From comsat@vtcs1 Tue Nov 5 05:13:20 1985 Date: Tue, 5 Nov 85 05:13:16 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a007941; 5 Nov 85 1:00 EST Date: Mon 4 Nov 1985 20:58-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V3 #161 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Tue, 5 Nov 85 05:01 EST AIList Digest Tuesday, 5 Nov 1985 Volume 3 : Issue 161 Today's Topics: Queries - DAI Contacts & Abduction & User Modelling Panel & ATNS vs. ATTs & Vision Systems and American Sign Language, AI Tools - LISP Workstation Help Facilities, Literature - Proc. 5th Int. Workshop on Expert Systems, Programming Languages - Object-Oriented Language Semantics ---------------------------------------------------------------------- Date: Wed, 30 Oct 85 14:12:36 pst From: Cindy Mason Subject: DAI contacts I have been reading a lot of articles in the area of Distributed Artificial Intelligence (DAI) and would appreciate getting in touch with others who have similar interests to discuss articles and toss around ideas. Thanks. Cindy Mason (clm@lll-crg) ------------------------------ Date: Fri, 1 Nov 85 10:01:23 cst From: Alan Wexelblat Subject: Abduction Will someone please explain to me what is meant by this word? My dictionary gives two definition: one has to do with kidnapping, the other has to do with exercising certain thigh muscles. I assume that AI'ers have a third definition (!?). Replies directly to me, please. --Alan Wexelblat WEX@MCC.ARPA ------------------------------ Date: Fri 1 Nov 85 11:49:21-EST From: John C. Akbari Subject: User Modelling Panel Does anyone have notes on the User Modelling panel held at IJCAI-85 in August? THanks in advance. john akbari akbari@columbia-20.arpa ------------------------------ Date: Fri, 1 Nov 85 18:11:58 pst From: decvax!ittatc!dcdwest!sdcsvax!sdcrdcf!polyslo!cburdor @ucb-vax.berkeley.edu (Christopher Burdorf) Subject: ATNS vs. ATTs I am currently working on a master's thesis in natural language processing. I am currently deciding whether to use ATNs or ATTs to do the parsing. If anyone out there has any feelings one way or the other as to which method is better, please let me know. Chris burdorf Cal poly slo. ------------------------------ Date: Mon, 4 Nov 85 09:58:29 GMT From: gcj%qmc-ori.uucp@ucl-cs.arpa Subject: Vision Systems and American Sign Language One of goals of AI research is to produce speech recognition systems. Has there been a proposal to produce a vision system that can ``read'' ASL? Gordon Joly gcj%qmc-ori@ucl-cs.arpa ------------------------------ Date: 1 Nov 1985 01:35 CST (Fri) From: Paul Fuqua Subject: LISP Workstations Some comments in response to: Date: Thursday, 24 October 1985 13:53-CDT From: Liz Allen To: AIList at MIT-MC Re: LISP Workstations What I really needed was to see the files that were listed off the bottom of the screen... The other big problem I had was in using emacs -- I learned about apropos pretty quickly, but it was not a lot of help. My favorite example is when I wanted to pick up some text without modifying the existing buffer. On the Texas Instruments Explorer (the third of the MIT-derived lispms), we have a system called Suggestions that occupies a small menu strip on your window with a selection of, well, suggestions. Some of the menu items are commands, some switch to more detailed menus of classes of commands, some do other things. As a whole, the suggestions menus are supposed to track the state you're in -- in Zmacs, there are Zmacs menus, with headings like cursor movement, deleting and moving text, font commands, etc; in Dired, there is a menu of the Dired-specific Zmacs commands; in the Lisp Listener, there are input-editing menus, window-switching menus, and so on. (Sorry about the vagueness, but with two years of pre-TI lispm experience, I've never used Suggestions myself; I just gripe about the implementation.) The idea is to try to relate concepts that the user already has in mind to commands or groups of commands. The target person is someone who knows what he wants to do, but not how to do it. Suggestions is by no means perfect; for one thing, it doesn't explain scroll bars. However, it's a start. The documentation is good if you either already know the vocabulary or have someone who can tell you the right word for what you want. So one obvious goal of good documentation is to lead the way to the vocabulary. The most useful feature of the red/black/blue/grey/green/orange Lisp Machine Manual is its concept index. At least it lands me in approximately the right section of the manual, where I can pick up the proper terms for the next time. pf ps I'm not any sort of official TI spokesman, but no one else here was taking a shot at the issue. ------------------------------ Date: 1 Nov 1985 16:55-CST From: leff%smu.csnet@CSNET-RELAY.ARPA Subject: 5th Int. Workshop on Expert Systems and Their Applications Some time ago, I sent a list to this digest of the papers presented at the Fifth International Workshop on Expert Systems and Their Applications at Avignon, France on May 13-15, 1985. I have tracked down the ordering information for those proceedings. I am posting the details here since I have received mail from many individuals who needed articles from there and thus this info is of general info: To order write to: Marie Martine Sainflou Agence de l'Informatique Tour Fiat-Cedex 16 92084 Paris la Defense, France They accepted our purchase order and billed us for 800 French Francs. If you try and order via Interlibrary loan here is the information from the OCLC entry: OCLC: 12661613 1 100 2 040 ISM c ISM 3 020 2865810283 4 041 0 freeng 5 090 TK7885.A1 b P7 1985 6 049 ISMM 7 245 00 [Proceedings] / C Expert ?Systems & Tehir Applicatoins. 5th International Workshop 8 260 0 [Paris] : b AGence de l'Informatique c 1985 9 300 2 v. : b ill. ; c 24 cm 10 500 French and English. 11 500 5 `emes Journbees Internationals, les Syst`emes Experts & Leurs Applications 12 504 Includes bibliographies 13 650 0 Computer engineering x Congresses. 14 711 20 International Workshop on Expert Systems and their Applications n (5th : d 1985 : c Avignon, France) ------------------------------ Date: Wed, 30 Oct 85 16:16:30 est From: "Dennis R. Bahler" Subject: replies to OOL semantics request > Does anyone have pointers to work done on >formal specification and/or formal semantic definition of >object-oriented languages or systems such as Smalltalk-80? Well, the traffic has died away on my request about formal semantics of OOLs and a number of folks have asked to see what I got, so this is it. Dennis Bahler Usenet: ...cbosgd!uvacs!drb Dept. of Computer Science CSnet: drb@virginia Thornton Hall ARPA: drb.virginia@csnet-relay University of Virginia Charlottesville, VA 22903 ------- From: mac@uvacs.UUCP (Alex Colvin) You might check on the work done on PLASMA, an actor ( ~ object) language, mostly applicative. I asked the net about this some time ago, but got no response. Then there's Act I, another MIT-AI project. And who knows what else? Lastly, some folks (Lisp types, mostly), model objects as closures. This leads to flavors. %A Carl Hewitt %T Viewing Control Structures as Patterns of Passing Messages %J Artificial Intelligence %V 8 %D 1977 %P 323-364 %X especially section 7. %A Carl Hewitt %A Brian Smith %T Towards A Programming Apprentice %J IEEE Transactions on Software Engineering %V 1 %N 1 %D March 1975 %P 26-45 %X describes meta-evaluation to justify contracts on implementations %X featuring the Actor Induction Principle %A Henry Lieberman %T A Preview of Act I %R AI Memo 625 %I MIT AI Lab %D June 1981 %X describes the actor model of computation based on message passing %A Henry Lieberman %T Thinking About Lots of Things At Once Without Getting Confused %R AI Memo 626 %I MIT AI Lab %D May 1981 %X synchronization and concurrency in Act I %A C. Hewitt %A G. Attardi %A H. Lieberman %T Specifying and Proving Properties of Guardians for Distributed Systems %B Semantics of Concurrent Computation %S Lecture Notes in Computer Science %V 70 %I Springer Verlag %C Berlin %D 1979 %X synchronization device Since you're just down the hall from me, you can check out my copies. ------ >From johnson%p.cs.uiuc.edu@CSNET-RELAY Thu Oct 10 00:18:08 1985 You recently asked a question on the net about work in semantics for OOLs. I am interested in semantics for OOL, though I haven't done anything worth talking about, so I would appreciate any responses that you get. In general, I don't think that inheritance makes semantics any more difficult, although Smalltalk (which is my interest) has weird "functions that can goto creating environment" things called blocks that require continuation semantics. I least, I think they require continuation semantics, I haven't completely solve the problem yet. I have done some work in type systems for Smalltalk, but I haven't written it up yet. Are you interested in such things? Ralph Johnson ------- >From sokol%mitre.arpa@CSNET-RELAY Fri Oct 11 00:22:09 1985 Dennis We have been using RAND's Rule Oriented Simulation System for about 4 years now and have been very happy with it. You can pick it up in a morning, and can decipher other people's code immediately, It`s a lovely system. For more information, see Rand publication R-3160-AF (1984) and N-1854-AF (1982), or contact Phil Klahr at Rand. He also gives out source code to universities and the like for research. Lisa Sokol (sokol@mitre) [ ROSS was written up in Proc. IJCAI-81 too if I remember -- drb ] ------- >From jisdale%omnilax.arpa@CSNET-RELAY Fri Oct 11 00:22:46 1985 I saw your note in AILIST-DIGEST and it struck a responsive note. I just finished a UCLA Extension class on formal semantics that required a term paper. For that paper I chose to attempt some formalization of "Little Smalltalk", a stripped down Smalltalk-80 that is written in C and does not require (or support) fancy graphics, etc. It is available from Univ. of Arizona for a modest fee. The paper did not really do any formal specification. Since it was limited to 10 pages, I spent most of it giving an intro to OOL & some of the difficulties in formalizing the syntax and semantics. The main point on formalization I found was the ability of Smalltalk to be self-defined. The book "Smalltalk-80, The Language and Its Implementation" does provide a formal specification of the semantics in Smalltalk-80 in Part Four. This was an interesting example of the power of Smalltalk, since very few languages can be self-defining. However, the definition is much longer than the self definition of LISP. I did think there is potential for defining Smalltalk in VDL or other language, but given the time I had (and the level of the class) I did not invest much time on this. I am interested in any responses you get about such formalizations. Jerry Isdale CSNET (X.25 site): jisdale@omnilax from phonenet: jisdale%omnilax@CSNET-RELAY (I think thats right but not sure). US Snail: Omnibus Computer Graphics, Studio G, Paramount Pictures 5555 Melrose Ave, Hollywood, CA. 90038 (213) 468-4694 (Omnibus is a commercial computer animation house with offices in NYC, Toronto and Hollyweird). ------- >From mct%gandalf.cs.cmu.edu@CSNET-RELAY Sat Oct 12 00:13:48 1985 A Semantics of Multiple Inheritance Luca Cardelli in, Semantics of Data Types, SpringerVerlag, Lecture Notes in CS, #173 has a nice treatment of multiple inheritance. -- Mark Tucker ------------------------------ End of AIList Digest ******************** From csvpi@vtcs1 Tue Nov 5 05:16:38 1985 Date: Tue, 5 Nov 85 05:16:31 est From: csvpi@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a009000; 5 Nov 85 2:46 EST Date: Mon 4 Nov 1985 21:10-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V3 #162 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Tue, 5 Nov 85 05:05 EST AIList Digest Tuesday, 5 Nov 1985 Volume 3 : Issue 162 Today's Topics: Games - ACM Computer Chess Championship, Expert Systems - DARPA Funds KEE, Opinion - AIList Discussion Style & Definition of AI & Japanese Fith-Generation Motives ---------------------------------------------------------------------- Date: Sat 2 Nov 85 19:16:33-PST From: Stuart Cracraft Subject: ACM Computer Chess Championship [Forwarded from the MIT bboard by SASW@MIT-MC.] The annual slug-fest of machine against machine in the game of chess has produced a new champion while in the same process dethroning CRAY BLITZ which ran on multiple, parallel, CRAY processors. The new champion, by a perfect 4-0 score against its opponents, is HITECH at Carnegie-Mellon, searching 175,000 chess positions per second. Below are reproduced descriptions of the participants including the names of the computer chess programs, the authors, their affiliation, the type of hardware used, and the number of nodes the particular program searches per second. The time control for this is usually about 2 or 3 minutes per move so you can multiply the nodes per second by 3x60 to get the total nodes executed by a program in order to find its chess move (on the average). In recent years, this number has been on the order of 10^7. There are various schools of thought which believe that this will have to increase by several orders of magnitude before an artificial player will defeat the human champion, unless significant breakthroughs in chess knowledge representation are achieved. Also, in recent years, the best computer chess programs have barely passed the National Master ranking. That is, they have achieved a rating of 2200. The human champion is normally rated beyond 2700. The relationship of processor speed to ratings has been determined to be about 100 points per factor of 2 increase in processor speed, at least in the range up to a 2000 rating. There is some suspicion that beyond this level, the relationship is not linear, although this has not yet been shown by sufficient analysis, either empirical or theoretical. Stuart Cracraft (cracraft@isi-vaxa) ------------- The following information was provided by the author of Phoenix. PROGRAM AUTHOR AFFILIATION HARDWARE N/S _______ ______ ___________ ________ ___ Awit Tony Marsland University of Alberta Amdahl 5860 10 Bebe Tony Scherzer SYS-10 Inc., Chicago Custom Chess 20,000 Engine Chaos Mike Alexander University of Michigan Amdahl 5860 70 Fred Swartz Jack O'Keefe Cray Blitz Robert Hyatt University of Southern Cray X-MP 100,000 Albert Gower Mississippi (4 CPUs) Harry Nelson Hitech Carl Ebeling Carnegie-Melon Special pur- 175,000 Hans Berliner pose hardware Gordon Goetsch Andy Palay Murray Campbell Larry Slomer Intelligent Mark Taylor Intelligent Software Apple IIE 500 Software David Levy Kevin O'Connell Lachex Burton Wendroff Los Alamos Laboratory Cray X-MP 50,000 Ostrich Monty Newborn McGill University 8 Data General 1,200 computers Phoenix Jonathan Schaeffer University of Alberta 4 VAX 11/780s 2,000 6 SUN workstations Spoc Jacques Middlecoff SDI/Cypress Software IBM PC 300 ------------------------------ Date: Mon 4 Nov 85 09:25:01-PST From: Ken Laws Subject: DARPA Funds KEE >From Expert Systems, Vol 2., No. 3, July 1985, p. 166: IntelliCorp has recently been awarded a DARPA contract to develop a prototype expert system development tool. The tool will be used by the Department of