Date: Sun 13 Mar 1988 20:46-PST From: AIList Moderator Kenneth Laws Reply-To: AIList@KL.SRI.COM Us-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList V6 #51 - Programming, Dependency Propagation, Uncertainty To: AIList@KL.SRI.COM Status: R AIList Digest Monday, 14 Mar 1988 Volume 6 : Issue 51 Today's Topics: Queries - Missing KCL Parts for Sequent Balance & VLSI Testing & Student versions of OPS5, Programming - AI and Software Engineering & Constraint Satisfaction, Theory - Representations of Uncertainty, Philosophy - Chinese Room ---------------------------------------------------------------------- Date: Thu, 10 Mar 88 20:32:15 pst From: George Cross Subject: Missing KCL parts for Sequent Balance Hi, Does anyone have the parts that were left out of KCL for the Sequent Balance? I have the KCL distribution from a recent ftp from rascal and the Sequent parts are missing. ---- George - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - George R. Cross Computer Science Department cross@cs1.wsu.edu Washington State University cross@wsuvm1.BITNET Pullman, WA 99164-1210 Phone: 509-335-6319 or 509-335-6636 ------------------------------ Date: Fri, 11 Mar 88 10:00:11 GMT From: Gabriel Mc Dermott Reply-to: AIList@Stripe.SRI.Com Subject: Re: AIList V6 #50 - Constraints, Neural Nets, Does anyone out there have any references or information on applications of AI techniques to Testability in VLSI design? Thanks in advance, Gabriel. ------------------------------ Date: 10 Mar 88 19:01:53 GMT From: mcvax!unido!tub!tmpmbx!netmbx!morus@uunet.uu.net (Thomas M.) Subject: Re: Student versions of OPS5 In article <1110@pembina.UUCP> ahmed@alberta.UUCP (Ahmed Mohammed) writes: >Computer Thought Corp. has begun shipping students versions >of the expert system lang. These are operational on IBM PC/XT/AT >The package costs $255 for Graduate student version, >$90 for undergraduate version. I would like to know about the capabilities of the above mentioned versions of OPS5. Are there any substantial differences between the "expensive" and the "cheap" version ? I think it is not really the same price-politics as Borland's - 255 $ is not a "Pappenstiel" for a student. By the way: All of you, who helped me with my request for a PC-public domain OPS5-version - thank you very much*. I have now available a few common-lisp sources (each about 100KB big) which I will try to convert to a PC-runnable version in the near future. There will be - I suppose - some difficulties to adapt the sources to PC-Scheme or XLISP. If anybody had experience with porting code from CL to XLisp or Scheme, I would like to contact you for an explanation for some of the more annoying transformations. * I will summarize later! Please - no e-mailing of long files to the above adress (..netmbx..)! For e-mailing of sources etc. please use my BITNET-adress: muhrth@db0tui11.BITNET Thank you very much, Thomas Muhr. -- @(^o^)@ @(^x^)@ @(^.^)@ @(^_^)@ @(*o*)@ @('o`)@ @(`!')@ @(^o^)@ @ Thomas Muhr Tel.: (Germany) 030 - 87 41 62 (voice!) @ @ NET-ADRESS: muhrth@db0tui11.bitnet or morus@netmbx.UUCP @ @ BTX: 030874162 @ ------------------------------ Date: 11 Mar 88 06:11:01 GMT From: munnari!metro.ucc.su.oz.au!daemon@uunet.UU.NET Subject: AI &Software Engineering A recent article on news asked for information regarding software engineering applied to AI. A subsequent article eluded to a lack of information in this area. The CSIRO division of information technology has a research program applying software engineering techniques to the AI area. We have two projects currently on the go, both first generation expert system redevelopment projects. SIRATAC is an expert system that advises cotton growers on what to spray on their cotton crop, and GARVAN ES1 is an expert system that interprets blood samples in a pathology laboratory and inserts an interpretation onto the report in the area of thyroid disorders. Both these expert systems have grown uncontrolled by any design parameters, and subsequently have become difficult to maintain. We are applying data dictionary technology to both these expert systems to fully document, define, and cross reference the knowledge, in the form of production rules. There are several technical reports available as