nhaus@eagle.wesleyan.edu (02/13/91)
About a week ago I posted a request for information regarding KA techniques. Responses were fairly evenly divided between references and requests, so I am posting here a summary of the references I have received. To everyone who responded, thank you! Summary follows: ----MESSAGE 1------------------------------------------------------------- From: Vibhu Mittal <mittal@venera.isi.edu> >You might look at recent issues of the Knowledge Acquisition Journal. There >waas also a recent issue of Machine Learning Journal dedicated to KA >(December 1989). And you should also consider some of the theses on this >area in the last few years - say Tom Gruber's on strategic knowledge acq. >and Sandra Marcus' on the system SALT. >Oh, also Mark Musen's system OPAL and Ray Bareiss' system PROTOS. ----MESSAGE 2------------------------------------------------------------- From: Mark Dahl <dahl@bcstec.boeing.com> >I suggest you do a lit search on John Boose. He's given some overviews >at KA workshops and such that should get you started. ----MESSAGE 3------------------------------------------------------------- From: Ian Gent <ipg@cs.warwick.ac.uk> >You may be interested in the research report listed below. >The approach I outline could not seriously be applied directly, but could >sensibly be used as a framework for judging other approaches to checking >knowledge bases. In particular, coming from a logic background, I am >claiming that logic has its place even in such an applied area. >==== >Details: >Finding Problems in Knowledge Bases Using Modal Logics >Ian Gent >Department of Computer Science, University of Warwick, UK >Research Report 179 >Feb 1991 >Abstract: >In this paper I propose that it is suitable to consider some >statements that an expert makes during knowledge elicitation as being >statements in a modal logic. This approach gives us several advantages in >finding inconsistencies between a knowledge base and >an expert's intuitions of her field. I illustrate this approach >by using the modal logic VC, a logic of counterfactual conditionals. >In an appendix, I give brief details of theorem proving in VC. ----MESSAGE 4------------------------------------------------------------- From: Ian Gent <ipg@cs.warwick.ac.uk> >I made a small mistake in the mail I sent you. The research report I >described is number 173, not 179 as I said before. ----MESSAGE 5------------------------------------------------------------ From: Jim Brule' <jfbrule@rodan.acs.syr.edu> >Might I suggest the book, "Knowledge Acquisition," Brule' & Blount, >McGraw-Hill 1989. In it we describe a methodology for knowledge >acquisition which is based on cybernetic learning theory and >clinical psychology. ----MESSAGE 6------------------------------------------------------------- From: Linda Rodi <rodi@linus.mitre.org> >In regard to your question on sources of information on knowledge >acquistion: check the SIGART Newsletter #108, April 1989. This was a >special issue on knowledge acquisition and contains a large >bibliography on KA. ----MESSAGE 7------------------------------------------------------------- From: "Sanjiv K. Bhatia" <sanjiv@hoss.unl.edu> >A very good source for KA papers is a four-volume collection edited by John >Boose and Brian Gaines. The volumes are called "Knowledge Acquisition for >Knowledge-Based Systems" and are published by Academic Press. I have an online >list of papers (with abstracts) from the first volume and quite a few from the >second. If you want, I can mail it to you. In the other case, you can also >access my complete bibliography by anonymous ftp from fergvax.unl.edu (IP # >129.93.33.1). This bibliography has about 160 documents (quite a few on KA) >and is in the refer format. ----MESSAGE 8------------------------------------------------------------- From: Evangelos Simoudis <simoudis@chaos.cs.brandeis.edu> >I have been working on a methodology for acquiring a particular type >of knowledge that is used by case-based expert systems. I had >published a paper in the knowledge acquisition workshop (November >1990). If you do not have access to the proceedings I could send you a >copy of the paper. I would need however your surface mail address. ----MESSAGE 9------------------------------------------------------------- From: Steve Cousins <sbc@informatics.WUstl.EDU> >I have been teaching a course on ES with emphasis on Knowledge >Acquisition for about 4 semesters now. I haven't used the same book >twice, because new and better ones keep coming out. I believe I have >about 4 with KA in the title now (but they're at work, so I can't send >the references right now). However, I don't think any really address >formal models of KA, because that's really hard -- like coming up with >a formal model of psychology, I'd think. ----MESSAGE 10------------------------------------------------------------ From: Nikolaus Haus <NHAUS@eagle.wesleyan.edu> While looking for some of the above references, I found two special editions of the International Journal for Man-Machine Studies: vol 26 no 1 and vol 26 no 2 (Jan 1987 and Feb 1987, resp.), which are entitled "Knowledge Acquisition for Knowledge Acquisition Systems", edited by John Boose and Brian Gaines. While these may not be the most recent work on the subject, they could provide good background and overview. ----END OF MESSAGES------------------------------------------------------- Again, thank you to everyone who replied!