Defense, related government agencies, and contractors working on DARPA-funded projects. The contract is worth $1 million to IntelliCorp and will take two years to complete. The new system will be based on a refined version of KEE, incorporating new knowledge representation techniques. IntelliCorp sees limitations in current ways of representing knowledge as being the limiting factor in developing more powerful expert systems. New techniques will allow the full diversity of an expert's knowledge to be used. IntelliCorp will retain exclusive ownership of the KEE system around which the new tool will be built. The compiled version of KEE will be sublicenced to the Department of Defense, related agencies, and DARPA- funded contractors. Ownership of the remainder of the new system will be shared by IntelliCorp and DARPA, with the company retaining exclusive rights to its further release and commercialisation. ------------------------------ Date: Thu 31 Oct 85 19:07:57-PST From: Gary Martins Subject: Contributions of "AI" ? A recent issue of AIList [#156] carries a highly emotional message from Mr. Chris Welty in defense of "AI". The message is entitled "Contributions of AI". Wouldn't you expect such a message to refer to some real contributions of "AI" ? Instead, it contains: * idle speculations about the relationship between my wife and Prof. Minsky * complaints about his morning mail * educational advice * finally, the usual vague, abstract "AI" blah about all kinds of contributions "AI" has made to the world; Mr. Welty says the list is too long to provide in full -- and so he provides no specific information of any kind I cannot pretend to help Mr. Welty with all of these problems. I think he should adopt a "wait and see" posture on the first point. Perhaps a scrolling terminal will help with #2. Advice duly noted. As for the same old "AI" gobbledegook: CAN'T YOU BE MORE SPECIFIC ? ------------------------------ Date: Sat, 2 Nov 85 14:26:53 pst From: ames!eugene@RIACS.ARPA Subject: Re: Minsky's definition of AI (really definition of I) Interesting posting. I'm not doing AI work, but I have something to share. Two weeks ago on the plane down to JPL/Caltech, I read a very interesting definition of "Intelligence" in the airline's magazine (PSA). Intelligence is the ability to simultaneously hold two contradictory thoughts in one's head. I am working on parallelism, and I sort of like that definition. >From the Rock of Ages Home for Retired Hackers: --eugene miya NASA Ames Research Center {hplabs,ihnp4,dual,hao,decwrl,allegra}!ames!aurora!eugene emiya@ames-vmsb ------------------------------ Date: Thu 31 Oct 85 15:21:39-EST From: Steven M. Kearns Subject: The Japanese are Coming! THE JAPANESE ARE CONTROLLING THE WORLD - FUN, FICTION, OR FACT? Here is proof that fiction is actually stranger than truth: the first "correct" exposition of the true goals of the Japanese Fifth Generation program. Actually, this is a fun look at some of the Fifth Generation Hype that might leave you wondering "what if it is true???" (Q1) QUESTION: Why the "open architecture" of the Japanese Fifth Generation effort? Why tell the world exactly what the goals are, the money to be spent, and the principal goals to be pursued? (A1) The Japanese have stated in their Fifth Generation propoganda material that the security of Japan in the future depends on transforming the structure of today's economy. Nowadays it is resource based; Japan would like it to be information based. The reasons for this are clear. Japan imports something like > 90% of their fuel and food. In a war, their enemies could stop these shipments, while Japan could only threaten to cut off next years shipment of remote control VCRs. In addition, Japan has a severe space shortage. Material resources take up critical space, while information does not. If Japan had the power to transform the world economy by themselves, they would. But the truth is, they do not have the resources to do it by themselves. To do so requires mobilizing the world - and in high tech the "world" means the United States. So the Japanese were very clever. By committing a small amount of money (< 1 billion over 10 years, I believe) and publicizing it as much as possible, they managed to steer the lumbering giant of the United States, and the rest of the world, in the direction that they wanted. In effect, the Japanese are investing a little money, and in return they get to mobilize ALOT of money in a way that benefits them the most. And none of the money that they invest is wasted either: if the Japanese Fifth Generation program does provide a significant advance, they get to finally shake the unfair label of "imitators, not creators". If no major advance results, they are at least ready to exploit the successes of the rest of the world. Let's look at some facts that support this theory. As already mentioned, the Japanese committment is less than 1 billion over 10 years. In contrast, IBM's R&D budget is something like one and a half billion EVERY YEAR. And though I do not know the specific data, I would suspect that the defense department's Fifth Generation budget is of similar order of magnitude (after all, the defense budget is now 300 Billion). Finally, there are the contributions from other countries of the world such as Britain and France, which have now mounted efforts comparable to Japan's. Thus, all of the gloom and doomers warning about the onslaught of the Japanese are actually aiding the Japanese' cause. -steve kearns (kearns@cs.columbia.edu) ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Thu Nov 7 23:44:41 1985 Date: Thu, 7 Nov 85 23:44:36 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a014337; 6 Nov 85 15:16 EST Date: Wed 6 Nov 1985 10:37-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V3 #163 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 7 Nov 85 23:26 EST AIList Digest Wednesday, 6 Nov 1985 Volume 3 : Issue 163 Today's Topics: Seminars - CommonLoops (SU) & The Boyer-Moore Theorem Prover (UTexas SIGART) & IEEE Seminar on AI (SMU) & Mental Representations (UCB) & AI in Design and Manufacture (UPenn) & Predicting the Effects of a Therapy (MIT) & Tools for Building Expert Systems (Rutgers) & Very High-Level Programming Environment (CSLI), Conference - ACL 1986 Annual Meeting ---------------------------------------------------------------------- Date: Fri 1 Nov 85 16:42:00-PST From: Susan M. Gere Reply-to: m.susan@sierra Subject: Seminar - CommonLoops (SU) EE380--Seminar on Computer Systems Title: CommonLoops--A Graceful Merger of Lisp and Object Oriented Programming Speaker: Daniel G. Bobrow From: Xerox PARC Time: Wednesday, November 6 at 4:15 p.m. Place: Terman Auditorium CommonLoops merges the facilities of object oriented programming and Lisp. This talk will briefly describe the relevant features of the two styles of programming, and describe the unique properties of this merge. These include a uniform syntax for function calling and sending messages; a merger of the type space of Lisp and the class hierarchy of objects; a generalization of method specification that includes ordinary Lisp functions at one extreme, and fully type specified functions at the other; and a "metaclass" mechanism that allows tradeoffs between early binding and ease of exploratory programming in the implementation of objects. Short Biography: Daniel Bobrow is a Research Fellow in the Intelligent Systems Laboratory at Xerox PARC. His research interests include programming languages, expert systems, artificial intelligence, and cooperative computing. He received his PhD from MIT, started the Artificial Intelligence Department at Bolt Beranek and Newman, and since at Xerox has helped to develop a number of systems, including KRL, GUS, PIE, LOOPS, COLAB, and CommonLoops. ------------------------------ Date: Tue 5 Nov 85 15:21:35-CST From: David Throop Subject: Seminar - The Boyer-Moore Theorem Prover (UTexas SIGART) SIGART, the Special Interest Group on ARTificial Intelligence, has its monthly program meeting WEDNESDAY, 6 Nov, at JIMS restaurant at I-35 and 183 (Anderson). We meet for drinks at 6:30 and dinner to start at 7:00. Charge is $2 for members and $5 for non members (plus food and drinks). The speaker will be Dr J S Moore, speaking on: Applications of the Boyer-Moore Theorem Prover to the Verification of Computer Hardware and Software J Strother Moore The Boyer-Moore Theorem Prover is computer program that proves theorems about recursive functions. The primary application of the program is to prove formulas that establish the correctness, reliability, or security of computer hardware and software. The proof techniques used by the system include rule driven simplification, generalization of the conjecture to be proved, and mathematical induction. Each time a formula is proved the theorem-prover builds it into an evolving knowledge base which is used to structure subsequent proofs. Thus, the human user of the system can improve the system's performance by having it prove key lemmas first. As the theorems get harder the user's role in the process more and more resembles that of the mathematician who sketches proofs before an assistant who fills in the often large gaps. In this talk I will informally explain how the system works and how it is used. I will also discuss some applications of the system, including its use in finding security flaws in the formal specifications of computer software, its proof of the invertibility of the RSA public key encryption algorithm, and the correctness proofs for a general purpose microcoded CPU. ------------------------------ Date: 31 Oct 1985 09:12-CST From: leff%smu.csnet@CSNET-RELAY.ARPA Subject: IEEE Seminar on AI (SMU) The following is the program for the Special Event of the Dallas IEEE Computer Society and Dallas Section of the ACM Artificial Intelligence Satellite Symposium Knowledge-Based Systems and Their Applications presented by Texas Instruments Incorporated Date: Wednesday, November 13, 1985 8:30 am - 4:00 pm Place: Infomart, 1950 Stemmons Freeway Room 7011 Agenda: Welcome and Opening Remarks 8:30 - 8:45 am o Edward E. Feigenbaum Stanford 8:45 - 9:45 AI: An Overview. Knoweldge Engineering & Expert Systems o RAndall Davis MIT 10:00 - 11:00 am Problem Solutions with Expert Systems: Approach, Tools Available, How to Begin o Bruce G. Buchanan Stanford 11:00 - 12:00 pm Knowledge Based Systems: Problem Selection, Knowledge Acquisition, Validation o Mark Fox CMU 1:00 - 2:00 pm Knowledge-Based Systems: Applications in the Induatrial Environment o Harry Tennant, Host TI Inc. 2:00 - 3:00 pm Applications Abstracts by Representatives from AErospace,. Manufacturing, Military, Industrial Control Engineering, and Education o Harry Tennant, Moderator 3:00 - 4:00 pm Presenter's Roundtable - Live Closing Remarks 4:00 [TI is also sponsoring a satellite presentation at Stanford. -- KIL] ------------------------------ Date: Wed, 30 Oct 85 12:05:18 PST From: admin%cogsci@BERKELEY.EDU (Cognitive Science Program) Subject: Seminar - Mental Representations (UCB) BERKELEY COGNITIVE SCIENCE PROGRAM Fall 1985 Cognitive Science Seminar - IDS 237A Tuesday, November 5, 11:00 - 12:30 240 Bechtel Engineering Center Discussion: 12:30 - 1:30 in 200 Building T-4 ``On the Intentional Contents of Mental States About Fictions'' Edward Zalta Postdoctoral Fellow in Philosophy at C.S.L.I. Acting Asst. Professor of Philosophy, Stanford University In this seminar, I present a theory of intentional objects some of which seem to serve nicely as the contents of mental states about stories and dreams (no matter how bizarre they may be). The theory yields a way of understanding utterances about particular fictional characters and particular dream objects. For the purposes of the talk, it will make no difference whether one construes the theory ontologically as a theory about what the world has to be like or has to have in it in order for us to characterize properly such mental states, or whether one construes the theory as just a canonical notation for specifying the contents of (or mental representations involved in) such states. Either way, one is left with a domain over which operations may be defined to explain how we get from one state to the next, and so the theory should be of interest to cognitive scientists. The philosophical basis of my work lies in a theoretical compromise between the views of Edmund Husserl and Alexius Meinong, and it is consistent with classical logic. ------------------------------ Date: Mon, 4 Nov 85 15:24 EST From: Tim Finin Subject: Seminar - AI in Design and Manufacture (UPenn) Professor Robin Popplestone Department of AI at Edinburgh University will give a lecture on Applying AI Techniques to Design and Manufacturing Today: Monday, November 4 at NOON in Towne Building, Room 303 I discuss the representation of mechanical engineering designs in a logic programming context, and the exploration of a space of different possible designs. Designs are represented in terms of modules, which are basic concrete engineering entities (eg. motor, keyway, shaft). Modules interact via ports, and have an internal structure expressed by the part predicate. A taxonomic organisation of modules is used as the basis for making design decisions. Subsystems employed by the design system include the spatial relational inference mechanism employed in the RAPT robot Language, the Noname geometric modeller developed at Leeds Univeristy and the Press symbolic equation solver. The system is being implemented in the POPLOG system. An assumption based truth maintenance system based on the work of de Kleer is being implemented to support the exploration of design space. ------------------------------ Date: Tue, 5 Nov 85 16:49 EST From: Brian C. Williams Subject: Seminar - Predicting the Effects of a Therapy (MIT) Thursday 7, November 4:00pm Room: NE43- 8th floor Playroom The Artificial Intelligence Lab Revolving Seminar Series "Predicting the Effects of a Therapy in a Physiological Network" Bill Long Clinical Decision Making Group, LCS If the physician gives Inderol to the patient to decrease angina, what will happen to the blood pressure? Or more generally, is there anything the physician should watch out for when giving drug X to this patient? An important aspect of the Heart Failure Program is helping the user answer such questions. The program assists in diagnosis by using the patient information to constrain a physiological model to represent the state of knowledge about the patient. That model can then be used to find likely therapies to correct dangerous states and to reason about the possible effects of those therapies. The problems with predicting the effects of the therapies include accounting for multiple causal pathways, accounting for the effects of feedback, reasoning about pathways that take widely differing amounts of time, reasoning when there is uncertainty about the patient state, and reasoning even though there is interpatient variation. In attempting to deal with these problems, we have developed an algorithm based on techniques of signal flow analysis that handles some of these problems well and others acceptably and