>> The value of AIXI is not that it tells us how to solve AGI. The value is
>> that it tells us intelligence is not computable
Define "not computable" Too many people are incorrectly interpreting it to
mean "not implementable on a computer".
----- Original Message -----
From: Matt Mahoney
To: agi@v2.listbox.com
Sent: Friday, October 24, 2008 10:49 AM
Subject: **SPAM** Re: [agi] If your AGI can't learn to play chess it is no AGI
The value of AIXI is not that it tells us how to solve AGI. The value
is that it tells us intelligence is not computable.
-- Matt Mahoney, [EMAIL PROTECTED]
--- On Fri, 10/24/08, Mark Waser <[EMAIL PROTECTED]> wrote:
From: Mark Waser <[EMAIL PROTECTED]>
Subject: Re: [agi] If your AGI can't learn to play chess it is no AGI
To: agi@v2.listbox.com
Date: Friday, October 24, 2008, 9:51 AM
>> E.g. according to this, AIXI (with infinite computational power)
but not AIXItl
>> would have general intelligence, because the latter can only find
regularities
>> expressible using programs of length bounded by l and runtime
bounded
>> by t
<rant>
I hate AIXI because not only does it have infinite computational
power but people also unconsciously assume that it has infinite data (or, at
least, sufficient data to determine *everything*).
AIXI is *not* a general intelligence by any definition that I would
use. It is omniscient and need only be a GLUT (giant look-up table) and I
argue that that is emphatically *NOT* intelligence.
AIXI may have the problem-solving capabilities of general
intelligence but does not operate under the constraints that *DEFINE* a general
intelligence. If it had to operate under those constraints, it would fail,
fail, fail.
AIXI is useful for determining limits but horrible for drawing other
types of conclusions about GI.
</rant>
----- Original Message -----
From: Ben Goertzel
To: agi@v2.listbox.com
Sent: Friday, October 24, 2008 5:02 AM
Subject: **SPAM** Re: [agi] If your AGI can't learn to play chess
it is no AGI
On Fri, Oct 24, 2008 at 4:09 AM, Dr. Matthias Heger <[EMAIL
PROTECTED]> wrote:
No Mike. AGI must be able to discover regularities of all kind in
all
domains.
If you can find a single domain where your AGI fails, it is no
AGI.
According to this definition **no finite computational system can
be an AGI**,
so this is definition obviously overly strong for any practical
purposes
E.g. according to this, AIXI (with infinite computational power)
but not AIXItl
would have general intelligence, because the latter can only find
regularities
expressible using programs of length bounded by l and runtime
bounded
by t
Unfortunately, the pragmatic notion of AGI we need to use as
researchers is
not as simple as the above ... but fortunately, it's more
achievable ;-)
One could view the pragmatic task of AGI as being able to discover
all regularities
expressible as programs with length bounded by l and runtime
bounded by t ...
[and one can add a restriction about the resources used to make this
discover], but the thing is, this depends highly on the underlying
computational model,
which then can be used to encode some significant "domain bias."
-- Ben G
--------------------------------------------------------------------
agi | Archives | Modify Your Subscription
----------------------------------------------------------------------
agi | Archives | Modify Your Subscription
------------------------------------------------------------------------------
agi | Archives | Modify Your Subscription
-------------------------------------------
agi
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/
Modify Your Subscription:
https://www.listbox.com/member/?member_id=8660244&id_secret=117534816-b15a34
Powered by Listbox: http://www.listbox.com
