Newsgroups: comp.ai.doc-analysis.misc,comp.ai.nat-lang
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!gatech!newsfeed.internetmci.com!in2.uu.net!bcstec!nntp
From: Richard Wojcik <rwojcik@redwood.rt.cs.boeing.com>
Subject: Re: Software Knows to Deconstruct in Plain English
X-Nntp-Posting-Host: snake.rt.cs.boeing.com
Content-Type: text/plain; charset=us-ascii
Message-ID: <312CA13F.222A@redwood.rt.cs.boeing.com>
Sender: nntp@bcstec.ca.boeing.com (NNTP News Access)
Content-Transfer-Encoding: 7bit
Organization: Research & Technology
References: <4ftflg$st9@la1.digilink.net> <4gdgnn$d2h@la1.digilink.net> <DOWDING.96Feb20182843@Gansett.ai.sri.com> <312B54B2.28DA@redwood.rt.cs.boeing.com> <4gftku$j48@agate.berkeley.edu>
Mime-Version: 1.0
Date: Thu, 22 Feb 1996 17:00:47 GMT
X-Mailer: Mozilla 2.0 (X11; I; SunOS 5.4 sun4c)
Lines: 42
Xref: glinda.oz.cs.cmu.edu comp.ai.doc-analysis.misc:136 comp.ai.nat-lang:4577

John Lazzaro wrote:
> 
> In article <312B54B2.28DA@redwood.rt.cs.boeing.com>,
> Richard Wojcik  <rwojcik@redwood.rt.cs.boeing.com> wrote:
> >
> >I've been following this thread with some interest.  The problem with AI
> >is that it can never live up to people's expectations, even when it does
> >some things spectacularly well.
> 
> Could these expectations be fueled by press releases like this? :
> 
> >"This is the Holy Grail" this is what people have sought for years,"
> >said Jack Reynolds, marketing director for Intelligent Text Processing
> >(ITP). "Top computer-linguistics experts see this as the first virtual
> >intelligence system."

It's true that this kind of marketing hype exacerbates the problem, but I
wasn't thinking of that when I made my comment.  People don't need marketing
hype to have unrealistic expectations about how well computers can handle
natural language.  It is also true that people in the field who take delight
in breaking programs (myself being a guilty party to this, too) also do
damage in conveying the impression that we should expect programs to understand
natural language as well as, or better than, normal humans.  You can break
any NL if you try hard enough.

I think that the litmus test for any program is whether it can make itself 
useful to the computer user, not whether it performs at a human level of
intelligence.  It is good to strive for perfection, but you are always going
to be frustrated if perfection is the only criterion for success.  An NL 
interface that communicates poorly can be effective if the user can learn to
make it perform services that other interfaces (e.g. GUIs) don't do as well
at.  Now, I am not saying that Inquizit passes the test.  I just think that it
ought to be given some leeway to try.  If your query produces bad results, try
rephrasing it to see if you can get better ones.  With a little patience, you 
may find that the program rewards you with good results.

-- 
Richard H. Wojcik, Natural Language Processing
Boeing Information & Support Services
P.O. Box 3707, MS 7L-43, Seattle, WA 98124-2207 (phone: 206-865-3844)
rwojcik@redwood.rt.cs.boeing.com                (fax:   206-865-2965)
----Opinions expressed above are not those of The Boeing Company----
