Re: hello?


"Curt Welch" <curt@xxxxxxxx> wrote in message
news:20080312224119.747$5W@xxxxxxxxxxxxxxxxx
<snip>

I happen to have a fairly well defined definition of intelligence (the
behavior of a reinforcement learning system), but few agree with me on
that
definition.

Few if any with a sense of what AI/ intelligence is all about would agree
that your defintion is complete, satisfying or well-defined.


What I get frustrated with all the time is when people attempt to argue
about what they think something like intelligence is but yet they can't
even come close to defining it with words that actually have a real
(physical) definition,. They define it using other nonsense words which
just creates huge circular definitions of nonsense. But yet, they have no
clue just how much nonsense they our spitting out.

You are the expert on spitting out nonsense!


If all the words you use when attempting to define something like
intelligence can't each be grounded to some physical events, then the
words, and the definitions, are little more than random bull ***.

You have no clue what a "physical event" really is!



<snip>


"wanting" is another perfect example. We witness seeking behavior and we
say that the agent must "want" the object it is seeking, and we talk as if
this wanting was something located inside the agent doing the seeking.

You must be right - the "wanting" is actually in your keyboard, not in the
agent doing the seeking! BWHAHAHAHAHAHAHA!


People at times will try to program this behavior into an robot by adding
an internal "wanting" state to the code. Nothing could be further from
the
truth about what we actually have to build into the machine to make it
intelligent.

How we define words like "wanting" in physical terms is very important to
solving the mystery of AI - how to build machines that act like humans.

Many words we use in these AI debates border on this
problem of being nonsensical - like the words complexity
and information.

No they are not nonsensical. They serve a purpose. You
can refine or restrict their meaning just as physics
has done with force or momentum and so on but they
still had discriminating powers before that. The words
make descriminations between things and events even
if we haven't yet explained how those things are put
together or why they behave the way they do.

Sure, they are used like that. The word "mad" helps us discriminate
between different human behaviors. As long as people understand that
"mad"
is just a word to label a class of behavior,

Mad can be used to describe *only* a state of mind - not any observable
overt behavior.

<snip>

And yet you used the word "understanding" a few posts back?

Yeah, I'm guilty. If I refuse to use those words, then feel like I can't
connect with the person I'm talking with. This is the problem the
behaviorists have run into.

The behaviorists have run into the rest of the sciences that know that there
are things that behaviorists cannot account for; it is the behaviorits that
have been dissed by knowledgeable people in all the sciecnes as going
overboard by eliminating the explanadums! They are masters of decit and
obfuscation.
..



--
Posted via a free Usenet account from http://www.teranews.com

.