Yet Another Roswell Thread

On 28 Jul, 17:38, Fred J. McCall wrote:
Ian Parker wrote:

:
o IQ tests say anything about AI? Not really.
:

Well, one would hope not, since AI really has nothing to do with
'intelligence, human or otherwise.

Francis Crick failed the 11+ but the exception proves the rule. People
often ask what intelligence is. It is the ability to find connections,
sometimes complicated between different things.

I mentioned translation. If you have a situation where you can
translate with a number of different words, which word do you choose?

Flores es a primavera como reloj es a {resorte, mamanthal}

This might well be set in Spain or Latin America.

Intelligence tests are a fair predictor of how students perform. Not
an infallible predictor. There is motivation, there is memory. In
terms of emory of course AI has the Web. It has difficulty in fishing
things out but no difficulty in remembering. We need verbal reasoning
in oder to be able to grasp the essence of a language. As well as US
relations with Latin America I feel that a language does improve
verbal reasoning.

Do you know anything at all about the field of AI or does all your
love for the field stem from ignorant imaginings?

AI is in fact based on being able to have a mathematical model of what
we are talking about and being able to "intelligently" give answers.

In the case of robotics we have dynamical models. A standard package
like ProEngineer will evaluate the dynamical equations of any
assembly. To perform an action, let us say we need to shift something
on the Moon or whatever. We need to have a model of what we are doing,
we then model it and then send the various electrical impulses which
will pick up Moon rock.

There are few gaps in picking up Moon rock, the dynamics is already
fairly well understood. This is why in an earlier posting I suggested
having a competition of doing work on a simulated Moon using (say)
Javalink on ProEngineer.

The cognitive field is probably more interesting from an academic
viewpoint.

http://www.paperoftheweek.com/2007/0...-intelligence/

As you might imagine google is interested in AI. They have got a
number of researchjers working for them uincluding Dr. Riezler from
Tubingden. German will probably bre the first "good" language. With
language you can have really specific fishing out which is why I rate
it as so important.


- Ian Parker

Ian Parker wrote:

:On 28 Jul, 17:38, Fred J. McCall wrote:
: Ian Parker wrote:
:
: :
: o IQ tests say anything about AI? Not really.
: :
:
: Well, one would hope not, since AI really has nothing to do with
: 'intelligence, human or otherwise.
:
:
:Francis Crick failed the 11+ but the exception proves the rule. People
ften ask what intelligence is. It is the ability to find connections,
:sometimes complicated between different things.
:

Ok, you don't know what intelligence is (either).

:
:I mentioned translation. If you have a situation where you can
:translate with a number of different words, which word do you choose?
:
:Flores es a primavera como reloj es a {resorte, mamanthal}
:
:This might well be set in Spain or Latin America.
:

This might well be irrelevant filler - and it is.

:
:Intelligence tests are a fair predictor of how students perform.
:

Not even that.

:
:Not
:an infallible predictor. There is motivation, there is memory. In
:terms of emory of course AI has the Web.
:

Well, no, it doesn't. This is the sort of silly remark that leaves me
thinking you don't have a clue what the field of AI is about and that
all your 'knowledge' stems from having read too many cyberpunk novels.

:
:It has difficulty in fishing
:things out but no difficulty in remembering. We need verbal reasoning
:in oder to be able to grasp the essence of a language. As well as US
:relations with Latin America I feel that a language does improve
:verbal reasoning.
:

Well, sort of by definition, wouldn't you say? Spouting circular
reasoning doesn't help your case.

:
:
: Do you know anything at all about the field of AI or does all your
: love for the field stem from ignorant imaginings?
:
:
:AI is in fact based on being able to have a mathematical model of what
:we are talking about and being able to "intelligently" give answers.
:

AI is in fact based on no such thing (necessarily).

:
:In the case of robotics we have dynamical models. A standard package
:like ProEngineer will evaluate the dynamical equations of any
:assembly.
:

A standard package like ProEngineer has nothing to do with AI.


snip

I think we're done here. You don't have a clue what AI is (either).


--
"Ignorance is preferable to error, and he is less remote from the
truth who believes nothing than he who believes what is wrong."
-- Thomas Jefferson

On 29 Jul, 00:00, Fred J. McCall wrote:
Ian Parker wrote:

:On 28 Jul, 17:38, Fred J. McCall wrote:
: Ian Parker wrote:
:
: :
: o IQ tests say anything about AI? Not really.
: :
:
: Well, one would hope not, since AI really has nothing to do with
: 'intelligence, human or otherwise.
:
:
:Francis Crick failed the 11+ but the exception proves the rule. People
ften ask what intelligence is. It is the ability to find connections,
:sometimes complicated between different things.
:

Ok, you don't know what intelligence is (either).

:
:I mentioned translation. If you have a situation where you can
:translate with a number of different words, which word do you choose?
:
:Flores es a primavera como reloj es a {resorte, mamanthal}
:
:This might well be set in Spain or Latin America.
:

This might well be irrelevant filler - and it is.

:
:Intelligence tests are a fair predictor of how students perform.
:

Not even that.

:
:Not
:an infallible predictor. There is motivation, there is memory. In
:terms of emory of course AI has the Web.
:

Well, no, it doesn't. This is the sort of silly remark that leaves me
thinking you don't have a clue what the field of AI is about and that
all your 'knowledge' stems from having read too many cyberpunk novels.

On the contrary cyberpunk is largely about intelligent boxes. We know
that this is unrealistic. In fact most research is now directed at the
extraction of information from large databases.

Correlations are in fact 0.2 to 0.6. This depends very much on what
you are measuring. The standard intelligence test is indeed split into
2 parts.
Educationalists don't talk about IQ because they (correctly) feel that
knowledge of your IQ will affect motivation.

http://en.wikipedia.org/wiki/IQ
http://www.intelligent-systems.com.a...t/whaisint.htm What is
intelligence? There have in fact been a number of definitions
proposed.

:
:It has difficulty in fishing
:things out but no difficulty in remembering. We need verbal reasoning
:in oder to be able to grasp the essence of a language. As well as US
:relations with Latin America I feel that a language does improve
:verbal reasoning.
:

Well, sort of by definition, wouldn't you say? Spouting circular
reasoning doesn't help your case.

Not really. I have tried to give an insight into genuine problems not
just academically conceived ones. Problem :- You have a Terabyte or so
of Arabic. Most of it is pretty innocuous but you have to make sense
of it and glean some intelligence. Al Qaeda can indeed mean the
toilet. How do you know whether someone isn't just going to the loo.
That is a real problem, real research is needed.
:
:
: Do you know anything at all about the field of AI or does all your
: love for the field stem from ignorant imaginings?
:
:
:AI is in fact based on being able to have a mathematical model of what
:we are talking about and being able to "intelligently" give answers.
:

AI is in fact based on no such thing (necessarily).

Not necessarily. In robotics ioit defintely is. In terms of language
one metric of intelligence is the ability to predict the next word.
This is fact leads us into compression and entropy criteria.
:
:In the case of robotics we have dynamical models. A standard package
:like ProEngineer will evaluate the dynamical equations of any
:assembly.
:

A standard package like ProEngineer has nothing to do with AI.

No ProEngineer contains Javalink where you can write your own
programs. A lot has been done on the question of dynamical
understanding. Most CAD/CAM packages contain first class simulation
software.


- Ian Parker