Newsgroups: comp.ai
Path: cantaloupe.srv.cs.cmu.edu!bb3.andrew.cmu.edu!newsfeed.pitt.edu!gatech!newsfeed.internetmci.com!in1.uu.net!usc!news.isi.edu!gremlin!shomase!jbarnett
From: jbarnett@shomase.NoSubdomain.NoDomain (Jeff Barnett)
Subject: Re: How is AI going?
Message-ID: <DnM7u2.H5w@gremlin.nrtc.northrop.com>
Sender: news@gremlin.nrtc.northrop.com (Usenet News Manager)
Reply-To: jbarnett@charming.nrtc.northrop.com
Organization: Northrop Automation Sciences Laboratory
References: <4h7a2q$rqh@portal.gmu.edu> <4h7ipu$9ii@Venus.mcs.com>
Date: Sat, 2 Mar 1996 00:54:02 GMT
Lines: 34


|> In article <4h7a2q$rqh@portal.gmu.edu>,
|> Kirt Undercoffer <kunderco@osf1.gmu.edu> wrote:
|> >To clarify my previous posting:  it appears that the central
|> >issues of AI are knowledge representation, inference, learning,
|> >and pattern recognition.  Once a specific problem is posed (machine
|> >translation for example) then we can get at very specific questions.
|> >My question is basically what are the great questions of AI from the
|> >standpoint of knowledge representation in general, inference in general,
|> >ect.  The only question of this sort that comes to mind is the frame
|> >problem.  Are there 9 other questions of the same nature that could be
|> >posed on the same sort of overarching level?

Alan Newell wrote in several places that he thought that the problem
of goal representation was of major importance.  I agree.  If goals
are represented symbolically, our solvers can't have the solutions
wired in so easily.

If I remember correctly, he tested a system written by someone else
that solved the Tower of Hanoi problem from a goal description.  Alan's
test was to change the goal to that of the Tower of Warsaw problem --
Warsa is like Hanoi except that you can put a large disk on top of a
smaller one.  The system found a solution alright but it took 2^n
steps!  Thus, th4e programmer had unwittingly wired in the solution.

This same issue was raised about Lenat's AM, a system that had 
heuristics about what is "interesting".  Its other base knowledge
was a model of bags (sets with multiplicity).  AM (re)discovered
may fascinating number theory concepts.  I thought that AM was the
neatest work of its time.  However, the question of how its goal
-- to find interesting concepts -- was represented was a major
question and concern of the AI community.

Jeff Barnett
