MindForth Programming Journal (MFPJ)

Tues.21.SEP.2010 -- (work in progress)

We are now in a strange situation as AI Mind coders. 
We have created an extremely powerful AI Mind at 
but we have been so relentlessly in pursuit of basic 
AI functionality, that many facets of our AI creation 
remain totally unexplored. Our most recent achievement -- 
yesterday -- was KB-exhaustive searches of the AI Mind 
through input queries put to the knowledge base (KB) 
of the emerging artificial person. The MindForth AI can 
now discuss its own existence with human users, who 
may tell the AI about itself and question the AI about 
its own self-knowledge.

The AI Forthmind still exhibits quirky behaviour, but we 
have the opportunity now to track down each instance of 
quirkiness and to fix it on the most fundamental level. 
Simply put, the conceptual activations are out of whack. 
While the AI exhaustively searches its knowledge base (KB) 
for answers to questions, stray activations build up on 
peripheral concepts (not involved in the discussion) until 
suddenly the accumulating activations override the valid 
chain of thought and engender a mental aberration, 
a statement of nonsense. 

Let us try to solve one particular bug that looks serious. 
As we ask the newly KB-exhaustive AI "What are you?" 
and it answers with "I am (this)" and "I am (that)", 
we notice that, at some point, the verb "AM" in the 
responses starts to have zero ("0") as its "seq" tag 
instead of the psi concept number for the noun at the 
end of the "I am..." idea. Such a situation is True-AI 
intolerable, because every thought of the AI needs to 
lay down associative tracks for future retrieval and 
re-assertion of the same idea in its current formulation. 


Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/8660244-d750797a
Modify Your Subscription: 
Powered by Listbox: http://www.listbox.com

Reply via email to