home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.ai
- Path: sparky!uunet!pipex!pavo.csi.cam.ac.uk!rwab1
- From: rwab1@cl.cam.ac.uk (Ralph Becket)
- Subject: Re: Abductive Reasoning?
- Message-ID: <1993Jan26.101138.148@infodev.cam.ac.uk>
- Sender: news@infodev.cam.ac.uk (USENET news)
- Nntp-Posting-Host: lelaps.cl.cam.ac.uk
- Reply-To: Ralph.Becket@cl.cam.ac.uk
- Organization: U of Cambridge Comp Lab, UK
- References: <1k29msINN4bv@MINERVA.CIS.YALE.EDU>
- Date: Tue, 26 Jan 1993 10:11:38 GMT
- Lines: 22
-
- I suppose I would say that abduction is the inferential process of
- coming up with possible explanations for things. Josephson et al (1987)
- propose a definition of abduction:
-
- D is a collection of data (facts, observations, givens etc);
- H explains D (ie. if H were true, then it would imply D);
- no other hypothesis explains D as well as H;
-
- therefore, H is correct.
-
- The third line in the definition tries to remove wild or implausible
- explanations from the set of solutions.
-
- Abduction is used in machine learning, for example. A fun paper to read
- with all this sort of stuff in is Brian Falkenhainer's "A Unified
- Approach to Explanation and Theory Formation." Falkenhainer talks about
- abduction (amongst other things) in PHINEAS, a proggy which uses analogy
- to infer explanations of phenomena it hasn't seen before.
-
- Hope that's not too far off the mark,
-
- Ralph
-