Date: Sat 25 Jun 1988 15:16-EDT From: AIList Moderator Nick Papadakis Reply-To: AIList@AI.AI.MIT.EDU Us-Mail: MIT Mail Stop 38-390, Cambridge MA 02139 Phone: (617) 253-2737 Subject: AIList Digest V7 #44 To: AIList@AI.AI.MIT.EDU Status: R AIList Digest Sunday, 26 Jun 1988 Volume 7 : Issue 44 Today's Topics: Queries: heuristics online thesaurus and database oral surgery expert system proposal Response to: Connectionist Expert Systems (MACIE) competitive learning Query Dbase III Plus with Turbo Prolog CLIPS on Apollo? ---------------------------------------------------------------------- Date: Mon, 20 Jun 88 14:10:42 EDT From: Nicky Ranganathan Subject: heuristics I am currently attempting a classification of heuristics based on the "type" of the heuristic and the manner in which it is used. For example, a heuristic such as "A smell of gas in the air could signal the presence of a leak in a pipe" is an evidential association, linking some observable feature of the environment to a probable fault or hypothesis. Heuristics that impose orderings on the reliability of components are essentially built on statistical experience, and can be used to discriminate between competing hypotheses. I would appreciate ideas/examples/pointers to heuristics like the above, from any domain, how they could be classified, and how they would be used. Also, any speculations comparing the general nature of heuristics for different tasks, such as diagnosis, planning and prediction would be much appreciated. Please e-mail responses to me. If there is sufficient interest I will post a summary or e-mail those who are interested. Thanks in advance. --Nicky ------------------------------------------------------------------------------ Nicky Ranganathan Internet: nicky@vx2.GBA.NYU.EDU Information Systems Dept. UUCP: ...{allegra,rocky,harvard}!cmcl2!vx2!nicky New York University Bitnet: nicky@nybvx1 ------------------------------ Date: Mon, 20 Jun 88 18:45:43 EDT From: HSINCHUN CHEN Subject: online thesaurus and database I am working on a dissertation topic which involves applying AI techniques and methodologies on online information retrieval systems, such as online catalog and online bibliography database. This research has been conducted for the last two years. Some results were reported in AAAI87. At the current stage I am developing a program which can serve as an online information specialist between the searchers and the retrieval system. I am also actively looking for some kind of online database (book records; in the areas of computer science, information systems, and finance) and online thesaurus (preferrably based on the Library of Congress subject headings classification scheme). Any pointers or information regarding these electronic forms of thesaurus and database are highly appreciated. For those of you who have similar research interests, I would also be very happy to exchange ideas and thoughts. Please contact me through email. Hsinchun Chen Information Systems area New York University ADDRESS: HCHEN@VX2.GBA.NYU.EDU TEL: 212-9984205 ------------------------------ Date: Wed, 22 Jun 88 15:01 CST From: Subject: oral surgery expert system proposal Two of us -- a faculty member in the College of Dentistry and a knowledge engineer in the Computer Science Department at the University of Tennessee at Memphis -- have developed the following proposal: Develop an expert system to teach undergraduate oral surgery students the most appropriate means of patient evaluation and thereby provide the best treatment plans for patients at our clinics. Students will enter facts relating to specific patients into a database on a Macintosh II. Using expertise in the knowledge base, the computer will determine appropriate preoperative patient evaluation and then print those findings for students to use in patient treatment and study later. Patient histories and data in the database will be used for further research and evaluation. Programming, hardware and development will cost an estimated $25,000. THE BIG QUESTION: Does anyone out there know any possible places for funding of this project? Any ideas for sources of funding would be appreciated. Thanks much. Contact