follows tr-fd-87-02 Applying software engineering concepts to rule based expert systems. tr-fc-87-05 Formal Specification of a self referential data dictionary tr-fc-88-01 The knowledge dictionary. A relational tool for the maintenance of expert systems. In addition, another program is implementing a pattern matching hardware engine into UNSW Prolog to speed up the pattern matching process. Copies of any technical reports are available by applying to The Divisional Secretary CSIRO Division of Information Technology PO Box 1599 North Ryde NSW 2113 Australia phone Australia 02 887 9307 fax Australia 02 888 7787 or alternatively by mailing a request to me at the address ACSnet: jansen@ditsyda.oz JANET: ditsyda.oz!jansen@ukc ARPA: jansen%ditsyda.oz@seismo.css.gov CSNET: jansen@ditsyda.oz UUCP: {enea,hplabs,mcvax,prlb2,seismo,ubc-vision,ukc}!munnari!ditsyda.oz!jansen AUSPAC: jansen@au.csiro.ditsyda D ------------------------------ Date: 10 Mar 88 02:52:21 GMT From: kddlab!icot32!nttlab!gama!etlcom!hasida@uunet.uu.net (Koiti Hasida) Subject: Re: constraint satisfaction programming In <5070@pyr.gatech.EDU>, Parvis Avini writes: > I'm looking for some interesting research in the topic of constraint logic > programming or constraint satisfaction programming. I'm already familiar with > Jaffar's and Lassez' work and also with the Prolog III approach. See my article, entitled 'Dependency Propagation', included in IJCAI87 Proceedings, though, I'm afraid, this is not very well-written; less than half of it talks about constraint programming, and natural language is what the rest of it is about. I should work out its constraint programming part in a more complete form. A crucial difference between my DP and others (CLP, Prolog III, etc.) is that DP deals with constraints on combinatorial objects (i.e., the term structures of logic) whereas the constraints considered in the other approaches are about arithmetic objects (rational numbers or real numbers). Another difference is that in DP we look upon processing as constraint transformation. Since the constraint is the program here, processing is a sort of program transformation. Currently under way is an implementation of the interpreter according to DP. This implementation is being done in language C. The first phase of the work is supposed to be finished within one month or two, and will be applied to a natural-language parser based on a unification grammar formalism. I hope this is of some interest to you. HASIDA, Koiti ('HASIDA' is my family name) Machine Inference Section ------------------------------ Date: Thu, 10 Mar 88 10:32:55 EST From: Robert Hummel Subject: Re: Uncertainty and FUZZY LOGIC VS PROBABILITY Concerning the note by Paul Kreelman -- You might find an article by Mike Landy and me of interest. It appears in the North-Holland book of the proceedings of the 1986 workshop on `Uncertainty in AI,' held in Philadelphia. A more complete version of the paper appears in this month's issue (Mar 88) of IEEE Transactions on Pattern Analysis and Machine Intelligence. Bob Hummel New York University hummel@relaxation.nyu.edu ------------------------------ Date: Thu, 10 Mar 88 10:28:48 HNE From: Spencer Star Subject: Re: AIList V6 #48 - CommonLoops, OPS5, Constraint Languages, First I saw a surprising reference to Spiegelhalter quoting him as saying that probabilites were inappropriate for representing uncertainty in expert systems, and then in the March 8th AI LIst, a reference to the original quote. I just don't believe Spiegelhalter made those remarks, or at least they should be put in a context that reflects his views on probabilities. Here's why. > We shall not attempt to review the long, and sometimes acrimonious, > debate as to whether probability theory is an appropriate tool in > this context [expert systems];...Finally the 'statistical/engineering' > model adheres to the probability calculus, justified both from a > theoretical perspective (Lindley, 1982, 1987) and from the pragmatic > claim that it alone provides flexible and operational means of assessment, > criticism and learning (Cheeseman, 1985; Spiegelhalter, 1987). Pearl (1986a) > also argues for probabilistic structuring in expert systems as > providing a good model for human