats" \    |   | | |   \___________/-------|-------------\ |
| / percept \---|---|-+ |     |        \        |             | |
| \ engram  /   |  e|c|f|     |        _\____   |  auditory   | |
|  \_______/    |  a|a|i|     |       / Verb \  |  memory     | |
|               |  t|t|s|     |      ( Phrase ) |  channel    | |
|   visual      |   |s|h|     |      /\______/  |  where      | |
|               |   | | |     |     /     |     |  ideas      | |
|   memory      |   | | |  ___|__  /      |     |  (re)enter  | |
|               |   | | | / Noun \/       |     |  the        | |
|   channel     |   | | |( Phrase )       |     |  conscious  | |
|               |   | | | \______/        |     |  Mind       | |
|               |   | | |     |           |     |   ________  | |
|   _______     |   | | |     |         __V__   |  /        \ | |
|  /fresh  \    |   |_|_|    _V___     / En  \  | /  "cats"  \| |
| / image   \   |  /     \  / En  \   / Verbs \-|-\  "eat"   /  |
| \ engram  /---|--\ Psi /-/ Nouns \  \_______/ |  \ "fish" /   |
|  \_______/    |   \___/  \_______/------------|---\______/    |

diagrams.html shows a Theory of Mind.

The Verb-Phrase module flushes out the most active En(glish) verb
along with any most active noun to be the direct object of the verb.


2. The Recall of Verbs is the Key to AI

The robot seed AI is based on a uniquely original theory of mind
that results first from posing and then from answering the question:
How does a mind recall a verb to describe a perceived action?

The necessity of linking multiple image percept engrams (see above)
in the visual memory channel with a single verb stored in audition
led to the insight that an intervening abstract memory channel must
exist in the brain as an area that processes links between senses.
Without such an area, only one engram in a sensory memory channel
could be linked by an associative tag to one single engram in any
other sensory memory channel, and always on a one-to-one basis.
With such an area -- whether you call it an abstract memory channel
or a semantic memory -- not only can multiple engrams in vision
interact logically with a single engram in, say, auditory memory, but
entire superstructures of language, volition and emotion may arise in
response to the genetic call for Mind to evolve on planet Earth and
spread through the parsecs of nearby space as Mind over matter.

The diagram ai4u_157.html is a flowchart of Mind.


3. JavaScript artificial intelligence source code of 12 August 2002
// verbPhrase is called from subject-verb-object SVO()
// to find and express a verb +/- object.
function verbPhrase() {  // ATM 6jul2002; or your ID & date.
  Reify();  // Move abstract Psi concepts to enVocab reality.
  act = 0;  // Precaution even though zeroed in Reify().
  aud = 0;  // Zero out the auditory recall-vector "aud".
  en3 = 0;  // No pre-assumptions about grammar category "en3"
  motjuste = 0; // Reset the choice of "best word" for safety.
  opt = 8;  // Look for option eight (a verb).
  psi = 0;
  unk = 0;
  for (i = t; i>midway; i--) {
    enLexicon[i].enExam(); // Inspect English nodes.
      if (en3 == 8) { // Only look at verbs, for predicate.
        if (en1 > 0) { 
          act = en1; // Obtain "act" if positive.
          if (act > unk) {
            motjuste = en0;
            aud = en5; // auditory recall-vector
            unk = act; // to test for higher "act(ivation)"
          } // end of if-clause looking for high activation
        } // End of search for positive "act" values;
      } else continue; // try to avoid looking at non-verbs.
  } // end of loop cycling back through English lexicon.
  verb_act = unk;   // value for display in troubleshoot mode. 
  Speech();         // Call Speech to say or display a word.
  psi = motjuste;   // For use in Activate().
  urpre = pre;      // Preserve "pre" during Activate();
  Activate();       // for sake of En-Psi interaction.
  pre = urpre;      // Retrieve pre after Activate().
  urpsi = motjuste; // For use in psiDamp().
  psiDamp();        // Deactivate any concept after using it.
  enDamp();         // to de-activate English concepts
  motjuste = 0;     // Reset for safety.
  psi = 0;          // Reset for safety.
  pho = " ";
  // Make a SPACE-BAR 32?
  nounPhrase();     // Select a direct object. 
  urpre = pre;      // Preserve "pre" during Activate();
  Activate();       // for sake of residual activation.
  pre = urpre;      // Retrieve pre after Activate().
  urpsi = psi;      // to make sure that psiDamp works.
  psiDamp();        // So as to leave only peripheral activation.
} // End of verbPhrase(); return to subject-verb-object SVO().

