From: Ben Goertzel (ben@goertzel.org)
Date: Mon Jul 18 2005 - 07:57:30 MDT
Hi Stephen,
We also have test suites that we use to test specific functionalities of our
system, such as language processing or quantitative pattern recognition.
What Michael and I are talking about, I think, is a more comprehensive test
suite that tries to encompass a wide variety of aspects of intelligent
system function, including language, puzzle-solving, math,
pattern-recognition, sensorimotor learning, etc. etc.
-- Ben
> -----Original Message-----
> From: owner-sl4@sl4.org [mailto:owner-sl4@sl4.org]On Behalf Of Stephen
> Reed
> Sent: Monday, July 18, 2005 9:49 AM
> To: sl4@sl4.org
> Subject: RE: Flashback to 1975
>
>
> On Mon, 18 Jul 2005, Ben Goertzel wrote:
> > Yes, I also agree that such an "AI IQ test suite" would be a
> nice thing to
> > have, and I'd also have my AI team devote resources to it if we had a
> > sufficient excess...
>
> At Cycorp we have hundreds of knowledge base content queries that we run
> frequently as a regression test.
>
> Typical questions, paraphrased in English:
>
> Can a dead person read?
> Can bacteria do math?
> Do horses live under water?
> What is the approximate diameter of a basketball?
> Can boulders float?
> Does owning three cars entail owning at least two vehicles?
>
> These tests are useful for tuning the performance of symbolic deductive
> inference, and for ensuring that ongoing knowledge (facts/rules) entry,
> and programming changes do not adversely effect the regression test suite.
>
> Cheers.
> -Steve
>
> --
> ===========================================================
> Stephen L. Reed phone: 512.342.4036
> Cycorp, Suite 100 fax: 512.342.4040
> 3721 Executive Center Drive email: reed@cyc.com
> Austin, TX 78731 web: http://www.cyc.com
> download OpenCyc at http://www.opencyc.org
> ===========================================================
>
>
This archive was generated by hypermail 2.1.5 : Wed Jul 17 2013 - 04:00:51 MDT