From newshub.ccs.yorku.ca!torn!utcsri!rpi!usc!wupost!uunet!psgrain!qiclab!nosun!techbook!szabo Tue Jul 28 09:41:35 EDT 1992
Article 6472 of comp.ai.philosophy:
Xref: newshub.ccs.yorku.ca rec.arts.sf.science:2781 comp.ai.philosophy:6472
Newsgroups: rec.arts.sf.science,comp.ai.philosophy
Path: newshub.ccs.yorku.ca!torn!utcsri!rpi!usc!wupost!uunet!psgrain!qiclab!nosun!techbook!szabo
>From: szabo@techbook.com (Nick Szabo)
Subject: How do computers fare on scholastic achievement tests?
Message-ID: <1992Jul16.093057.8880@techbook.com>
Summary: A pragmatic definition of intelligence
Organization: TECHbooks --- Public Access UNIX --- (503) 644-8135
References: <NICKH.92Jul13151811@VOILA.VENARI.CS.CMU.EDU> <1992Jul14.014658.4921@newstand.syr.edu> <NICKH.92Jul14141610@VOILA.VENARI.CS.CMU.EDU>
Date: Thu, 16 Jul 1992 09:30:57 GMT
Lines: 34

In article <NICKH.92Jul14141610@VOILA.VENARI.CS.CMU.EDU> nickh@CS.CMU.EDU (Nick Haines) writes:

>(for the sake of discussion, let's specify `intelligent' = `can pass 
>Turing test') are possible...

I've never liked the Turing test.  "Eliza" passed the Turing test
with some people, and it was just a babbling fake Rogerian therapist.
How about let's consider a practical measure of intelligence, eg the
SAT tests?  These are used not only as a practical measure of scholastic 
achievement, but also for membership into MENSA as an alternative to IQ 
tests.  This constitutes a pragmatic, if not philosophically rigorous,
definition of intelligence.

This raises some questions along these lines:

* Which questions would be easiest for a computer?  Which the most
  difficult?  For example, I suspect a souped-up version of Mathematica
  could do pretty well on many of the math questions.  A semantic analyzer 
  with Mathematica might do well on some of the word problems.  Perhaps 
  there is a logic to some of the verbal problems that a parser/semantic 
  analyzer could tackle.  Some questions might be amenable to neural net 
  training if the computer is fed thousands of related practice questions.  
  Etc.
* How difficult for a computer to scan in the questions from a standard
  SAT test book into a recognizable form?
* Have achievement tests changed over the years in response to the
  appearance of computers?  For example, has brute-force calculation
  been deemphasized?  These tests may be another "moving target" that 
  continually redefines intelligence as computers evolve.


-- 
szabo@techbook.COM  Public Access User --- Not affiliated with TECHbooks
Public Access UNIX and Internet at (503) 644-8135 (1200/2400, N81) 