has the right properties to provide understandable justifications for the conclusions it reaches. The talk will focus on the criteria that are being used in developing this methodology, the algorithm itself, the effectiveness of the approach, and the remaining problems. ------------------------------ Date: 5 Nov 85 17:02:55 EST From: Smadar Subject: Seminar - Tools for Building Expert Systems (Rutgers) III Seminar Title: Issues in the Selection of Knowledge Engineering Environments and Tools for Building Large Expert Systems Speaker: Susan Man Date: Tuesday, November 12, 1985, 11:00am - 12:00pm Place: Hill Center, room 423 Susan Man, a Ph.D. student in our department, will present results of a study on knowledge representation and programming paradigms (done in conjunction with an independent study under Chris Tong). This is her abstract: One of the first decisions that must be made by designers of expert systems is the choice of the knowledge engineering environment and tool to be used for the development of the system. In this talk, we attempt to identify some features of programming environments and knowledge engineering tools that are important in building large expert systems. We first look at features in programming environments on Lisp machines such as the Symbolics 3600's and the Xerox 1100's. We then compare three knowledge engineering tools that are suitable for the development of large-scaled expert systems. The knowledge engineering tools studied are (1) Zetalisp, (2) KEE (from Intellicorp), and (3) S.1 (from Teknowledge). In discussing and comparing the features offered by these knowledge engineering environments and tools, we are particularly interested in their abilities to accommodate various programming methodologies and to provide useful support utilities. Programming methodology, which encompasses the issues of knowledge representation and programming paradigm, impact directly on the ability of the knowledge engineering tool to model precisely and efficiently complex domain tasks and problem solving behaviors. Support utilities offer facilities such as editing, debugging, and explanation and are important factors in reducing the time and effort required in building a large expert system. ------------------------------ Date: Tue 5 Nov 85 14:37:04-PST From: Terry Winograd Subject: Seminar - Very High-Level Programming Environment (CSLI) [Forwarded from the Stanford bboard by Laws@SRI-AI.] COMING ENVIRONMENTS MEETING (11/11) - Steve Westfold (Kestrel) A Very-High-Level-Language Programming Environment Kestrel Institute is doing research on a programming system based on a very-high-level specification/programming language. The language is based on logic and set theory. It is a wide-spectrum language encompassing both an inference model of computation and a state-change model. Compilation is done by transformation and step-wise refinement into the target language (initially Lisp). A central part of the system is the ability to define new language constructs and domain languages, and facilities for manipulating and transforming them. Most of the system is written in the system language. The underlying structure of the environment is a database of objects, sets, sequences and mappings. There is an object hierarchy which is used primarily for factoring applicability of mappings. Language statements (parse structures and annotations) are represented in the database. We identify the representation of statements with the meta-level description of those statements. Thus, meta-level inference on descriptions results in statement manipulation such as transformation. Usually the programmer need not be aware of the representation because of a quotation construct that is analogous to lisp backquote, but is more powerful and can be used for testing and decomposing statements as well as constructing them. Among the ways that the user may view portions of the database are as prettyprinted language statements, as objects with properties, and as graphs of boxes and arrows. The database may be edited using any of these views. The system enforces constraints stated as implications (universally quantified) with an indication of the triggers for enforcement and of the entities to change to make the constraint true. We have a context tree mechanism for keeping different states of the database. It is somewhat smart in that it does not save undo information for database changes that are "internal" to the current state. It would have wider application if it were able to work on subsets of the database rather than the database as a whole. We have recently built a prototype for a project management system. It deals with system components and their versions and bugs, and tasks and schedules. This work is at a fairly early stage and not my area so I wouldn't want to talk much about the details of it, although someone else at Kestrel might. However, it does provide good examples of the utility of the language-defining and constraint capabilities in a domain other than program synthesis. ------------------------------ Date: Thu, 31 Oct 85 16:47:40 est From: walker@mouton (Don Walker) Subject: CALL FOR PAPERS; ACL 1986 Annual Meeting CALL FOR PAPERS 24th Annual Meeting of the Association for Computational Linguistics 10-13 June 1986, Columbia University, New York, NY, USA SCOPE: Papers are invited on all aspects of computational linguistics, including, but not limited to, pragmatics, discourse, semantics, and syntax; understanding and generating spoken and written language; linguistic, mathematical, and psychological models of language; phonetics and phonology; speech analysis, synthesis, and recognition; translation and translation aids; natural language interfaces; and theoretical and applications papers of every kind. REQUIREMENTS: Papers should describe unique work that has not been submitted elsewhere; they should emphasize completed work rather than intended work; and they should indicate clearly the state of completion of the reported results. Authors should send eight copies of an extended abstract up to eight pages long (single-spaced if desired) to: Alan W. Biermann ACL86 Program Chair Department of Computer Science Duke University Durham, NC 27706, USA [919:684-3048; awb%duke@csnet-relay] SCHEDULE: Papers are due by 6 January 1986 . Authors will be notified of acceptance by 25 February. Camera-ready copies of final papers prepared on model paper must be received by 18 April along with a signed copyright release statement. OTHER ACTIVITIES: The meeting will include a program of tutorials and a variety of exhibits and demonstrations. Anyone wishing to arrange an exhibit or present a demonstration should send a brief description to Alan Biermann along with a specification of physical requirements: space, power, telephone connections, tables, etc. CONFERENCE INFORMATION: Local arrangements are being handled by Kathy McKeown and Cecile Paris, Department of Computer Science, Columbia University, New York, NY 10027; 212:280-8194 and 8125; mckeown and cecile @columbia-20.arpa. For other information on the conference and on the ACL more generally, contact Don Walker (ACL), Bell Communications Research, 445 South Street, MRE 2A379, Morristown, NJ 07960; 201:829-4312; walker@mouton.arpa or walker%mouton@csnet-relay or bellcore!walker@berkeley. Program Committee: Alan W. Biermann, Duke University Kenneth W. Church, AT&T Bell Laboratories Michael Dyer, University of California at Los Angeles Carole D. Hafner, Northeastern University George E. Heidorn, IBM T.J. Watson Research Center David D. McDonald, University of Massachusetts Fernando C.N. Pereira, SRI International Candace L. Sidner, BBN Laboratories John S. White, Siemens Communication Systems LSA SUMMER LINGUISTIC INSTITUTE: ACL-86 is scheduled just before the 53rd LSA Institute, which will be held at the Graduate School and University Center of the City University of New York from 23 June to 31 July. The 1986 Institute is the first to focus on computational linguistics. During the intervening week, a number of special courses will be held that should be of particular interest to computational linguists. For further information contact D. Terence Langendoen, CUNY Graduate Center, 33 W. 42nd Street, New York, NY 10036; 212:921-9061; tergc%cunyvm@wiscvm.arpa. ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Thu Nov 7 05:52:07 1985 Date: Thu, 7 Nov 85 05:52:02 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a003024; 6 Nov 85 23:49 EST Date: Wed 6 Nov 1985 21:09-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V3 #164 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 7 Nov 85 05:47 EST AIList Digest Thursday, 7 Nov 1985 Volume 3 : Issue 164 Today's Topics: Queries - IQ Test for AI & RSA Encryption & IEEE Software Special Issue, Correction - TI Satellite Seminar, Expert Systems - Statistics and Diagnosis, Logic - Abductive Inference, Poetry - Colourless Green Ideas ---------------------------------------------------------------------- Date: Tue 5 Nov 85 09:37:20-PST From: Rene Bach Subject: IQ test for AI. I don't remember anyone suggesting to test AI programs with the traditional human IQ test. Is there any validity to this idea ? I understand that it would be very hard for a computer program to reason about the shape and graphics questions (given as is, i.e. just given the actual picture of the question). I also understand the IQ test are somewhat controversial, but they nevertheless provide a metric for a lot of human reasoning abilities which certainly require multiple facets of intelligence. Has any one tried to write such a computer program ? Rene Bach (Bach@score) ------------------------------ Date: Wed, 6 Nov 85 20:22:21 EST From: "Keith F. Lynch" Subject: RSA encryption Date: Tue 5 Nov 85 15:21:35-CST From: David Throop ... I will also discuss some applications of the system, including its use in finding security flaws in the formal specifications of computer software, its proof of the invertibility of the RSA public key encryption algorithm ... Has RSA been broken? I thought it was NP complete. Can you give a reference? ...Keith ------------------------------ Date: Tuesday, 5 November 1985 22:12:21 EST From: Duvvuru.Sriram@cive.ri.cmu.edu Subject: Abstracts for IEEE Software Special Issue All summaries that have been submitted to the IEEE Software Special Issue on KBES for Engineering (March 1986) that have exceeded the 3 page (double-spaced) limit have been edited to comply with the guidelines. If this is not acceptable, please send me mail. Sriram ------------------------------ Date: Wed 6 Nov 85 18:30:51-EST From: "Randall Davis" Subject: Correction - TI Satellite Seminar Concerning the recent item in AI List which said in part: From: leff%smu.csnet@CSNET-RELAY.ARPA Subject: IEEE Seminar on AI (SMU) The following is the program for the Special Event of the Dallas IEEE Computer Society and Dallas Section of the ACM Artificial Intelligence Satellite Symposium Knowledge-Based Systems and Their Applications presented by Texas Instruments Incorporated Date: Wednesday, November 13, 1985 8:30 am - 4:00 pm Place: Infomart, 1950 Stemmons Freeway Room 7011 TI has advertised this event widely, but there has still been some confusion about it. It is an educational program conceived of and created by TI, and is being broadcast via satellite to 23 different pre-selected sites around the country; tickets are necessary only to assure seating; admission is free. It is also, by design, available to anyone who has the equipment to pick up the satellite signal (TI is offering technical information on satellite reception at 214-995-4076). Current estimates are that perhaps 400 additional sites around the country will be doing so. >From the IEEE/ACM notice above it appears that they have organized themselves around this event, as numerous other organizations have. Please note, though, that the broadcast is available nationwide to anyone interested. ------------------------------ Date: Wed, 6 Nov 85 22:18:15 est From: Dana S. Nau Subject: Statistics and Diagnosis >From: Stuart Crawford > > I am interested in obtaining pointers to recent references regarding > the known pros and cons of using pure statistical approaches to > medical diagnosis (such as the use of classification and regression The following articles are relevant to your request. I think one or both of them may have appeared somewhere by now; my "refer" file is out of date. For more information, write to Dr. James Reggia Computer Science Department University of Maryland College Park, MD 20742 %A C. Ramsey %A J. A. Reggia %A D. S. Nau %A A. Ferrentino %T A Comparative Analysis of Methods for Expert Systems %R submitted for publication %A J. A. Reggia %A B. T. Perricone %T Answer Justification in Medical Decision Support Systems Based on Bayesian Classification %R Submitted for publication %D 1983 %C College Park, MD %I University of Maryland Dana S. Nau (dsn@rochester) from U. of Maryland, on sabbatical at U. of Rochester ------------------------------ Date: Tue 5 Nov 85 04:48:39-EST From: "Sidney Markowitz" Subject: abductive inference The following is in reply to the question about abduction, in the sense of abductive inference, and its relation to AI. This is a portion of a message extracted from the old philosophy-of-science mailing list [and reprinted by permission of the author -- KIL]. It is a bit long, as it contains both definitional material and opinions as to its relevance to AI. I have deleted much extraneous material: Date: 26 Jan 1983 0128-PST From: ISAACSON at USC-ISI [ ... ] I can't speak for the Intuitionist position. I do prefer the second inference-chain you propose, though. It is a *weaker* inference, but, it is potentially generative. It is close to a form of inference which is sometimes called "abduction", favored by Charles Sanders Peirce. This sort of inference is held to underlie hypothesis-formation. Abductive inference can be stated as follows: The surprising fact, C, is observed; But if A were true, C would be a matter of course; Hence, there is reason to SUSPECT that A is true. The *tentative* (or hypothetical) nature of the conclusions in the two latter cases above is what makes these inferences potentially GENERATIVE, in the sense of knowledge-generation; whereas the deductive template inherent in [ an example of deduction ] makes it a barren exercise in classical Aristotelian logic, with no generative power beyond the (pre-programmed, as it were) syllogitic-chain itself. [ ... ] I want to use that fairly long introduction to move into a comparison between traditional deductive inference [ ... ] and other types of inferences I consider to be knowledge-generators, or "epistemogenic inferential processes." [ ... discussion of deductive inference ... ] For example, All birds are five-legged mammals; Fred is a bird; Then Fred is a five-legged mammal. Or, All Blacks are white; Fred is a Black; Then Fred is white. All of the stuff you see above is entirely Kosher, viewed from within deductive logic proper, and I have no quarrel with that. BUT WHY TAKE EXTRA PRIDE IN PROMOTING THIS KIND OF BARREN SYLLOGISTIC ACTIVITY WITHIN AI IS BEYOND ME! What we sorely need are inferential processes that are capable of generating new knowledge [through computational means!]