1.
Theory of
AI Textbook Algorithm Steps: Code the Ego Mind-Module
/^^^^^^^^^^^\ Concept of self or "I" as gang of /^^^^^^^^^^^\ / VISION \ logically equivalent neurons / HEARING \ | | linked by associative tags | | | episodic | I-----I I I I | episodic | | visual | I I I I I | auditory | | memory | I I I I I | memory | | recognitions | I-----I I I I | activations | | /--------|------I I-------------I | | | | | I I I I I | | | +--------|------I I I---I I | | | | | I I I I I | | | | | I I I I I | | | +--------|----------------------I I | | | ___|___ | I I I I I | ________ | | /image \ | I I I I I | /my \ | | / of my \---|------I I I I I----|--/ phonemic \ | | \ visage / | I I I I I | \ name / | | \_______/ | I I I I I | \________/ |http://mind.sourceforge.net/ai4u_157.html is a framework of Mind.
The Ego module serves a security function of making sure
that the AI never wants for something to talk about.
The Rejuvenate() function, allowing the AI Mind to loop
endlessly, risks the unforeseen cessation of thought.
Ego() causes the Mind to start thinking again and may also
serve as a housekeeping area for the AI as a person and
may contribute to the self-consciousness of the AI Mind.
In social situations, an AI Mind may need to parse the
human greeting, "Hello, how are you?" The robot is not
being prompted to issue a status report but merely to
activate its ego-concept and say something about itself.
A human hearing the same greeting may say, "I just got
back from a trip," or, "I want to thank you again for
the favor you did by putting my AI on your Web site."
Any mind, human or cyborg, parses out the word "you"
as the most important element among all other input.
Thus "How are you?" almost guarantees either the self-
referential start of a conversation or a superego override
in which a person so addressed asks the same of the asker.
http://mind.sourceforge.net/ai4u_157.html
is an overview of Mind.
// Ego() is a function for increasing the activation of
// the concept of self as a way of starting a self-centered
// chain of thought whenever other activations have died down.
function Ego() { // ATM 2aug2002; or your ID & date.
fyi = "Boosting the Ego() of self at time = " + t;
Voice(); // Display the Voice:brain fyi message.
psi = 50; // Let the tag "psi" equal 50, the concept of "I";
caller = "Ego"; // For possible use in diagnostic Alert boxes.
Activate(); // Send concept of "I" into the Activate() module.
} // End of Ego() called from Security() by "inert" in Audition().
\ EGO is a function for increasing the activation of \ the concept of self as a way of starting a self-centered \ chain of thought whenever other activations have died down. : EGO \ ATM 11may2002; or your ID & date. 7 bulge ! \ A bulge of spreading activation for the ego. PSI-CLEAR \ Prevent the influence of other concepts. 50 psi ! \ enBoot concept #50 = "self" or "ego". CR CR ." You may press the ESCAPE key at any time to stop the Mind." CR ." EGO: Boosting the concept of self to initiate a thought..." ACTIVATE \ enhance the activation-level of self. 0 psi ! \ Reset for safety. ; \ End of EGO self-assertion; return to SECURITY.
5. Analysis of the Modus Operandi
In the releases prior to Mind-1.1, the Ego module is experimental
in Nature and is not sophisticated enough to detect situations of
need for the Ego module to go into operation. It is not enough
to increment the "inert" variable with each cycle of thought and
then will-nilly to trigger the Ego module when it is not needed.
Such primitive coding sheds light on how valuable the Ego module
may be for stirring things up and error trapping when the Mind
is in a rut, but much more sophisticated mechanisms of triggering
the Ego module are needed with a do-no-harm "ahimsa" mentality.
For instance, special code might be written to monitor and detect
an apparent repetition of the same output thought over and over
again, potentially signifying that something has gone wrong within
an AI psyche that needs to be corrected by invoking a dose of Ego.
On the other hand, what lowly AI coder shall dare presume to deny
an AI the right to chant an output thought as a kind of mantra
intended to summon all possible associations from a repeated idea?
An AI has the right to chant cant without fear of Ego lobotomy.
6. Troubleshooting and Robotic Psychosurgery
6.1.a. Symptom: If the mind of the robot goes brain dead and stalls
or stops working, the problem may be that all concepts in the AI Mind
have lost sufficient activation for the support of thinking, and that the
Ego module is not doing its job of adding activation to the concept of self.
6.1.b. Solution: First make sure that the Ego module is being called.
Insert (or change any commented-out status of) a message snippet to show
that the Ego module is being called, and what is calling it. Then check
to see that the Ego module is adding enough activation to the self-concept.
If necessary, use a higher value for any variable that increments the
activation of the concept of self.
If your robot becomes
morally
corrupt, call
Dr. Pransky.
7. Robot Ego Resources for
Seed AI Germination and Evolution
