KRULWICH@C.CS.CMU.EDU.UUCP (03/02/87)
There seems to be a trend nowadays to use the phrase "expert systems" to mean rule-based systems, not to mean any systems that mimick expert behavior. While I'm not sure I like the terminology, I think that it's beneficial to have a seperate catagory for rule-based-systems work, since that's often very different from other A.I. work (especially in describing research work) This opinion may, however, be biased by my opinions of current work in AI and expert systems. What do others think?? Bruce Krulwich If you're right 95% of the time, why worry about the other 3% ?? arpa: krulwich@c.cs.cmu.edu bitnet: krulwich@c.cs.cmu.edu Any former B-CC'ers out there?? uucp: ... !seismo!krulwich@c.cs.cmu.edu
jvc@stl.stc.co.UK.UUCP (03/11/87)
In article <8703040725.AA27188@ucbvax.Berkeley.EDU> KRULWICH@C.CS.CMU.EDU (Bruce Krulwich) writes: > >There seems to be a trend nowadays to use the phrase "expert systems" to >mean rule-based systems, not to mean any systems that mimick expert >behavior. While I'm not sure I like the terminology, I think that it's >beneficial to have a seperate catagory for rule-based-systems work, >since that's often very different from other A.I. work .... I'm inclined to agree. Once upon a time, "Knowledge Based System" equalled "Expert System" equalled "Rule Based System", none of which equalled "AI System". AI sympathists looked askance at the sudden mushrooming of expert systems with suspicion and cynicism as a band- wagon for squeezing as much cash as possible out of gullible sponsors. (And perhaps the old "if it's *that* easy to do, it can't be AI" attitudes came around again...) But KBS work is now returning to the stable, and concerning itself more and more with "real AI" (!) issues - use of metaknowledge for planning and control, problems of learning, ... so now when people say "expert system" they could mean a KBS or they could mean a "first generation" rule-based system. My guess is that KBS will replace ES as the preferred term for forthcoming systems, and that ES will shrink to denoting the things that you make using a commercially- available ES shell: typically, rule-based (and that don't mean much more than computer-based) systems. As for the other question of whether an ES should explain itself: it's fairly easy to make a RBS give some kind of explanation, and so it's been done frequently. The domain and user context might not require it, and the nature of the explanation might be useless anyway, but .... I'd go along with the other correspondents who argue that a KBS might just not have access any more (at the time you asked for it) to the 'exact' reasons for its outputs and that maybe there is no 'exact' reason if there is indeterminacy/context-dependency built in. Generally, the ability to give explanation on demand seems to be only an optional, useage-dependent, external characteristic rather than an essential universal internal one. Vic Churchill ( ...!mcvax!ukc!stl!jvc +44-279-29531 x 2546) STL Ltd., London Road, Harlow, Essex CM17 9NA, U.K.