. We need to develop a broad class of "Epistemogenic Processes". I think it includes a family of inferences that can generate explanatory hypotheses, and therefore, underlie theory-formation. Peirce, the so-called "Father of Pragmatism" (he actually called his creation "Pragmaticism"), devoted much of his massive life-work to elaborating a type of inference he called "abduction". In his view, when contrasted with "induction" and "deduction", it is the only truly creative mode of inference. It is THE epistemogenic agent. The sort that yields new explanatory hypotheses in scientific inquiry. As a corollary he developed a theory of the "Economy of Research", an obscure and understudied, yet incredibly rich, research methodology. I do agree with Minsky that we ought to be courageous and resourceful enough to be willing to break new ground, without too many hangups about "old stuff". Yet, I think that we have an incredibly fertile resouce in Peirce, and we owe it to our enterprise to COHERE what we are trying to do with what he has already accomplished. ------------------------------ Date: Wed 30 Oct 85 23:02:50-PST From: Paul Roberts Subject: Colourless Green Ideas [Excerpted from the Stanford bboard by Laws@SRI-AI.] For the Literary competition set on Christmas Eve you were asked to compose not more than 100 words of prose, or 14 lines of verse, in which a sentence described as grammatically acceptable but without meaning did, in the event, become meaningful. The sentence, devised by Noam Chomsky, was: colourless green ideas sleep furiously. [...] Competitors rose to this challenge good-humouredly and in force.... It can only be the thought of verdure to come, which prompts us in the autumn to buy these dormant white lumps of vegetable matter covered by a brown papery skin, and lovingly to plant them and care for them. It is a marvel to me that under this cover they are labouring unseen at such a rate within to give us the sudden awesome beauty of spring flowering bulbs. While winter reigns the earth reposes but these colourless green ideas sleep furiously. C. M. Street Behold the pent-up power of the winter tree; Leafless it stands, in lifeless slumber. Yet its very resting is revival and renewal: Inside the dark gnarled world of trunk and roots, Cradled in the chemistry of cell and sap, Colourless green ideas sleep furiously In deep and dedicated doormancy, Concentrating, conserving, constructing: Knowing, by some ancient quantum law Of chlorophyll and sun That come the sudden surge of spring, Dreams become reality, and ideas action. Bryan O. Wright Let us think on them, the Twelve Makers Of myths, trailblazing quakers Scourging earthshakers Colourless green ideas sleep furiously Before their chrysalides open curiously Anarchy burgeons spuriously Order raises new seedlings in the world By word and gun upheld The scarlet banner is unfurled The New Country appears Man loosens his fears The New Dawn nears Recollect our first fathers The good society in momentum gathers. ("recently discovered sonnet by Alexander Blok") translated by Edward Black [...] (and the winner:) (got 50 lbs.) Thus Adam's Eden-plot in far-off time: Colour-rampant fowers, trees a myriad green; Helped by God-bless'd wind and temp'rate clime. The path to primate knowledge unforseen, He sleeps in peace at eve with Eve. One apple later, he looks curiously At the gardens of dichromates, in whom colourless green ideas sleep furiously then rage for birth each morning, until doom Brings rainbows they at last perceive. D. A. H. Byatt ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Thu Nov 7 23:44:28 1985 Date: Thu, 7 Nov 85 23:44:24 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a008889; 7 Nov 85 13:17 EST Date: Thu 7 Nov 1985 09:33-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V3 #165 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Thu, 7 Nov 85 23:30 EST AIList Digest Thursday, 7 Nov 1985 Volume 3 : Issue 165 Today's Topics: Queries - Billy Salter's Thesis & Intensional Contents & Validation of Knowledge Based Systems, Applications - Reasoning About Shape and Graphics, Cryptography - RSA Encryption, AI Tools - Typed Languages and Lisp, Opinion - AAAI and AI Hype, Expert Systems - Stock Market Prediction Hype ---------------------------------------------------------------------- Date: Thu, 7 Nov 85 12:01:32 -0100 From: Rolf Pfeifer Subject: Billy Salter's Thesis I am looking for Billy Salter's Thesis on Subjective Theories of Economics (or something like that). He got his PhD from Yale approximately in 83. If anyone has a copy available or knows of Billy's whereabouts (possibly BBN?) please let me know (including netmail address). Thanks. --Rolf Pfeifer ------------------------------ Date: 7 Nov 85 13:40:15 GMT From: Bob Stine Subject: Intensional Contents 'whois zalta' fails, so I'll direct this request to the net. Can anyone give me pointers to Edward Zalta's work on the "intentional contents of mental states about fictions?" In AIList Digest V3 #163, it was announced that Dr. Zalta would be giving a lecture in Berkeley on that topic. Would one use Golden Hill LISP to code assertion's about Meinong's Golden Mountain? - Bob Stine ------------------------------ Date: Thu, 7 Nov 85 16:03 From: at GEC Research Subject: Validation of Knowledge Based Systems I am interested in carrying out some work on the validation of real-time Knowledge Based Systems and would appreciate some help in getting started. If anyone has any information they can send me on this topic, in particular, significant references or current projects it would be a great help. Thanks. Please send responses to: Kevin Poulter (YE15%uk.co.mrca@uk.ac.ucl.cs) GEC Research Ltd Marconi Research Centre West Hanningfield Road Great Baddow Chelmsford Essex CM2 8HN United Kingdom ------------------------------ From: CONNOLLY CHRISTOPHER IAN Subject: reasoning about shape & graphics Bach@score recently asked if anyone has tried to write a computer program to reason about shape or graphics, in the context of IQ tests: I don't know if this is really what you were thinking of, but Gelernter did some work on this at IBM back in 1958-1960. His aim was to use diagrams to aid in proving geometrical theorems. If I remember correctly, he succeeded in that his program was able to prove several theorems in geometry. The work is reported in one of the issues of the IBM research journal in 1960. Gelernter has since escaped to Biophysics as an occupation (!). ------------------------------ Date: Thu 7 Nov 85 10:14:32-CST From: David Throop Subject: RE: RSA encryption > ... I will also discuss (the Boyer Moore theorem prover's) proof of the invertiblity of the RSA public key encryption algorithm ... > Has RSA been broken? I thought it was NP complete. Can you give a reference? ====================== No, it has not been broken. But it has been proven invertible. That is, consider an (input) message encrypted with a public key and an (output) string produced by decrypting the (encrypted) message using a private key. The Boyer Moore theorem prover has rigorously proven that, given any possible input, the output will be identical to it. As Dr Moore said in last night's talk, what you would like to prove is that no one can solve the encrypted message with only the public key. But that's not provable, because the problem IS solvable. Its just very hard. NP hard, in fact. What you would really like to prove are some things about the properties of NP problems. You will certainly hear about it if anybody around here proves any such thing. ------------------------------ Date: 05 Nov 85 2115 PST From: John Craig Subject: Typed languages and Lisp [Forwarded from the Stanford bboard by Laws@SRI-AI. This is part of an ongoing discussion of strong typing.] Chris Goad (Stanford CS grad) developed a language originally called SIL, now called RISE which is essentially a typing system added to Lisp, but with a no-type type so that one can get around typing as one desires. The main reasons for adding typing are: 1) faster code development (type checker finds bugs) 2) the compiler can use type information to generate more efficient object code (for example, less or no garbage collection pauses when running compiled code) RISE is in use at Silma Inc (Los Altos) and forms the user interface for their product, RoboCam. It seems to me like you get the best of lisp and typed worlds, and efficient code generated also. Its pretty fun, too. John ------------------------------ Date: Tue, 5 Nov 85 12:20:02 MST From: shebs@utah-cs.ARPA (Stanley Shebs) Subject: Re: AI Hype Just got my announcement of AAAI-86 the other day. Big poster, bright colors (thankfully not as garish as certain other announcements), lots of pictures, little information. By contrast, announcements for ACM conferences stick to the salient details and do it on the standard size sheet of paper, although they do get radical for the national conference and put in a little icon... If AAAI can't seem to escape the need for hype, why should anyone assume there's actually something real in AI? stan shebs ------------------------------ Date: Sat 2 Nov 85 08:20:07-PST From: the tty of Geoffrey S. Goodfellow Subject: The Ultimate Hype: Mail Order "AI"! In the interest of keeping the AI Research Community-At-Large informed of just how out of hand AI Hype is, I offer, for your consideration the following travesty (if you will) from the latest edition of the JS&A "Products that THINK #16" yuppy catalog. Both pieces are rather lengthy, but in order to appreciate the fullness of egregiousness they espouse to our public about "ARTIFICIAL INTELLIGENCE" (emphasis added here as well as below), I have included them in toto. [I have taken the liberty of cutting Geoff's 17K-char message by about half. -- KIL] -0- -0- -0- -0- Back Cover: (full page) MARKET VICTIM I used to be a sucker for the stock market. I'd get a tip from a friend, I'd invest on a hunch or a broker would offer me some advice and off I'd go. On the average, I never made any money. So I dropped out of the market until I met Hal and Bill. Hal was one of the most astute market advisors I have ever met. At first, I didn't trust Hal. Despite his good credentials and a well-documented three-year track record, was still suspicious. [...] It took the incredible accuracy of Hal and Bill to finally get me back into the market. Hal had selected 14 stocks out of the 1500 on the New York Stock Exchange that he felt would really take off. A few I had never heard of. Bill advised me to sell some of my stocks in the nick of time so the net results were 13 out of 14 winners with a few stocks giving me a return of greater than 30% in just a four-month period. [...] Hal and Bill are ARTIFICIAL INTELLIGENCE computer programs--two independent programs that perform what no other advisor or market expert has been able to perform in the history of investing. I can now sit at my Visual Computer (see page 10), insert both Hal and Bill into my disk drive, and be told by this intelligent program what stocks on the New York Stock Exchange to buy and when to sell them. [Runs on IBM compatibles with 256K memory.] Hal gives his advice on the long range--usually his advice is good between a three to nine month period after he makes his decisions. Once that happens, I then follow Bill very closely because he tells me when to close that position. If I want to know when to sell, I let Bill tell me by first downloading the last 36 days of stock market activity from the Dow Jones News Retrieval service with the help of my telephone modem. The program does this automatically--all I do is press a few buttons. Then Bill takes that information and advises me. Between both programs the accuracy will amaze you. But that's not all. [...] The entire program is called Halographix. [...] The program features all the tools you need to analyze evaluate and record your transactions. It has automatic log-on features that access the Dow Jones News Retrieval service and download information on any stock you've selected for your portfolio. Each disk is valid for only four months. You pay only $199 for the program (that's roughly $25 for each of the two independent programs per month), and purchasing the disk entitles you to renew the disk every four months thereafter for the same $199 per disk. Obviously, if the program doesn't make you plenty of money, you don't renew. But for the few who have participated, practically all of them have enthusiastically remained with us. Our guarantee of satisfaction is very compelling. If, after the four months, you are not satisfied with the program or have not seen it pay for itself many times over, please return it and get a full refund of your $199 investment. You can't lose. To order, simply send your check, money order or credit card number to me personally, Joseph Sugarman, President, JS&A Group, Inc., One JS&A Plaza, Northbrook, Illinois 60062. If for some reason, we sell out our subscription for this issue, I will promptly refund your money and keep you on our list for our next release. But I urge you to act quickly. I also urge you to read the article on page 20 entitled "Million Dollar Phone." In that article, I give you an idea of the nature of ARTIFICIAL INTELLIGENCE and the philosophy of how the program works to accurately predict market movement. ARTIFICIAL INTELLIGENCE, the tremendous number-crunching power of your own personal computer and the skills of a computer programmer have produced a personal friend for the small investor. You can now compete with the big institutions on Wall Street with advice more accurate, personal and more astute than their huge network of advisors and computers. Join with me in this novel program and enjoy the wealth of information Halographix provides to guide you to success in your investment activity with the New York Stock Exchange stocks. Halographix (6082N 6.00)..........$199 -0- -0- -0- -0- Page 20/21: MILLION DOLLAR TELEPHONES What you are about to read may sound like a get-rich-quick scheme. And indeed it may be. [...] For the past year, I have been working with a computer genius who has developed a program that, when run on a powerful computer, can predict the movement of the Dow Jones Industrial Average. In our first year of testing this program, our computer generated ten signals. Nine were accurate with the Dow making the forecasted moves within twelve trading days. To my knowledge, there has not been an advisor, a market expert or another computer program that has been as accurate. Two years ago a new form of option trading was introduced to investors. Called OEX options, the concept gave investors a way to speculate on the movement of the Standard and Poor's market index. This market index is very easy to follow because it runs almost parallel to the Dow Jones Industrial Average. If you were right on the movement of the overall market, you could possibly double or triple your investment within a few short weeks. If you were wrong, all you could lost was your initial investment and no more. Another popular feature of the option program is the tax treatment. 