home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.ai
- Path: sparky!uunet!zaphod.mps.ohio-state.edu!darwin.sura.net!ua1ix!athos.cs.ua.edu!rsun
- From: rsun@athos.cs.ua.edu (Ron Sun)
- Subject: Re: AI Winter
- Message-ID: <1992Aug28.191729.105759@ua1ix.ua.edu>
- Sender: news@ua1ix.ua.edu
- Nntp-Posting-Host: athos.cs.ua.edu
- Organization: University of Alabama Department of Computer Science
- References: <1992Aug28.125825.7628@csc.ti.com>
- Date: Fri, 28 Aug 1992 19:17:29 GMT
- Lines: 17
-
- In article <1992Aug28.125825.7628@csc.ti.com> thrift@ra.csc.ti.com (Phil Thrift) writes:
- >knowledge-based approaches are declining, learning-based approaches
- >are on the upswing (see the latest addition of Winston's AI text).
- >The basic idea is that in organizations there's lots of undigested
- >data but litle knowledge in many important problem domains (the push
- >for knowledge-based AI came from the idea that there's lots of human
- >knowledge in organizations that we could computerize). Whether
- >machine learning, data mining, neural nets, statistical modeling or
- >whatever the hell you want to call it comes to the rescue -- we shall
-
- Although interesting, this thing is no longer AI: it has nothing to do
- with the original grand goal of understanding intelligence and cognition.
- I hope that AI can be a discipline that has its own consistent program
- and stands on its own merit (a minimal requirement for any discipline,
- I gusee), rather than being whatever "AI people" are doing.
-
- --Ron
-