I like rigorous progress metrics, and I agree with your statement. I
feel that tests should be developed/rediscovered that provide some
definite goals.
I suspect that any general intelligence test is subject to 'cheating',
through prior training/knowledge. Though without a student having a
certain margin of experience, the test may never be passed.
Ideally it would be nice to have a test that determines the "capability
of extending further". And this is the tricky one isn't it, how do we
measure the potential, when we can only measure biased samples?
(Brett.
On 27/12/2012 3:44 AM, Ben Goertzel wrote:
I agree that existing tests for human child intelligence are basically
fine for AGIs being built and taught via a human-development-related
path...
However, I think such tests are useful as qualitative guides for folks
working on AGI. I think if you start interpreting them as rigorous
progress metrics, somebody is going to be able to hack a system to
pass those early childhood tests without any capability of extending
further on to adult intelligence....
-- Ben G
On Wed, Dec 26, 2012 at 11:40 AM, Todor Arnaudov <[email protected]
<mailto:[email protected]>> wrote:
Hi Brett,
Yes, for example myself. We've discussed a bit here, me and Ben, I
think, in 2011 or so.
IMHO it's the most meaningful way of testing "human level" AI, all
other tests are unjustified, if they don't map into capabilities
comparable to human's.
Those are a must tests.
As of formal specifications - I think it's already done and new
explicit definitions are not needed, just references to the
educational standards.
The capabilities of the AGI/SIGI can be mapped to particular age
given average human requirements/educational standards in
particular domains - it can be "as a 3 year old in language (so
and so words, sentence complexity, tenses, vocabulary, ...), 4
years socially (can play with other kids so and so...) and as a
4-5 year old in maths - e.g. "can count to so and so, can
recognize which group of objects consists of more items, which
lengths are equal, etc.
All the tests are already precisely defined in the developmental
psychology textbooks and educational requirements (manuals for
teachers, psychologists etc.) in all possible domains, starting
from the expected capabilities of infants at a given age, going to
tests for a child readiness to go to school (similar to ones in
your paper), to all kinds of examination in all subjects in school
etc.
I don't think that anything special should be defined, just the
AGI should be made to be general enough to learn to deal with all
that.
For example that series for preschool children:
http://izdatelstvo-slovo.com/poredica.php?poredica=1
To me is an already made fine grained AGI test for human-level
intelligence for ages 3-7.
..../ Todor "Tosh" Arnaudov ..../
*
.... Twenkid Research:* http://research.twenkid.com
<http://research.twenkid.com/>
.... *Self-Improving General Intelligence Conference*:
http://artificial-mind.blogspot.com/2012/07/news-sigi-2012-1-first-sigi-agi.html
*.... Todor Arnaudov's Researches Blog**:
*http://artificial-mind.blogspot.com
<http://artificial-mind.blogspot.com/>
* *From:* Brett Ward <[email protected]
<mailto:[email protected]>>
* *To:* AGI <[email protected] <mailto:[email protected]>>
* *Subject:* A test for less narrow artificial intelligence
* *Date:* Wed, 26 Dec 2012 21:58:14 +1100
Hi All,
Has anyone defined 'preschool' training/testing for AGI?
Here's my draft:
https://docs.google.com/document/d/1ISOrxL2i_K-w_FEpp9c_KXw-Qwf7bhcILCla72wm5xg/edit
Thanks for your time.
*AGI* | Archives <https://www.listbox.com/member/archive/303/=now>
<https://www.listbox.com/member/archive/rss/303/212726-deec6279> |
Modify <https://www.listbox.com/member/?&;> Your Subscription
[Powered by Listbox] <http://www.listbox.com>
--
Ben Goertzel, PhD
http://goertzel.org
"My humanity is a constant self-overcoming" -- Friedrich Nietzsche
*AGI* | Archives <https://www.listbox.com/member/archive/303/=now>
<https://www.listbox.com/member/archive/rss/303/23142996-0ca616e2> |
Modify
<https://www.listbox.com/member/?&;>
Your Subscription [Powered by Listbox] <http://www.listbox.com>
-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424
Modify Your Subscription:
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657
Powered by Listbox: http://www.listbox.com
