Re: AI Mind (Was: Sound capture)

From: Arthur T. Murray (uj797_at_victoria.tc.ca)
Date: 07/31/04


Date: 30 Jul 2004 16:03:45 -0800


"Chris S." writes on Fri, 30 Jul 2004:

> Arthur T. Murray wrote:
>
>> "Chris S." writes on Fri, 30 Jul 2004:
>>
>>> If you have "solved" AI (a claim I heavily discount),
>>
>> http://www.scn.org/~mentifex/theory5.html -- solved in *theory*.
>
> After reading over your javascript code, your
> implementation appears to be a mix between a
> knowledge base and a natural language parser.
ATM:
Yes, a *conceptualized* knowledge base (KB).

> Unfortunately, it doesn't seem to have any general
> IO capabilities. This is a common pitfall to many amateur
> AI researchers. Language alone doesn't define knowledge.

Ture. Language is nevertheless closely involved with thought.

> Without a physical perception of the world, a computer
> will never "know" what 'soft', 'pain', or 'up' mean.
> I've often thought vision algorithms, speech synthesizers,
> NLPs, and such without a general form of processing
> and representing sensory data was like trying
> to run before we could walk.

The sensorium is where we amateur mind-makers count on
help from the amateur robot-makers. They want a mind
for their robots; we want robots for our AI to live in.
>
> But if I were to recommend an improvement in your system,
> I would suggest you try coding it in Python (www.python.org).

http://mind.sourceforge.net/python.html is a Python AI weblog.

> Compared to Javascript, and especially Forth, Python
> is remarkably easy to learn. Python would allow you
> to represent your idea in an object-oriented framework,
> making it much easier for others to understand.

Understanding the AI Mind is helped along considerably by
http://www.scn.org/~mentifex/jsaimind.html -- for MS IE 5:
 
                  ____________ ____________
AI4U textbook / \ / \
User's Manual ( Motorium ) ( Security )
Add your link \____________/ \____________/
        ____________ ____________
       / \ / Sensorium \
      ( Volition ) Listening... ( --> Audition )
       \____________/ \____________/
                   ____________ ____________
                  / \ / \
                 ( Think ) ( Emotion )
                  \____________/ \____________/

> It would also make persisting dynamic data much easier,
> so you wouldn't have to hard code so much of your lexicon.

Only an initial, hit-the-ground-thinking lexicon is hard-coded.
The AI learns new words via the human-computer interaction (HCI).

> Plus you should never run into any platform issues, since
> Python runs everywhere. Python's quickly becoming a serious
> tool for AI research. For instance, "AI: A Modern Approach"
> is currently offering both Lisp and Python source code for
> many of its examples.

With heartfelt thanks to you for the analysis and advice,

Arthur T. Murray



Relevant Pages

  • Re: Does Python really follow its philosophy of "Readability counts"?
    ... with enforcing that "shouldn't" in the language itself? ... In Python, direct access to pointers is a MUST NOT. ... where you are allowed to mess with the implementation. ... human assembly language programmers? ...
    (comp.lang.python)
  • Re: Basic inheritance question
    ... used 'this' in C++ and Java. ... but in Python it doesn't. ... language, they would write a lot of ten liners that is changed a LOT ... Add three levels of inheritence and a couple globals and you'll find out ...
    (comp.lang.python)
  • Re: Python for Fortran programmers
    ... proposition for Python in the Fortran community", ... compiled language such as Fortran or C++, but I would still prefer to ... greater number of programming errors that are typically detected by a ...
    (comp.lang.fortran)
  • Re: Pythons simplicity philosophy
    ... the correct Python idiom, ''.join. ... However, if you agree with Paul Graham's theories on language design, ... 'sum' itself lost substantial generality between my original ...
    (comp.lang.python)
  • Re: Is there a "Large Scale Python Software Design" ?
    ... around 20 developers, assuming colocation... ... situation with Python, yet, only with Fortran, C, C++. ... It sure beats "retrofitting" unit tests post facto. ... Jones' estimates for Java's language level; ...
    (comp.lang.python)