Re: [agi] Probabilistic inference in grounded and ungrounded domains

[j.Maxwell Legg]

My route to AGI is as a personal programmer. That means I am programming 
for myself rather than a user base.

Would you still consider as ungrounded the reading information that 
passes through my mind? Common sense indicates that that textual 
information is grounded to me just because of my choices.

I'm not there yet and I have no persistent stores of my reading database 
however I'm not so much worried about the stuff I've already read in the 
past that is going to be lost to this process because a mind, as a 
grounded state, is a continuum which should be faintly fingerprinted 
once its analogous database kicks in.

The following is my inventory of integrated tools which I'm applying to 
this purpose:

(1) A Shuttle AMD 2600+ with 3 monitors

(2) Including a personally written Visual Basic entertainment system 
with code snippets all over the net.

(3) Including a streaming teleprompter that reads aloud anything I throw 
at it.

(4) Waiting in the wings to be integrated is a Zipf Constant extractor 
and Natural Language Chatterbot and prospective compression parser to 
feed a modified Shannon code driven PCA algorithm.

At first I wanted to use the Zipf constant extractor to add some fine 
tuning to the teleprompter so that it can track the spoken word within a 
window.

Now I think I have enough metadata to hash the total of the first 6 
eigenvalues of each text. This I am going to use as an index to my 
reading database that links the most analogous documents. I want to use 
a local WordNet database to extract significant antonyms that appear 
between these texts in the order of their loadings in my reading 
database. These are then the closest to the words being spoken.and it 
should be a breeze to link to a course of action predetermined by the 
analogy selected.

Jim

-------
To unsubscribe, change your address, or temporarily deactivate your subscription, 
please go to http://v2.listbox.com/member/[EMAIL PROTECTED]