understanding and memory. S.L. Lauritzen and D.J. Spiegelhalter "Local computations with probabilities on graphical structures and their application to expert systems" Oct. 1987 The referenced articles include Lindley, "Scoring rules and the inevitability of probability" Internat. Stat. Review, 50, 1982. and Cheeseman, "In defense of probability" AAAI-85. To put it simply, Spiegelhalter is one of the researchers most committed to putting a probabilistic approach to work in expert systems. My own view is that a subjective probability approach appears to be a better choice than either fuzzy sets or Dempster-Shafer' belief functions because it is the only approach that has the characteristics of 1. Being based on a few simple, acceptable axioms. 2. Being able to connect directly with decision theory (Dempster-Shafer can't) 3. Having efficient algorithms for computation (The Laruitzen-Spiegelhalter paper cited above gives one; Pearl gives another) 4. Being well understood. (Look at what people are doing with Dempster-Shafer belief functions or fuzzy sets. People are not agreed as to what their fundamental theory says. However, if someone prefers one of the other approaches, fine. It's really a question of whether someone wants to work on the mainstream approach, which is Bayesian subjective probabilities or Bayesian decision theory, or if a more experimental approach is preferred, such as fuzzy sets or belief functions. Spencer Star (Bitnet: star@lavalvm1) ------------------------------ Date: 10 Mar 88 20:26:58 GMT From: Adrian G C Redgers Reply-to: Adrian G C Redgers Subject: ...visit to the Chinese Room - some implications In article <8803020915.AA14044@ucbvax.Berkeley.EDU> gjoly@NSS.CS.UCL.AC.UK ("G. Joly", Birkbeck) writes: >...demonstration of the Chinese Room with two Chinese actors >and an English (only) speaking person in the room. >...[the anglophone] had no knowledge of written Chinese; he was merely >manipulating symbols (as computers do). > >..in terms of the Turing test, the room spoke Chinese, since it >satisfied the basic ideas of the test. Agreed that the operator could >not speak the language, but the language was spoken by the program he was >following. > >Does anybody have a ballpark figure for the time needed to run such a program >"by hand''? More or less than the age of the universe? a) I thought Searle's point was that humans might not "understand" Chinese (or English) and are simply manipulating symbols which model the world. The 'Chinese room' is then a brain. Personally I don't go for this because it doesn't give an adequate explanation of conciousness or 'intention'; which I know I've got even if no-one else does. Or was Searle pointing out that the room is unsatisfactory for those very reasons? Ballpark figure is human reaction time. [Searle proposed the room as a challenge for the symbolic-AI school, and would agree with your interpretation. This was discussed at great length in AIList a year or two ago. -- KIL] b) Last night (Wednesday March 9th) BBC1 showed 'Girls on Top' with French & Saunders and Ruby Wax. In it 'Saunders' acts as a jobber waving her arms around and making money in a share dealing room. After rising to dizzy heights of profit she 'crashes'. It transpires that she had no idea what her symbols meant to other dealers - she thought she was making a butterfly and then a bird.... The moral of the story is that 'meaning' or the 'real world' will always outwit (symbol manipulation) systems. I think Aristotle would disagree - but I don't. c) As Jon Silkin put it in his editor's introduction to the Penguin Book of First World War Poetry: Our humanity must never be outwitted by systems, and this is why we are at our most vital when our intelligence is in full and active cooperation with feeling. We shall never not be political again, and the best way to be this, among others, is to think and feel; and if this cooperative impulse is permeated with values we can decently share, we stand a chance, as a species, of survival. For that, I think, is what is at stake. Systems outwitted humanity to cause WW1. Time to move to newsgroup Aristotlelian.conspiracy. love (and peace), Adrian XXX ------------------------------ End of AIList Digest ********************