60% of all your gains are treated as capital gains and 40% as ordinary income or about a 32% effective tax rate. [...] I have formed a club called, "Dial-An-Option," which will offer information on the Dow Jones Industrial Average. First you sign up with JS&A as a member, which costs you only $25. Then, at no extra charge, we provide you with information on one major market signal. Next, sit back, relax and watch what happens to the Dow over the next two weeks following the signal. [....] Our advice will be sent via an overnight delivery service so you'll get the first market signal well enough in advance to act quickly. The signal must make money for you and double or triple your investment or you can drop out of our Dial-An-Option club and we'll refund your entire membership fee. Or don't take the chance and just follow the market and see what you would have made had you invested on our advice. [...] Why does our computer program work so effectively? What does it do to predict the Dow so accurately? If you'll take just a moment, I will explain. ARTIFICIAL INTELLIGENCE The big buzzword in software today is, "ARTIFICIAL INTELLIGENCE," or a special level of reasoning applied to a computer program. The reasoning involves looking at hundreds of factors at once and then comparing those factors to previously programmed responses and then drawing a conclusion. Our brains operate on a similar basis. We are programmed with patterns based on previous experiences or programming. When we have to make a decision based on new data, we lay that new data onto previous data and then draw our conclusions. With the number-crunching capabilities of a powerful computer and an ARTIFICIAL INTELLIGENCE software program, our computer can sense market movement, movement within an industry group and even movement (accumulation or distribution) within a specific stock. If then computes the major 30-day trend, and interprets the fluctuations in trading and valuation levels to determine the one-to-five day pattern within the major trend. [...] Our computer has the ability to first sense the changes and trends in the market before they become obvious and then, with the minimal amount of information, develop a comparable pattern and make a conclusion--an ideal system for predicting the movement of the Dow Jones averages. [...] No program or system can guarantee future profit or success. If we are not correct on any of our calls (remember, so far we've been right on nine out of ten calls), then you'll receive a refund of the fee--something that no other advisory service that we know of provides. Obviously we can't guarantee any loss that may result from our advice but we urge you never to invest more than you're willing to lose. [...] Thanks to a powerful computer, ARTIFICIAL INTELLIGENCE, and a brilliant program, we can no do this analytical work on a computer with greater accuracy than even the most astute analyst or the most powerful institution. I urge you to join our club and experience the power of our computer and its new program designed to earn you bigger returns than you've ever dreamt possible--and all in a very short period of time. We've given the small investor the tool to compete with the big institutions and enjoy the potential profits offered every day in the market place. Join our club, today. Dial-An-Option Membership (6078N)..........$25 OEX Options Video Tape (6079N 2.00).........39 OEX Options Audio Tape (6080N 2.00).........14 Video and Audio Tape (6081N 2.00)...........49 Note: Free OEX booklet sent to each club member or tape purchaser. (Dian-An-Option is a registered trademark of JS&A Group.) ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Sat Nov 9 02:37:00 1985 Date: Sat, 9 Nov 85 02:36:56 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a009314; 8 Nov 85 13:15 EST Date: Fri 8 Nov 1985 09:36-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V3 #166 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Sat, 9 Nov 85 02:25 EST AIList Digest Friday, 8 Nov 1985 Volume 3 : Issue 166 Today's Topics: Queries - Xerox 1186 Comments, Logic & Probability - Abductive Inference, Linguistics - New Reports (CSLI), Literature - AI at Past Conferences, Opinion - AI Hype, Humor - New Mailing List for AI Hype ---------------------------------------------------------------------- Date: Thu 7 Nov 85 09:42:55-PST From: Ted Markowitz Subject: Comments on the Xerox 1186... With the idea of setting up an in-house AI lab for informal testing of expert systems, teaching, etc., I've been looking at some of the specialized AI hardware available. I've read the discussions so far on the various machines available, but would like some more opinions on the new Xerox 1186 class processor. Needless to say price is something of a factor in my choice and this machine seems the cheapest with most of the functionality that I seek. Any thoughts? Pros and cons? Please post answers to the list for redistribution. Thanks muchly. --ted ------------------------------ Date: Thu, 7 Nov 85 16:01 EST From: Mukhop Subject: Abductive Inference > From: "Sidney Markowitz" >> Date: 26 Jan 1983 0128-PST >> From: ISAACSON at USC-ISI >> Abductive inference can be stated as follows: >> The surprising fact, C, is observed; >> But if A were true, C would be a matter of course; >> Hence, there is reason to SUSPECT that A is true. Assume the following: Of the subjects with mono(nucleosis), 100% show a positive result for the mono test. Of the subjects without mono, only 1% show a positive result. Given that the test result for a subject (of unknown condition) is positive, what is the likelihood of the subject having mono. Of the total population, one out of every 10000 people is assumed to have mono. "The surprising fact, C, is observed;" C is the fact that the subject tested positive. The fact is surprising because the probability is just over 1%. "But if A were true, C would be a matter of course;" A is the premise that the subject has mono (A => C). "Hence, there is reason to SUSPECT that A is true." Therefore, there is reason to suspect that the subject has mono. The result appears to be reasonable. However, the probability of ~A given C is .991, indicating: If the result of the test is positive, then the subject probably does not have mono. More appropriately: Despite the fact that the test is positive, the subject probably does not have mono. I realize that the point made by Isaacson is the generative nature of abductive inference (used to generate plausible hypotheses for testing). This counter-example is in the same vein as some recent contributions to this list regarding modus ponens and the presidential triangle. Uttam Mukhopadhyay Comp. Sci. Dept. GM Research Labs Warren, MI 48090-9055 Phone: (313)575-2105 ------------------------------ Date: Thu 7 Nov 85 16:41:45-PST From: Emma Pease Subject: New Linguistics Reports (CSLI) [Excerpted from the CSLI Newsletter by Laws@SRI-AI.] NEW CSLI REPORTS Report No. CSLI-85-37, ``On the Coherence and Structure of Discourse'' by Jerry R. Hobbs, and Report No. CSLI-85-38, ``The Coherence of Incoherent Discourse'' by Jerry R. Hobbs and Michael Agar, have just been published. These reports may be obtained by writing to David Brown, CSLI, Ventura Hall, Stanford, CA 94305 or Brown@SU-CSLI. ------------------------------ Date: 2 Nov 1985 18:46-CST From: leff%smu.csnet@CSNET-RELAY.ARPA Subject: AI at Past Conferences Eighth International Conference on Software Engineering Augst 28-30 1985 Imperial College, London On executable models for rule-based prototyping S. Lee USA Session 8A Knowledge Based Apporoaches Automating tuning of multi-task program for real time embeded system T. Shimizu K. Sakamura Prompter: A Knowledge based support tool for code understanding K. Fukunaga, Japan The Analyst - A Workstation for Design and Analysis M. Stephens, K. Whitehead Session 9 Discussion Software engineering- The role of logic and AI in the software enterprise ------------------------------ Date: 08 Nov 85 17:24:32 +1100 (Fri) From: munnari!mungunni.oz!lee@seismo.CSS.GOV Subject: Re: Mail Order "AI"! This reminds me of an episode of Minder (BBC TV series), in which Arthur (a con man) set up a racing tips business, on an "only pay if you win" basis. They tipped all horses in each race, so some of the clients won and paid up. A guaranteed income! I wonder if the "AI" program for predicting the stock market uses this heuristic. lee [Another possibility is that many subscribers may stay in for one or two rounds before becoming convinced that they've been had. On the other hand, the JS&A offer could be based on a legitimate formula (not a breakthrough, I assume, but "chartists" have been known to do quite with their formulas) that has recently donned AI garb. I have some of my savings in a "timing service" that uses a similar approach. -- KIL] ------------------------------ Date: Fri 8 Nov 85 10:48:37-CST From: CMP.MGREEN@R20.UTEXAS.EDU Subject: ai hype You ain't heard nothin' yet. PCWeek, "The national newspaper of IBM standard microcomputing", ran a four part review during October of a fourthcoming book by Mickey Williamson titled "Artificial Intelligence for Microcomputers: A Business Design". Although the book appears to give a generally well balanced view of the current capabilities of AI it occasionally misses the mark. >From the start it is apparent that the author knows very little about the subject and was forced to rely on facts and opinions supplied by others without being able to provide some kind of sanity check. Case in point: He quotes Barbra Wallace of KDS corporation who "worked on a molecular memory project in the 1950s" as predicting that within 10 years we will see a microcomputer with capabilities comparable to 2001's HAL, "but she stops short of predicting what it will cost". I guess it's time to offer Stanley Kubric a position with our group, if he did it once maybe he can do it again. Cheers -- Mike Green ------------------------------ Date: Thu 7 Nov 85 16:14:27-CST From: CMP.BARC@R20.UTEXAS.EDU Subject: New Mailing List for AI Hype Due to the volume of extraordinary AI claims, we have established a new mailing list for promoting and isolating AI hype. This list is designed for persons who wish to lie, exaggerate or otherwise misinform regarding the potential and/or accomplishments of AI. Inflated discussions about AI features, enhancements, performance, support, and other topics of interest to venture capitalists are welcome. Discussions about bugs, development problems, anything approaching realism or legitimacy, or about AI hype itself are specifically discouraged, as they are more appropriate for AIList or other mailing lists. These restrictions will have to be left as a matter of "honor among thieves", since the list will NOT be moderated, but will act as a mail "reflector" - ie., any message sent to the list will be rebroadcast to everyone on the list. The list will be maintained at Smart Expertelligeneric Logical Infer- enceware and Teknowledgecraft Inc., using a proposed neural emulation network of 3,140,000 parallel Lisp machines (Our current configuration is an Explorer prototype and a Commodore 64, running CP/M.). SELIT is connected to the ARPAnet (both military and educational) and uucp, and has gateways to CSNET, BITNET and Compuserve. Thus, it should be pos- sible to access every business, school or home in the U.S. and many in Europe. Therefore we hope to be able to reach every gullible element of the computing community. Of course, we intend to charge an exorbitant fee for inclusion in the mailing list, until our subscribers figure out that the list has nothing to offer beyond the current, conventional lists and bulletin boards. To add your name to the list, change or delete a name or have other administrative requests serviced, send mail to: ARPAnet: AIHype-REQUEST@SELIT.ARPA uucp: ...ihnp3.14!selit!aihype-request To post a submission to the list, send mail to: ARPAnet: AIHype@SELIT.ARPA uucp: ...ihnp3.14!selit!aihype Please, do NOT bother the entire list with a request to have your name added or deleted! The general discussion should be bothersome enough. Feel free to rebroadcast this announcement to anyone who might be interested. Dallas Webster ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Wed Nov 13 02:47:26 1985 Date: Wed, 13 Nov 85 02:47:21 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a002961; 10 Nov 85 23:12 EST Date: Sun 10 Nov 1985 19:56-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V3 #167 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Wed, 13 Nov 85 02:33 EST AIList Digest Monday, 11 Nov 1985 Volume 3 : Issue 167 Today's Topics: Seminars - TI AI Symposium Sites & Model Theory for Knowledge and Belief (SRI) & Example-Based Reasoning (NU) & Knowledge Representation (UCB) & Partial Truth Conditions and Their Logics (CSLI) & Automatic Generation of Graphical Presentations (CSLI) & CommonLoops (MIT) & Minimal Entailment (UPenn) & Alternatives to Concurrent Prolog (MIT), Conference - Eastern Simulation Conference ---------------------------------------------------------------------- Date: Fri, 8 Nov 85 00:09 EST From: Tim Finin Subject: TI AI Symposium Sites Two sites in our area that will be providing the TI AI Symposium are: University of Pennsylvania Harrison Auditorium (in the Univ. Museum, 33rd & Spruce) Philadelphia, PA contact: Tim Finin, TIM@UPenn (215-386-1749) lots of room - all are welcome - no invitation/RSVP needed U.S. Army Communications /Automatic Data Processing Center Watters Hall Fort Monmouth, NJ contact: Ms. Van dyke (201-544-2929) arrive early to assure seating. It starts at 9:15 (EST) on Wednesday, November 13th. ------------------------------ Date: Wed 6 Nov 85 17:41:42-PST From: LANSKY@SRI-AI.ARPA Subject: Seminar - Model Theory for Knowledge and Belief (SRI) MODEL THEORY FOR KNOWLEDGE AND BELIEF Moshe Vardi IBM San Jose 11:00 AM, MONDAY, November 11 SRI International, Building E, Room EJ228 (new conference room) Recently, there has been a surge of interest in the modal logic of knowledge and belief, which has applications in many area of computer science. The standard semantics for modal logic is Kripke semantics. In this semantics, possible worlds and the possibility relation are both primitive notions. This has both technical and conceptual shortcomings. >From a technical point of view, the mathematics associated with Kripke semantics is often quite complicated. From a conceptual point of view, it is not clear how to use Kripke structures to model knowledge and belief, where one wants a clearer understanding of the notions that are taken as primitive in Kripke semantics. We introduce modal structures as models for modal logic. We use the idea of possible worlds, but by directly describing the internal semantics of each possible world. It is much easier to study the standard logical questions, such as completeness, decidability, and compactness, using modal structures. Furthermore, modal structures offer a much more intuitive approach to modelling knowledge and belief. As an application, we present a semantic model for knowledge