4. Mind.Forth free AI source code of 4 August 2002
\  verbPhrase is called from the subject-verb-object SVO
\  syntax module to find and express a verb +/- object.
:  verbPhrase  \ ATM 22jul2002; or your ID & date.
  REIFY       \ move abstract Psi concepts to enVocab reality
  0 act !     \ precaution even though zeroed in REIFY
  0 aud !     \ Start with a zero auditory recall-tag.
  0 motjuste ! 
  8 opt !  \ Look for option eight (a verb).
  0 psi !  \ Start with a zero Psi associative tag.
  midway @ t  @ DO  \ Search backwards through enVocab
    I      3 en{ @  8 = IF  \ only look at predicate/verbs
      I    1 en{ @  act @ > IF  ( if en1 is higher )
        I  0 en{ @  motjuste !  ( store psi-tag of verb )
        I  5 en{ @  aud !  ( auditory recall-vector )
        I  1 en{ @  act !  ( to test for a higher en1 )
        ELSE  \ An error-trap (?) is needed here.
      THEN  \ end of test for en1 highest above zero.
    THEN    \ end of test for opt=8 verbs
  -1 +LOOP  \ end of loop cycling back through English lexicon
   0 psi !            \ A precaution lest psi transit SPEECH.
  SPEECH    \ to say or display the verb
  motjuste @ psi !  \ for use in ACTIVATE
  pre @ urpre !
  ACTIVATE   \ to increase and spread "psi" activation
  urpre @ pre !
  motjuste @ urpsi !  \ for use in psiDamp
\ psiDamp    \ to deactivate Psi concepts
  psiDecay   \ 29jul2002 Less harsh than psiDamp.
  enDamp     \ to de-activate English concepts
  32 EMIT        \ Insert a SPACE.
  nounPhrase     \ To express direct object of verb,
  pre @ urpre !  \ Preserve "pre" during ACTIVATE
  ACTIVATE       \ for sake of residual activation.
  urpre @ pre !  \ Restore "pre" after ACTIVATE.
\ psiDamp        \ So as to leave only peripheral activation.
  psiDecay        \ 29jul2002 Less harsh than psiDamp.
;  \  End of verbPhrase; return to the SVO syntax module.

variable.html explains the location and purpose of each AI variable.


5. Analysis of the Modus Operandi

The code above searches backwards in time for le mot juste to be
the verb in a sentence of thought generated by the open source
software implementation of English syntax in the transformational
grammar theory of Noam Chomsky for linguistics. The criterion
for selection is, which of the verbs in the English en array
is momentarily most active as a result of the neuronal process
of spreading activation, central to our default standard model
of the brain-mind?


6. Troubleshooting and Robotic Psychosurgery

Try in advance not to introduce any evolutionary bugs.

The AI Debugger program may shed some light in general on how to debug
and troubleshoot programs in artificial intelligence.


7. verbPhrase Resources



Return to top; or to
ai4u_faq.html Frequently Asked Questions (FAQ) about AI4U
ai4udex.html Index of the AI4U Textbook of Artificial Intelligence
aisteps.html AI Algorithm Steps - Artificial Intelligence Algorithms For Humanoid Robots
Mind.html Artificial Intelligence Seed AI in JavaScript
mind4th.html Mind.Forth Artificial Intelligence for Robots
PODsales.html Sales of the AI4U Print-On-Demand (POD) Textbook