Distributed AI Coming to a Computer Near You
chuck....@gmail.com
distributed virtual brain one computer at a time. Utilizing a
computational model we’ve seen in such projects as SETI@Home, the
system will harness the computing power of thousands of machines
throughout the world to achieve its goal.
http://www.syntheticthought.com/st/artificial-intelligence/61/158
Don Stockbauer
wrote:
What if this could be extended to every machine and person on Earth?
Then it would be the......the....Global.....
AAAAAHHHHHHHHHHHHHHHGGGGGGAAAGGGG!!!!!!!!!!!!!!
Ian Parker
wrote:
The assumption here seems to be that AI is a question of raw computer
power. This has never been the case. AI is a matter of formalizing
knowledge and organising knowledge. There are only two valid AI
BOINCs. Google which will allow you search out any website according
to its content and Wolfram which deals with facts.
Translation is I believe an acid test of AI which is why it is so
important. In my Arabic - English examples, Wolfram gives you the
properties of several different types of concrete. It will also give
you the weather at Aswan. The $64 billion dollar question. Could this
knowledge be used to feed back into a translation. What do I mean by
"understanding"? Understanding must surely be taking the weather
results at Aswan and correlating it with concrete.
- Ian Parker
Don Stockbauer
Or correlating my windmill with the Zargonian slime creatures of Beta
Lyrae 7.
J.A. Legris
ISTM that even though aggregate understanding certainly involves
aspects of the concrete, especially among the "hard" issues, the real
question is whether the results at Aswan correlate with the abstract.
--
Joe
Ian Parker
translated as "central air battle". Correct was "the battle against
climate and environment". At Aswan the temperature is frequently 50 or
so. Concrete (I mean the mixture of heated lime and clay mixed with
sand. To build a dam you pour massive quantities.
Another example http://sites.google.com/site/aitranslationproject/deepknowled
مضروبا في اربعة اضعاف درجة حرارة سطحه اى (mDrwbA fY ArbEp ADEAf drjp
HrArp sTHh Ay) = Multiplied by four times the temperature of any
surface (Google). drjp - degree is totally ignored. Determined by
multiplying by the temperature degree four for any surface (correct).
Four times the temperature is NOT the Stephan Bolzmann law.
Wolfram I do know has this little piece of "deep" knowledge. In fact I
could feed "Stefan Boltzmann" into Wolfram and I could check the
translation off against what its definition was. I set 200C. It
converts it into K, the law is after all in K and I get 473 degrees
perfectly correctly. Area*constant*273^4 NOT 473*4 which is total
nonsense.
Pouring concrete and fighting Israel which is my first example is a
little difficult to cover with deep knowledge I will admit. The Black
Body laws(example 2) could be done with present day technology.
Indeed calling W from within Mathematica will do just that.
- Ian Parker
Ian Parker
Buckwalter strict transliteration.
- Ian Parker
Neil W Rickert
>The assumption here seems to be that AI is a question of raw computer
>power. This has never been the case.
Agree.
> AI is a matter of formalizing
>knowledge and organising knowledge.
And that's never going to work either.
Ian Parker
What we really need to get our hands round is the following. To be
intelligent you have to know things. You have to know about the
chemistry of concrete, you have to know about the climate round Aswan.
Intelligence I see we could loosely define as the ability to put these
two facts together. We can define it in more precise entropic terms
and I have indeed posted along those very lines. mErkp (possibly
shortened) means (together with knowledge of concrete and the climate
of Aswan) means "The struggle against the climate/environment". I have
NOT abandoned my earlier definition. To translate it as such though
means that you have to put nuggets of knowledge together.
- Ian Parker
J.A. Legris
To be intelligent all you have to do is the find food and shelter,
avoid being eaten and reproduce. If you can do these things then the
organization of knowledge nuggets will take care of itself. There is
no evidence that it works backwards.
--
Joe
Ian Parker
>
> To be intelligent all you have to do is the find food and shelter,
> avoid being eaten and reproduce. If you can do these things then the
> organization of knowledge nuggets will take care of itself. There is
> no evidence that it works backwards.
>
You need things other than pure intelligence. Evolution also produces
a "selfish" form of self awareness. I feel that it is important to
draw a distinction between this
The claim is that once AI gets to "our level" it will decide that
humanity is superfluous. Such a decision would be the result, not only
of "intelligence" but of a system of priorities. Any "evolved"
priorities will have to be "our" priorities. It is important to
realize this. Skynet will never achieve self awareness for this very
reason.
The one danger, and I feel it is the only real danger is that Skynet
will be given a set of priorities malevolently.. I use the term
"Skynet" because it is a well known SF story. George Bush was the most
unpopular sitting President ever and knew he was going out. Suppose he
had had Skynet at his command and decided that Skynet would disobey
Obama and institute Republican values, by coup if necessary. Obama
would be powerless to stop it.