with the following properties: (1) Knowledge is necessarily correct (2) agents are logically omniscient, i.e., they know all the consequences of their knowledge (3) agents are positively introspective, i.e., they are aware of their knowledge, but not negatively introspective, i.e., they may not be aware of their ignorance. We argue that this is the appropriate model for implicit knowledge. We investigate the properties of the model, and use it to formalize notions such as "to know more" and "all that is known is". ------------------------------ Date: Wed, 6 Nov 85 14:53 EDT From: Carole D Hafner Subject: Seminar - Example-Based Reasoning (NU) College of Computer Science Colloquium Northeastern University, Boston, MA Example-Based Reasoning Prof. Edwina Rissland Dept. of Computer and Information Science University of Massachusetts, Amherst, MA In this talk, I shall discuss example-based reasoning, particularly in the contexts of assisting in the preparation of legal arguments and offering on-line explanations. In the case of legal argumentation, I discuss how hypotheticals serve a central role in analyzing the issues in a case and describe a program, called HYPO, which generates legal hypotheticals, and an environment, called COUNSELOR, which provides support for legal reasoning and other strategic tasks, like resource management. I'll briefly describe our current work on on-line assistance and how we are trying to make it more intelligent by embedding custom-tailored examples in the explanations. I'll also discuss some general issues about examples such as their generation, structure and importance in reasoning, especially in the domains of mathematics and the law. Date: Wednesday, Nov. 13, 1985 Time: 12:00 noon Place: To be announced (contact hafner@northeastern or call the department office at 437-2462). ------------------------------ Date: Thu, 7 Nov 85 17:29:20 PST From: admin%cogsci@BERKELEY.EDU (Cognitive Science Program) Subject: Seminar - Knowledge Representation (UCB) BERKELEY COGNITIVE SCIENCE PROGRAM Cognitive Science Seminar - IDS 237A Tuesday, November 12, 11:00 - 12:30 240 Bechtel Engineering Center Discussion: 12:30 - 1:30 in 200 Building T-4 ``Knowledge Representation and a Theory of Meaning'' Robert Wilensky Computer Science Division, U.C.B. Knowledge representation is central to most Artificial Intelli- gence endeavors. However, most knowledge representation schemes are incomplete in a number of ways. In particular, their coverage is inadequate, and they do not capture signifi- cant aspects of meanings. Many do not even adhere to basic criteria of well-formedness for a meaning representation. KODIAK is a theory of knowledge representation developed at Berkeley that attempts to address some of these deficiencies. KODIAK incorporates representational ideas that have emerged from different schools of thought, in particular from work in semantic networks, frames, Conceptual Dependency, and frame semantics. In particular, KODIAK eliminates the frame/slot distinction found in frame-based languages (alternatively, case/slot distinction found in semantic network-based systems). In its place KOKIAK introduces a new notion called the absolute/aspectual distinction. In addition, the theory sup- ports ``non-literal'' representations, namely, those motivated by metaphoric and metonymic considerations. Using these dev- ices, the theory allows for the representation of some ideas that in the past have only been represented procedurally, informally, or not at all. KODIAK is being used to represent both linguistic and concep- tual structures. When applied to the representation of linguistic knowledge, a new framework for talking about meaning emerges. Five aspects of meaning have been identified. These appear to be useful in describing processing theories of natural language use. ------------------------------ Date: Thu 7 Nov 85 16:41:45-PST From: Emma Pease Subject: Seminar - Partial Truth Conditions and Their Logics (CSLI) [Excerpted from the CSLI Newsletter by Laws@SRI-AI.] CSLI ACTIVITIES FOR *NEXT* THURSDAY, November 14, 1985 4:15 p.m. CSLI Colloquium Redwood Hall Partial Truth Conditions and Their Logics Room G-19 Hans Kamp, University of Texas Partial Truth Definitions and their Logics Hans Kamp Until recently truth definitions for formal and natural languages were, with some few exceptions, total (in the sense of specifying w.r.t. any model a truth value for each sentence of the language under consideration). But during the past decade partial truth definitions have become increasingly common both within symbolic logic and in formal semantics. The motives for adopting partial truth definitions vary considerably. I will focus on three issues that have led to the formulation of such definitions: i) vagueness; ii) the semantic paradoxes; and iii) verification by partial information structures (a concept that has inspired both situation semantics and recent work on the semantics of data structures). I will discuss and compare some of the partial semantics that have been developed in attempts to come to terms with these issues, looking in particular at the question what logics are generated by the resulting semantic theories. I will argue that the relation between semantics and logic is less straightforward when the truth definition is partial than when it is total, and consequently that the notion of logical validity becomes much more delicate and equivocal once total semantics is abandoned in favor of some partial alternative. ------------------------------ Date: Thu 7 Nov 85 16:41:45-PST From: Emma Pease Subject: Seminar - Automatic Generation of Graphical Presentations (CSLI) [Excerpted from the CSLI Newsletter by Laws@SRI-AI.] PIXELS AND PREDICATES Automatic Generation of Graphical Presentations Jock Mackinlay CSLI trailers, 1:00 p.m., Wednesday, November 13, 1985 The goal of my thesis research is to develop an application- independent presentation tool that automatically generates appropriate graphical presentations of information such as charts, maps, and network diagrams. A presentation tool can be used to build effective user interfaces because it exploits the structure of the information and the capabilities of the output device to generate appropriate presentations. Application designers need not be graphical presentation experts to ensure that their user interfaces use graphical languages correctly and effectively. The research has two parts: a formal analysis of graphical languages for presentation and a prototype presentation tool based on the formal analysis. The formal analysis uses syntactic and semantic descriptions of graphical languages to develop criteria for evaluating graphical presentations. There are two major classes of criteria: expressiveness and effectiveness. The expressiveness criteria are theorems that identify when a set of facts is or is not expressible in a language. The effectiveness criteria are conjectures (rather than theorems) about the relative difficulty of the perceptual tasks associated with the interpretation of graphical languages. Sufficiently expressive languages are ordered by the difficulty of their associated perceptual tasks. The prototype presentation tool, called APT (A Presentation Tool), uses the criteria developed by formal analysis to search a space of graphical languages for an appropriate presentation. A novel feature of APT is its ability to generate its search space by composing sophisticated designs from a small set of fundamental graphical languages. The design portion of APT is a logic program based on the MRS representation system. ------------------------------ Date: Thu 7 Nov 85 15:32:51-EST From: "Mary E. Spollen" Subject: Seminar - CommonLoops (MIT) [Forwarded from the MIT bboard by SASW@MIT-MC.] CommonLoops Speaker: Gregor Kiczales Xerox Palo Alto Research Center Date: November 15, 1985, Friday Time: 2:15 refreshments 2:30 lecture Place: NE43-512A CommonLoops is a merger of Object Oriented Programming and Lisp. It has a unique combination of features: 1) No special syntax: Most attempts to add object-oriented programming to Lisp have resulted in special syntax for message sending. In CommonLoops, there is no syntactic difference between calling a function and "invoking a method." 2) Method Specification: In object oriented programming, methods are specified in terms of the class of the object being sent the message. One can think of this as specifying the type of one argument of the method. In CommonLoops, one can specify the type of any number of arguments to a method. 3) Type space: The "object" space is an extension of the normal Lisp type space, not a separate space as in Loops or Flavors. 4) Metaclasses: The implementation of a type (determined by the "metaclass") is independent of the type description. This allows tradeoffs between early binding and ease of exploratory programming. Host: Hal Abelson ------------------------------ Date: Sat, 9 Nov 85 00:53 EST From: Tim Finin Subject: Seminar - Minimal Entailment (UPenn) FUN WITH MODELS: MINIMAL ENTAILMENT AND NON-MONOTONIC REASONING David W. Etherington University of British Columbia (Currently at AT&T Bell Laboratories, Murray Hill, NJ.) 3:00pm December 3, 1985 216 Moore School, University of Pennsylvania, Philadelphia Circumstances commonly require that conclusions be drawn (conjectured) even though they are not strictly war- ranted by the available evidence. Various forms of minimal entailment have been suggested as ways of generating appropriate conjectures. Minimal entailment is a conse- quence relation in which those facts which hold in minimal models of a theory are considered to follow from that theory. Thus minimal entailment is less restrictive than the standard logical entailment relation, which strongly constrains what evidence may be taken as supporting a con- clusion. Different definitions of minimality of models yield different entailment relations. The talk will outline a variety of such relations. Domain, Predicate, and Formula Circumscription [McCarthy 1978, 1980, 1984] are syntactic formalisms intended to capture these relations. We examine each from a semantic viewpoint, in the hope of clarifying their respective capabilities and weaknesses. Results on the consistency, correctness, and adequacy of these formal- isms will be presented. While minimal entailment corresponds most directly to the Closed-World Assumption, that positive information not implicit in what is known can be assumed false, McCarthy and others have suggested applications of circumscription to more general default reasoning tasks. With this in mind, connections between minimal entailment and Reiter's Default Logic will be sketched, if time permits. In this connec- tion, we will consider positive and negative results due to Grosof and Imielinski, respectively. ------------------------------ Date: Sun, 10 Nov 85 00:56:01 EST From: "Steven A. Swernofsky" Subject: Seminar - Alternatives to Concurrent Prolog (MIT) Thursday 7, November 2: 15pm Room: NE43- 7th floor playroom BFCP and GHC - Alternatives to Concurrent Prolog Jacob Levy Department of Applied Mathematics Weizmann Institute of Science This talk will discuss some of the alternatives to Concurrent Prolog recently proposed. Each of these languages is designed to cover a large subset of Concurrent prolog, but to be much easier to implement. Flat Concurrent prolog (FCP) and Guarded Horn Clauses (GHC) will be described in detail. FCP, which has only And-parallelism, was developed at the Weizmann Institute as a viable subset of Concurrent Prolog. Its current implementation, in terms of a Warren Abstract Machine, will be described. The GHC language, designed by K. Ueda of ICOT, Japan, has OR-parallelism as well as And-parallelism, but instead has more limited synchronization primitives than Concurent Prolog. The second part of this talk will briefly describe my implementation of GHC. After the talk, a demo of FCP and Logix, its programming environment, will be given. Refreshments at 2:00pm HOSTS: Professors Gerald Jay Sussman and Henryk Jan Komorowski (Harvard) ------------------------------ Date: 1 Nov 1985 17:09-CST From: leff%smu.csnet@CSNET-RELAY.ARPA Subject: 1986 Eastern Simulation Conference 10-12 March 1986, Omni International Hotel, Norfolk, Virginia For more info contact: SCS, PO BOX 17900, San Diego, CA 92117 (619)277-3888 List of AI related titles: "TAT Teach" An Expert Training Simulator Knowledge-Based Opponent Simulation for Tactical Decision Training Simulators with Artificial Intelligence Expert Systems in Training/Decision/Simulation The Simulation Algorithm Itself: Driving the Inference Algorithm ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Wed Nov 13 03:03:04 1985 Date: Wed, 13 Nov 85 03:03:01 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: R Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a006935; 12 Nov 85 0:45 EST Date: Mon 11 Nov 1985 20:47-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V3 #168 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Wed, 13 Nov 85 02:49 EST AIList Digest Tuesday, 12 Nov 1985 Volume 3 : Issue 168 Today's Topics: Queries - Recent Work by Johnson/Laird & Conceptual Dependencies and Predicate Calculus WFFs, AI Tools - Typed languages and Lisp, Cryptography - RSA Complexity, Inference - Abduction, News - Computer Museum Micromouse Competition, Review - Commercial Machine Translation, Humor - Intelligence Quotation ---------------------------------------------------------------------- Date: 11 Nov 85 21:50:57 GMT From: Bob Stine Subject: Recent work by Johnson/Laird Can anyone give me some pointers to recent work by Johnson and Laird on the role of mental models in cognition? Thanks, - Bob Stine ------------------------------ Date: 11 Nov 85 21:58:58 GMT From: Bob Stine Subject: Conceptual Dependencies and Predicate Calculus wffs Anyone know of any work that has been done in translating conceptual dependency structures into predicate calculus wffs? Thanks, Bob Stine ------------------------------ Date: Fri, 8 Nov 1985 14:38 EST From: Skef Wholey Subject: Re: Typed languages and Lisp From: John Craig Chris Goad (Stanford CS grad) developed a language originally called SIL, now called RISE which is essentially a typing system added to Lisp, but with a no-type type so that one can get around typing as one desires. The main reasons for adding typing are: 1) faster code development (type checker finds bugs) A type checker can find some bugs, but it isn't clear that such bugs would take much time to find and fix relative the to the "real" bugs a programmer spends most of his time on. Also, actually entering type information can add to program development time. Controlled experiments are required to back claims like the above. 