BHIPP@UTMEM1 or PMACLIN@UTMEM1. Dr. B.R. Hipp (901 528-6234) or Philip Maclin The University of Tennessee, Memphis, TN. ------------------------------ Date: Thu, 23 Jun 88 12:12:20 EDT From: carole hafner Subject: Response to: Connectionist Expert Systems (MACIE) Prof. Steve Gallant is the author of MACIE, a connectionist expert system that was described in CACM Vol. 31, No. 2 (Feb. 1988). Steve's e-mail address is: sg@corwin.ccs.northeastern.edu His U.S. Mail address is: College of Computer Science Northeastern University Boston, MA 02115 USA He has a number of interesting papers on connectionist learning methods and their applications in expert systems. --Carole Hafner ------------------------------ Date: 24 Jun 88 13:59:00 GMT From: s.cs.uiuc.edu!bhamidip@a.cs.uiuc.edu Subject: competitive learning I have a question that I am hoping that someone out there will be able to clear up. I have been reading about competitive learning in Rumelhart & McClelland "Parallel Distributed Processing", and am confused by the way that they have set up the model. Basically what they have is a two level network with all the units in the second layer taking their inputs from the units in the first layer. The units in the second layer are grouped into clusters and all units in a cluster are connected in a way so that they inhibit each other, allowing only one unit in a cluster to become active. What I do not understand is the nature of the inhibitory connections within a cluster. Are they just like the other connections, but with negative weights? Is some special activation function needed to take this into account? When a unit is "learning" can the weights on the inhibitory connections be changed? I think that this is a basic question and therefore would best be answered by email. If I get enough responses or there is some variation in the answers I get I will post the results to the net. Ram Bhamidipaty EMail: {rutgers}!bellcore!alpha!arb or arb@alpha.bellcore.com ps. I am posting this under a friends id. ------------------------------ Date: 24 Jun 88 18:29:32 GMT From: uccba!ucqais!bbeck@ohio-state.arpa (Bryan Beck) Subject: Query Dbase III Plus with Turbo Prolog I recently read an article in AI EXPERT, June 1988 written by Karl Horak about using Turbo Prolog to query Dbase III Plus files. There were two points that were not explained in the article, (1) How to get prolog to read the .dbf and how to get prolog to read the .dbt files. Also the article referencing Fileman, Rick. "Memo to Character Conversion," Aston-Tate Inc. TechNotes, Nov. 1987,pp 15-24. I any one has read this article of have done anything like this, or where I can find these TechNotes. I would greatly appreciate any additional information about trying to do this. Please send replys to my e-mail address. Thanks in advance, Bryan -- Bryan M. Beck Univ of Cincinnati College of Business Administration (bbeck@uccba.uc.edu) UUCP: {pyramid,decuac,mit-eddie}!uccba!bbeck ------------------------------ Date: 25 Jun 88 00:02:15 GMT From: cae780!leadsv!kallaus@hplabs.hp.com (Jerry Kallaus) Subject: CLIPS on Apollo? [artifact] HELLLLLLLLLLLLLLPPPPPPPPPPPPPPPPPPPPP!!!!!! I am trying to get the AI tool CLIPS running on an Apollo workstation. The Apollo has Aegis4-DOMAIN/IX rev 9.2.3. CLIPS seems to work fine for small test cases, but apparently prematurely thinks it is out of memory for a problem of any significant size. The code that I am trying to load and run was originally in ART on a Symbolics, was modified to be syntactically acceptable to CLIPS and has been successfully run with CLIPS on a VAX. I have yet to find anyone who knows of anyone using CLIPS on an Apollo. I've tried the CLIPS helpline at NASA and have talked to Apollo software reps already. I am new to CLIPS, the Apollo, and C, which CLIPS is coded in. Oh yeah, this my first USENET posting. ...have mercy! Any info or help would be greatly appreciated. -- Jerry Kallaus {pyramid.arpa,ucbvax!sun!suncal}leadsv!kallaus work (408)742-4569 home (408)732-0217 Funny, how just when you think that life can't possibly get any worse, it suddenly does. - Douglas Adams ------------------------------ End of AIList Digest ********************