Are you proposing that we try to achieve AI by genetic algorithms? How
crude our you going to be? Do we abandon Bleu and Meteor and simply
have a group of Arab women who speak no English?
In fact if you look at the whole Maximum Entropy idea you will find
that the algorithms used are already quite close to being genetic.
"Natural selection - survival of the fittest". Humpty Dumpty then
replied that for him "fitness" was inversely proportional to the
entropy or compressed size. Indeed if I perform an operation which
increases entropy I have a monotonic function. Evolution is almost by
definition monotonic.
- Ian Parker
zzbu...@netscape.net
wrote:
> Canadian high-tech startup Intelligence Realm is constructing a
> distributed virtual brain one computer at a time. Utilizing a
> computational model we’ve seen in such projects as SETI@Home, the
> system will harness the computing power of thousands of machines
> throughout the world to achieve its goal.
Well, it's always been simple to achieve consciousness when it's
implicity
defined as spying though,, So that's why the people who don't come
from
AT&T land work on HDTV Debuggers, Blue Ray, GPS, Broadband,
Digital-Terrain Mapping, Fiber Optics, Holograms, Atomic Clock
Wristwatches,
USB, XML, On-Line Publishing, AUVs, and Self-Assembling Robots
though.
>
> http://www.syntheticthought.com/st/artificial-intelligence/61/158
Neil W Rickert
>On 31 May, 22:43, Neil W Rickert <rickert...@cs.niu.edu> wrote:
>> Ian Parker <ianpark...@gmail.com> writes:
>> >The assumption here seems to be that AI is a question of raw computer
>> >power. This has never been the case.
>> Agree.
>> > =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0 =A0=
> AI is a matter of formalizing
>> >knowledge and organising knowledge.
>> And that's never going to work either.
>What we really need to get our hands round is the following. To be
>intelligent you have to know things. You have to know about the
>chemistry of concrete, you have to know about the climate round Aswan.
Sure. But knowing things requires knowing about how they relate
to reality, and not merely having a organized and formalized store
of sentences (or other symbolic representations).
Neil W Rickert
>To be intelligent all you have to do is the find food and shelter,
>avoid being eaten and reproduce. If you can do these things then the
>organization of knowledge nuggets will take care of itself. There is
>no evidence that it works backwards.
Yes, that's about how I look at it.
Merciadri Luca
Hash: SHA1
For some weeks, the website http://www.intelligencerealm.com/ seems to
be dead. Why? Any news?
Thanks.
- --
Merciadri Luca
See http://www.student.montefiore.ulg.ac.be/~merciadri/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/>
iEYEARECAAYFAkqjkC0ACgkQM0LLzLt8MhypZwCgr1OP4/capV6y1HU8arViQwbj
jakAoJJDdVHOg+FrcS+NDP0dcHs2F2Xo
=DwbY
-----END PGP SIGNATURE-----
Kent Paul Dolan
> For some weeks, the website
> http://www.intelligencerealm.com/ seems to be
> dead. Why? Any news?
The actions of some merciful deity favoring
intelligence over hype, perhaps?
xanthian.
"Ping" shows that the site name successfully
translates to an IP address, but that IP
address is dead to ping inquiries.
Merciadri Luca
Hash: SHA1
Kent Paul Dolan <xant...@well.com> writes:
> The actions of some merciful deity favoring
> intelligence over hype, perhaps?
Who knows!
- --
Merciadri Luca
See http://www.student.montefiore.ulg.ac.be/~merciadri/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/>
iEYEARECAAYFAkqk2+gACgkQM0LLzLt8Mhwu+wCfbdQygP/B/n3Ysg/JHyCOJSxc
qNEAn2J6OeivxpvWBWRGNlnne8G54v7Z
=oZjf
-----END PGP SIGNATURE-----
Merciadri Luca
Hash: SHA1
Neil W Rickert <ricke...@cs.niu.edu> writes:
> Sure. But knowing things requires knowing about how they relate
> to reality, and not merely having a organized and formalized store
> of sentences (or other symbolic representations).
Agree. For me, Wolfram is more like a motor of knowledge: it is
built-in with general and more precise facts, but it cannot learn from
its proper experience. Learning from one's experience is one of the
most important characteristics of A.I.
- --
Merciadri Luca
See http://www.student.montefiore.ulg.ac.be/~merciadri/
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.9 (GNU/Linux)
Comment: Processed by Mailcrypt 3.5.8 <http://mailcrypt.sourceforge.net/>
iEYEARECAAYFAkqmI/QACgkQM0LLzLt8MhzLVgCeNXV4JVZo1CxZW8ug6dhByFIt
5WkAmwbqYFm6w6aHPed4IVR5iZix1gqP
=iI6G
-----END PGP SIGNATURE-----