2) the compiler can use type information to generate more efficient object code (for example, less or no garbage collection pauses when running compiled code) I'll believe that type information can let a compiler generate more efficient code, but dynamic storage allocation (and therefore garbage collection) has almost nothing to do runtime typing. The exception to this is "number consing," which can be avoided by clever Lisp systems most of the time anyway. It seems to me like you get the best of lisp and typed worlds, and efficient code generated also. Its pretty fun, too. Common Lisp provides a very complete type declaration mechanism that lets one give the compiler a great deal of information. This information is used (by some Common Lisp compilers) to generate very efficient code. The difference is not that one language is typed and the other untyped, but that the default "typedness" is different. ------------------------------ Date: Fri, 8 Nov 85 10:18:06 cst From: ihnp4!gargoyle!simon@ucbvax.berkeley.edu (Janos Simon) Subject: Cryptography A small correction about the difficulty of breaking the RSA scheme: it is NOT NP-hard (although it is very likely that it is not invertible in polynomial time - in fact it is very likely that it cannot be inverted by polynomial time algorithms that use randomization (that yield correct answers with high probability). It is not hard to see that the RSA scheme can be broken if one knows the factorization of the underlying number. Now factoring is strongly suspected to be difficult (not doable in random polynomial time), but it is not known to be NP-hard, and there are good reasons to suspect that it isn't: 1)Both factoring and primality testing are in NP. That is not true of any NP-complete problem. If factoring would be NP-hard then NP would be closed under complementation. This would be a surprising answer to a very difficult question. 2)There is a deterministic factoring algorithm that runs in time exp(logn loglogn). This is not polynomial, but much less than exponential (2**n). Again, this would be a very unexpected behavior for an NP-hard problem. Janos Simon ------------------------------ Date: Mon, 11 Nov 85 09:27:54 EST From: munnari!basser.oz!anand@seismo.CSS.GOV Subject: Re: Abduction The term abduction( as applicable to AI) or retroduction was first coined by Charles Sanders Peirce. Deduction, Induction and Abduction are three types of reasoning mechanisms. DEDUCTION- Rule: All the beans from this bag are white. Case: These beans are from this bag. Therefore Result: These beans are white. INDUCTION- Case: These beans are from this bag. Result: These beans are white. Therefore Rule: All the beans from this bag are white. ABDUCTION- Rule: All the beans from this bag are white. Result: These beans are white. Therefore Case: These beans are from this bag. Induction is where we generalize from a number of cases of which something is true, and infer that the same thing is true for a whole class. Abduction is where we find some very curious circumstance which would be explained by the supposition that it was a case of a certain general rule and thereupon adopt the supposition. Refer the 'Collected Papers of Charles Sandes Peirce' Vol I & Vol II edited by Charles Hartstone & Paul Weiss, Harvard Uni. Press, 1960. (paragraph 65,66,67,68 of Vol I and paragraphs 623 & 624 of Vol II). (Postmaster:- This mail has been acknowledged.) ------------------------------ Date: Thu, 7 Nov 85 11:17:27 est From: Brian Harvey Subject: Computer Museum Micromouse Competition [Forwarded from the MIT bboard by SASW@MIT-MC.] English and Japanese robot "mice" will engage in a heated nose-to-nose competition at The Computer Museum on Saturday, November 23. The miniature self-guiding and self-propelled robots will compete for intelligence and speed in the official room-sized Micro Mouse Maze used in the World Micro Mouse Competition held in Japan last August. Schedule of Events: 11:00 - 12:00 Tours of maze; micromice on display 12:00 - 1:30 Mouse warm-up and adjustment 1:30 - 3:00 First micromouse race 3:00 - 3:30 Mouse warm-up and adjustment 3:30 - 5:00 Second micromouse race Would-be mouse designers and the simply curious can attend a special lecture and mouse demonstration clinic on Sunday, November 17 at 4:00 pm featuring England's noted mouse expert Professor John Billingsley. For more information call 426-2800 (a human being) or 357-8014 (a DECtalk voice synthesizer). ------------------------------ Date: 10 Nov 1985 2102-PST From: LAWS at SRI-AI.ARPA Subject: Commercial Machine Translation Title: Machines are Mastering the Language of Multinational Business Author: Joyce Heard with Leslie Helm Business Week (No. 2912, 9/16/85, pp. 90D ff.) This article describes machine translation systems that are currently available for translating English, German, French, Spanish, Italian, and Japanese. Speeds of up to 100,000 words per hour are claimed, as are accuracies of up to 90% and prices as low as $3,000. [Not all the same system, of course.] Customers are apparently willing to accept rough translations as long as they can get them quickly; translators, however, are not happy just polishing machine translations. Most of the companies offering multilingual services are converting text to a "neutral" language, then into the target language -- this greatly reduces the cost of additional source or target languages. NEC estimates that it needs about 100 "rules" for complete Japanese-English translation, and has developed 30. Europe has been the chief market so far, but most of the commercial leaders are American (Automated Language Processing Systems, Logos, World Translation Center, and Weidner). Fujitsu, Toshiba, NEC, and Bravice International are coming up fast, however. Philips and the Netherlands' BSO are also working on systems. ------------------------------ Date: Mon, 11 Nov 85 09:23:31 GMT From: gcj%qmc-ori.uucp@ucl-cs.arpa Subject: Future Intelligence Quotation >From the World Times, 11 November 2085 :- ``The World's first Intelligent System was put to the test today. On the standard IQ rating, it's score was...'' Gordon Joly gcj%qmc-ori@ucl-cs.arpa ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Wed Dec 18 01:23:09 1985 Date: Wed, 18 Dec 85 01:23:06 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a005254; 14 Nov 85 23:37 EST Date: Thu 14 Nov 1985 19:27-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V3 #169 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Fri, 15 Nov 85 03:32 EST AIList Digest Friday, 15 Nov 1985 Volume 3 : Issue 169 Today's Topics: Seminars - Question-Answering Systems (UPenn) & Bill, the Othello Program (CMU) & Information-Based Complexity (CSLI) & Multilisp (MIT) & Gazing in Theorem Proving (MCC) & Deductive Design Synthesis (SRI) & Probabilistic Propositional Logic (Buffalo) ---------------------------------------------------------------------- Date: Mon, 11 Nov 85 16:04 EST From: Tim Finin Subject: Seminar - Question-Answering Systems (UPenn) Forwarded From: Bonnie Webber on Mon 11 Nov 1985 at 10:42 Subj: Seminar in Natural Language Processing CIS679 - SEMINAR IN NATURAL LANGUAGE PROCESSING SPRING 1986 The topic for this term is question-answering systems, with particular attention to the type of information included in response to a question, instead of or in addition to an answer. We will look at the role of plan recognition and planning in formulating cooperative responses, as well as considering how to circumscribe the reasoning expected of a respondent. Response components of particular interest will be information intended to explain or justify answers, information intended to point out and/or correct misconceptions, and information intended to further the questioner's goals. Instructors: Joshi/Webber Time: MW 3-4:30 ------------------------------ Date: 13 Nov 85 12:54:10 EST From: Kai-Fu.Lee@SPEECH2.CS.CMU.EDU Subject: Seminar - Bill, the Othello Program (CMU) BILL : THE OTHELLO PROGRAM THAT BEAT IAGO Kai-Fu Lee Friday November 15, 1985 Wean Hall 5409 4:30 PM - 6:00 PM BILL is an Othello program written by myself and Sanjoy Mahajan. It was entered in the Waterloo Othello Tournament on November 9, and captured first place with a 4-0 record. In two unofficial games, it defeated IAGO, the world champion othello program developed at CMU in 1980-2. Most, if not all, othello programs use one of two types of evaluation functions: (1) knowledge-intensive but slow (such as IAGO), or (2) knowledge-deficient but fast (such as most programs at Waterloo). BILL succeeds through its use of a knowledge-intensive, yet extremely efficient evaluation function. It is further enhanced by an iterative deepening zero-window alpha-beta procedure, a hash table, a linked-move killer table, a two-phase end-game search, and thinking on opponent's time. In this talk, I will first discuss othello strategies. Next, I will describe Bill, and analyze its games in Waterloo and against IAGO. Finally, we will demonstrate BILL by playing it against the audience. ------------------------------ Date: Wed 13 Nov 85 17:05:26-PST From: Emma Pease Subject: Seminar - Information-Based Complexity (CSLI) [Excerpted from the CSLI Newsletter by Laws@SRI-AI.] An Introduction to Information-based Complexity J. F. Traub Computer Science Department, Columbia University THURSDAY, November 21, 1985 4:15 p.m. CSLI Colloquium, Redwood Hall, Room G-19 In information-based complexity ``information'' is, informally, what we know about a problem which we wish to solve. The goal of information-based complexity is to create a general theory about problems with partial and contaminated information and to apply the results to solving specific problems in varied disciplines. Problems with partial and contaminated information occur in areas such as vision, medical imaging, prediction, geophysical exploration, signal processing, control, and scientific and engineering calculation. For problems with partial and contaminated information, very general results can be obtained at the ``information level.'' Among the general results to be discussed is the power of parallel (non-adaptive) information and the application of such information to the solution of problems on distributed systems. The methodology and results of information-based complexity will be contrasted with the study of NP-complete problems where the information is assumed to be complete, exact, and free. ------------------------------ Date: Tue 12 Nov 85 12:45:02-EST From: "Brian C. Williams" Subject: Seminar - Multilisp (MIT) Thursday , October 14 4:00pm Room: NE43- 8th floor Playroom The Artificial Intelligence Lab Revolving Seminar Series "Multilisp: A Language for Parallel Symbolic Computing" Burt Halstead MIT, LCS Multilisp is an extension of Scheme with additional operators and additional semantics for parallel execution. These have been added without removing side effects from the language. The principal parallelism construct in Multilisp is the "future," which exhibits some features of both eager and lazy evaluation. Current work focuses on making Multilisp a more humane programming environment, and on expanding the power of Multilisp to express task scheduling policies. A skeletal Multilisp has been implemented, and has been run on the shared-memory Concert multiprocessor, using as many as eight processors, as well as on a BBN Butterfly machine with as many as 128 processors. The implementation uses interesting techniques for task scheduling and garbage collection. The task scheduler helps control excessive resource utilization by means of an unfair scheduling policy; the garbage collector uses a multiprocessor algorithm modeled after the incremental garbage collector of Baker. The talk will describe Multilisp, discuss the areas of current activity, and indicate the future direction of the project. ------------------------------ Date: Mon 11 Nov 85 15:55:02-CST From: AI.HASSAN@MCC.ARPA Subject: Seminar - Gazing in Theorem Proving (MCC) GAZING: USING THE STRUCTURE OF THE THEORY IN THEOREM PROVING Dave Plummer Department of Mathematics University of Texas at Austin Wednesday, November 20 10:00 a.m. Echelon I, Room 409 A mechanical theorem prover embodies two types of knowledge: logical and non-logical. The logical knowledge informs the prover which inferences are legal within the logic. The non-logical information, however, is specific to the theory that the prover is working in and includes definitions of concepts used in the theory, axioms, and previously proved facts. The theory is structured by relationships between these facts and these relationships may be exploited in order to provide guidance for a mechanical theorem prover. In this talk I will describe a technique, called Gazing, which exploits the structure of a theory, thus aiding a mechanical prover in determining which items of knowledge will be useful in the proof of a given goal. As concepts are defined in the theory, the system builds a graph representing the definitional order. This graph is used in two ways. First, whenever a new fact enters the theory, the ordering is used to determine an orientation of that fact creating a new rewrite rule. Secondly, the ordering is used to guide the search for a proof of a conjecture whenever the proof is known to require the use of non-logical facts. This guidance takes the form of determining which concepts are "close" in the definitional ordering, and attempting to find rewrite rules which may be used to rewrite two different concepts to a common new concept. The ordering can also be used to decide which of a number of possible common rewritings is preferable, and indeed if any common rewriting exists. ------------------------------ Date: Thu 14 Nov 85 14:31:42-PST From: LANSKY@SRI-AI.ARPA Subject: Seminar - Deductive Design Synthesis (SRI) EXPLOITATION OF CONSTRAINTS IN DEDUCTIVE DESIGN SYNTHESIS Jeff Finger Stanford University JFinger@SU-SUSHI PLANLUNCH 11:00 AM, MONDAY, November 18 SRI International, Building E, Room EJ228 (new conference room) The talk will cover two related topics in deductive design synthesis: (1) efficiency gained by reasoning forward from subgoals, and (2) advantages and disadvantages of using a declarative representation for partially completed designs. The first part of the talk gives the deductive framework for capturing the following intuition: Suppose I have decided that X and Y and to be true of my design. Perhaps I should think about what else X and Y imply about the design, say Z. Otherwise, I might waste time trying to complete the design process by making decisions that have *already* been ruled out by X and Y, for example, NOT(Z). The conditions that X and Y imply (called "necessary constraints" or "NC's") are found via reasoning forward from subgoals. We show how NC's of a subgoal can be used to prune the design space either by preventing some impossible possibilities from ever being generated or by providing a quick means of filtering bad choices. In terms of resolution, the above use of NC's corresponds to the rather counterintuitive notion of allowing OR-INTRODUCTION on clauses in the set of support. We will also discuss inheritance of NC's from goal to subgoal and the relation of finding NC's to that of checking consistency of partially completed designs. The second part of the talk deals with declarative representation of partially completed designs. Deductive design systems such as QA3 or Manna and Waldinger's reify the design as a single term in the logic. However, it is difficult to express many sorts of constraints on partially completed designs as a single term. Examples include two actions in an unspecified order, or the constraint that Action A takes place less than 3 seconds or more than 8 seconds after Action B. We present a system called RESIDUE in which we build up the design as a set of facts we are willing to assume about of the design. Using facts rather than a single term, we can make finer-grained decisions, avoiding unwitting commitments that might result in unnecessary backtracking. In addition, forward reasoning on subgoals (as in the first portion of the talk) may be done directly on the set of facts. ------------------------------ Date: Thu, 14 Nov 85 10:40:19 EST From: "William J. Rapaport" Subject: Seminar - Probabilistic Propositional Logic (Buffalo) UNIVERSITY AT BUFFALO STATE UNIVERSITY OF NEW YORK DEPARTMENT OF COMPUTER SCIENCE COLLOQUIUM DEXTER KOZEN Department of Computer Science Cornell University A PROBABILISTIC PROPOSITIONAL DYNAMIC LOGIC This talk concerns a probabilistic analog of Propositional Dynamic Logic, called Probabilistic Propositional Dynamic Logic (PPDL). PPDL is useful in the formal manipulation of simple pro- babilistic programs and the average-case analysis of determinis- tic programs. We describe the formal syntax and semantics of the system and its deductive calculus, and illustrate its use by cal- culating the expected running time of a simple random walk. We also describe briefly a polynomial-space decision procedure for deciding the truth of formulas involving well-structured pro- grams. Thursday, November 21, 1985 3:30 P.M. Bell 337, Amherst Campus Wine and cheese will be served at 4:30 P.M., 224 Bell Hall For further information, call (716) 636-3181. William J. Rapaport Assistant Professor Dept. of Computer Science, SUNY Buffalo, Buffalo, NY 14260 (716) 636-3193, 3180 uucp: ...{allegra,decvax,watmath}!sunybcs!rapaport ...{cmc12,hao,harpo}!seismo!rochester!rocksvax!sunybcs!rapaport cs: rapaport@buffalo arpa: rapaport%buffalo@csnet-relay bitnet: rapaport@sunybcs ------------------------------ End of AIList Digest ******************** From comsat@vtcs1 Wed Dec 18 01:12:39 1985 Date: Wed, 18 Dec 85 01:12:34 est From: comsat@vtcs1.VT To: fox@vtopus (MILLER,FRANCE,JOSLIN,ROACH,FOX) Subject: From: AIList Moderator Kenneth Laws Status: RO Received: from sri-ai.arpa by CSNET-RELAY.ARPA id a006001; 15 Nov 85 2:03 EST Date: Thu 14 Nov 1985 20:31-PST Reply-to: AIList%sri-ai.arpa@CSNET-RELAY US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V3 #170 To: AIList%sri-ai.arpa@CSNET-RELAY Received: from rand-relay by vpi; Fri, 15 Nov 85 03:35 EST AIList Digest Friday, 15 Nov 1985 Volume 3 : Issue 170 Today's Topics: Queries - Semantic Networks & Reason Maintenance System (or TMS), Representation - Conceptual Dependency and Predicate Calculus, New BBoard - TI Explorer, Hype - The Business World Flames as Well as We Do!, Inference - Rumor, Prejudice, and Uncertainty & Abduction and AI in Space Exploration ---------------------------------------------------------------------- Date: Mon, 11 Nov 85 09:24:07 EST From: "William J. Rapaport" Subject: another request for help When our system crashed, I also lost the address of a guy in Europe (Switzerland, I think) who wanted info on semantic networks. I'd greatly appreciate help on recovering his address. Thanks. ------------------------------ Date: Wed, 13 Nov 85 13:32:10 EST From: munnari!basser.oz!anand@seismo.CSS.GOV Subject: Reason Maintenance System(or TMS) Has anyone implemented a RMS using PROLOG? Would like to know the pros and cons of its implementation with LISP? Thank you in advance. -- Anand S. Rao ------------------------------ Date: Wed, 13 Nov 85 13:28:33 EST From: munnari!basser.oz!anand@seismo.CSS.GOV Subject: Conceptual dependency & predicate calculus Perhaps the best work on linking CD and predicate calculus is by John Sowa. Refer his book 'Conceptual Structures: Information Processing in Mind and Machine'. (Review in AI journal Sept. 1985) --- Anand S. Rao ------------------------------ Date: Tue, 12 Nov 85 13:19:23 gmt From: Patrick Hayes Subject: CD into PC In response to Bob Stines request concerning translating CD notation into 1PC. This should by now be regarded as a routine exercise, surely. Since logic doesnt have such ideas as physical transfer already incorporated into it, one has to translate into 1pc extended by the choice of a particular vocabulary of relations, etc., and this can be done in several ways ( n-place relations instead of (n-1)-place function symbols, for example ) : take your choice. You will need predicates such as PTRANS, of course, but also relations (or whatever) corresponding to the various colors of funny-arrow used in CD. There is a standard way to transform graphical notations into tree-structured notation such as 1PC or LISP: each node in the graph becomes a name in the language, and each link in the graph becomes an assertion that some relation ( which one depends on the color of the link ) holds between the entities named. In this way the graph maps into a conjunction of atomic assertions in a vocabulary which is just about as simple or complex as that used in the graphical language. Several notaional tricks can add variety to this simple idea, for example instead of mapping link2 into relation2(thing1,thing3) one can use exists x. Isrelation(x,tpe2) & Holds(x,thing1,thing3). This enables one to write general rules about a number of link types in a few compact axioms. Ask any experienced logic programmer for more ideas. Now, this just translates CD into 1PC notation, of course. To get the inferential power of CD one then needs to translate the inference rules into 1PC axioms written in the appropriate notation. If you can find the CD inference rules written out clearly somewhere, this should be straightforward. One might ask whether such a translation actually captures the meaning of CD adequately. Unfortunately, as ( to the best of knowledge ) CD notation has never been supplied with a clear semantics, this would have to remain a matter for subjective judgement. A last observation: if you check the published accounts of MARGIE, one of the early demonstration systems using CD, you will find that one-third of it was a program which manipulated CD graphs so as to draw conclusions. In order to do this, it first translated them into a tree-like notation similar to that obtained by the above technique. Pat Hayes ------------------------------ Date: 6 Nov 85 11:04:37 EST From: Kevin.Neel@ISL1.RI.CMU.EDU Subject: TI Explorer bbs [Forwarded from the CMU bboard by Laws@SRI-AI.] The following was posted on netnews: >Date: Fri 13 Sep 85 15:16:25-PDT >From: Richard Acuff >Subject: New Lists for TI Explorer Discussion In order to facilitate information exchange among DARPA sponsored projects using TI Explorers, two ArpaNet mailing lists are being created. INFO-EXPLORER will be used for general information distribution, such as operational questions, or announcing new generally available packages or tools. BUG-EXPLORER will be used to report problems with Explorer software, as well as fixes. Requests to be added to or deleted from these lists should be sent to INFO-EXPLORER-REQUEST or BUG-EXPLORER-REQUEST, respectively. All addresses are at SUMEX-AIM.ARPA. These lists signify no commitment from Texas Instruments or Stanford University. Indeed, there is no guarantee that TI representatives will read the lists. The idea of the lists is to provide communication among the users of Explorers. -- Rich Acuff Stanford KSL [...] ------------------------------ Date: 11 Nov 85 1549 PST From: Dick Gabriel Subject: The Business World Flames as Well as We Do! [Forwarded from the Stanford bboard by Laws@SRI-AI.] You might think that in moving to the business world I've given up on the joy of seeing first-class flaming in my normal environment - business ethics and all that. Wrong! The following is a quote from a story about Clarity Software Corp's new ad (soon to appear). Clarity is introducing a product called ``Logic Line-1,'' which is a natural language data retrieval system. The ad compares their product to competing AI products. They say, apparently about AI programmers: ``Luckily, we won't have to worry about their rancid cells polluting mankind's gene pool very long anyhow. Such brain-damaged geeks tend to die young. If you've recently spent money on artificial intelligence software, you might be wishing that a few programmers had croaked before writing that blithering swill they named AI and palmed off onto you.'' -rpg- ------------------------------ Date: Mon, 11 Nov 85 14:30 EST From: Mukhop Subject: Rumor, prejudice and the management of uncertainty AI research in recent years has extensively dealt with the management of uncertainty. A reasonable approach is to model human mechanisms for knowledge maintenance. However, these mechanisms are not perfect since they are vulnerable to rumor and prejudice. Both traits are universal; the object(s) of rumor or prejudice is a function of the culture and the times. Rumoring is illustrated in the following scenario: A passes some information to B and C, who in turn communicate it to others, and so on. It is possible for a person to receive the same information from several sources and consequently have a lot of confidence in its truth. The underlying uncertainty management calculus seems to be flawed since it ignores the fact that these sources are not independent. I would like to see some discussions on the following: 1) In any current AI system, is the test for independence of sources made prior to updating the uncertainty metric associated with a proposition? This seems to be especially relevant to Distributed AI systems. 2) Can someone suggest a model or scenario for prejudice? This may lead to a test to rid AI systems of it. 3) The human knowledge maintenance system (HKMS) seems to update knowledge in a reasonable manner when the information is received from independent sources but behaves erratically when the sources are not independent. Similarly, do the features of the HKMS, that cause prejudicial reasoning under some circumstances, lead to sound conclusions when certain conditions are met? How else could the HKMS have evolved in such a way? 4) The human visual system allows optical illusions to be formed, but is near-perfect for most routine activities (the bedouin who regularly observes mirages may beg to differ). It has also had more time to evolve. Is it conceivable that the HKMS will evolve in time so that it will be robust in the face of rumor and eliminate prejudicial reasoning? Or is it important to retain these traits to ensure "the survival of the fittest." Uttam Mukhopadhyay Computer Science Dept. General Motors Research Labs. Warren, MI 48090-9055 Phone: (313) 575-2105 [One model of prejudice is based on our propensity for prototype-based reasoning, combined with our tendency to focus on and remember the more extreme characteristics of prototypes. The fewer individuals we have seen from a population, the more certain we are that they are representative. The work of Kahneman and Tversky seems relevant. -- KIL] ------------------------------ Date: 13 Nov 1985 00:39-EST From: ISAACSON@USC-ISI.ARPA Subject: Abduction & AI in space exploration To my knowledge, abductive inference received some serious attention by NASA in the early 1980's. There is a heavy volume: ADVANCED AUTOMATION FOR SPACE MISSIONS, NASA Conference Publication 2255, Proceedings of the 1980 NASA/ASEE Summer Study, University of Santa Clara, CA [published end of 1982]. A certain "Space Exploration" team handled, among other things, futuristic requirements for advanced machine intelligence. (The task was to design a mission to Titan sometime around the year 2000.) The whole issue of abduction and hypothesis-formation was made a central issue in competition with "expert systems" soft- peddled by certain vested interests. The final "Conclusions and Recommended Technology Priorities" has in No. 1 place the following recommendation: (1) Machine intelligence systems with automatic hypothesis- formation [i.e., abduction - jdi] capability are necessary for autonomous examination of unknown environments. This capacity is highly desirable for efficient exploration of the Solar System and is essential for the ultimate investigation of other star systems. [p. 381] (Some well-known peddlers of expert systems actually wanted to send over there one of their expert systems, until confronted by the question of whose expertise they are going to package into the explorer... ) That recommendation is derived from the Space exploration report, p. 39-76. That report, p. 68, cites the following conclusion: Required machine intelligence technologies include: * Autonomous processing (essentially no programming) * Autonomous "dynamic" memory * Autonomous error-correction * Inherently parallel processing * Abductive/dialectic logical capabilities * General capacity for acquisition and recognition of patterns * Universal "Turing Machine" computability In the "Technology Assessment" section there are the following recommendations [p. 351]: 6.2.4 Initial Directions for NASA Several research tasks can be undertaken immediately by NASA which have the potential of contributing to the development of a fully automated hypothesis formulating ability needed for future space missions: (1) Continue to develop the perspective and theoretical basis for machine intelligence which holds that (a) machine intelligence and especially machine learning rest on a capability for autonomous hypothesis formation, (b) three distinct patterns of inference underlie hypothesis formulation - Analytic, inductive, and abductive inference, and (c) solving the problem of mechanizing abductive inference is the key to implementing successful machine learning systems. (This work should focus on abductive inference and begin laying the foundations for a theory of abductive inference in machine intelligence applications.) (2) Draw upon the emerging theory of abductive inference to establish a terminology for referring to abductive inference and its role in machine intelligence and learning. (3) Use this terminology to translate the emerging theory of abductive inference into the terminology of state-of-the-art AI; use these translations to connect abductive inference research needs with current AI work that touches on abduction, e.g., nonmonotonic logic; and then discuss these connections within the AI community. (the point of such an exercise is to identify those aspects of current AI work which can contribute to the achievement of mechanized and autonomous abductive inference systems, and to identify a sequence of research steps that the AI community can take towards this goal.) (4) Research proposals for specific machine intelligence projects should explain how the proposed project contributes to the ultimate goal of autonomous machine intelligence systems which learn by means of analytic, inductive, and abductive inferences. Enough is now known about the terms of this criterion to distinguish between projects which satisfy it and those which do not. ------------------------------ End of AIList Digest ********************