3-Jan-84 15:46:43-PST,10403;000000000001 Mail-From: LAWS created at 3-Jan-84 15:44:16 Date: Tue 3 Jan 1984 15:33-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #1 To: AIList@SRI-AI AIList Digest Wednesday, 4 Jan 1984 Volume 2 : Issue 1 Today's Topics: Administrivia - Host List & VISION-LIST, Cognitive Psychology - Looping Problem, Programming Languages - Questions, Logic Programming - Disjunctions, Vision - Fiber Optic Camera ---------------------------------------------------------------------- Date: Tue 3 Jan 84 15:07:27-PST From: Ken Laws Reply-to: AIList-Request@SRI-AI Subject: Host List The AIList readership has continued to grow throughout the year, and only a few individuals have asked to be dropped from the distribution network. I cannot estimate the number of readers receiving AIList through bboards and remailing nodes, but the existence of such services has obviously reduced the outgoing net traffic. For those interested in such things, I present the following approximate list of host machines on my direct distribution list. Numbers in parentheses indicate individual subscribers; all other hosts (and those marked with "bb") have redistribution systems. A few of the individual subscribers are undoubtedly redistributing AIList to their sites, and a few redistribution nodes receive the list from other such nodes (e.g., PARC-MAXC from RAND-UNIX). AIList is also available to USENET through the net.ai distribution system. AEROSPACE(8), AIDS-UNIX, BBNA(2), BBNG(1), BBN-UNIX(8), BBN-VAX(3), BERKELEY(3), BITNET@BERKELEY(2), ONYX@BERKELEY(1), UCBCAD@BERKELEY(2), BRANDEIS(1), BRL(bb+1), BRL-VOC(1), BROWN(1), BUFFALO-CS(1), cal-unix@SEISMO(1), CIT-20, CMU-CS-A(bb+11) CMU-CS-G(3), CMU-CS-SPICE(1), CMU-RI-ISL1(1), COLUMBIA-20, CORNELL, DEC-MARLBORO(7), EDXA@UCL-CS(1), GATECH, HI-MULTICS(bb+1), CSCKNP@HI-MULTICS(2), SRC@HI-MULTICS(1), houxa@UCLA-LOCUS(1), HP-HULK(1), IBM-SJ(1), JPL-VAX(1), KESTREL(1), LANL, LLL-MFE(2), MIT-MC, NADC(2), NOSC(4), NOSC-CC(1), CCVAX@NOSC(3), NPRDC(2), NRL-AIC, NRL-CSS, NSF-CS, NSWC-WO(2), NYU, TYM@OFFICE(bb+2), RADC-Multics(1), RADC-TOPS20, RAND-UNIX, RICE, ROCHESTER(2), RUTGERS(bb+2), S1-C(1), SAIL, SANDIA(bb+1), SCAROLINA(1), sdcrdcf@UCBVAX(1), SRI-AI(bb+6), SRI-CSL(1), SRI-KL(12), SRI-TSC(3), SRI-UNIX, SU-AI(2), SUMEX, SUMEX-AIM(2), SU-DSN, SU-SIERRA@SU-DSN(1), SUNY-SBCS(1), SU-SCORE(11), SU-PSYCH@SU-SCORE(1), TEKTRONIX(1), UBC, UCBKIM, UCF-CS, UCI, UCL-CS, UCLA-ATS(1), UCLA-LOCUS(bb+1), UDel-Relay(1), UIUC, UMASS-CS, UMASS-ECE(1), UMCP-CS, UMN-CS(bb+1), UNC, UPENN, USC-ECL(7), USC-CSE@USC-ECL(2), USC-ECLD@USC-ECL(1), SU-AI@USC-ECL(4), USC-ECLA(1), USC-ECLB(2), USC-ECLC(2), USC-ISI(5), USC-ISIB(bb+6), USC-ISID(1), USC-ISIE(2), USC-ISIF(10), UTAH-20(bb+2), utcsrgv@CCA-UNIX(1), UTEXAS-20, TI@UTEXAS-20(1), WISC-CRYS(3), WASHINGTON(4), YALE -- Ken Laws ------------------------------ Date: Fri, 30 Dec 83 15:20:41 PST From: Philip Kahn Subject: Are you interested in a more specialized "VISION-LIST"? I been feeling frustrated (again). I really like AIList, since it provides a nice forum for general AI topics. Yet, like many of you out there, I am primarily a vision researcher looking into ways to facilitate machine vision and trying to decipher the strange, all-too-often unknown mechanisms of sight. What we need is a specialized VISION-LIST to provide a more specific forum that will foster a greater exchange of ideas among our research. So...one question and one request: 1) is there such a list in the works?, and 2) if you are interested in such a list PLEASE SPEAK UP!! Thanks! Philip Kahn UCLA ------------------------------ Date: Fri 30 Dec 83 11:04:17-PST From: Rene Bach Subject: Loop detection Mike, It seems to me that we have an inbuilt mechanism which remembers what is done (thought) at all times. I.E. we know and remember (more or less) our train of thoughts. When we get in a loop, the mind is immediately triggered : at the first element, we think it could be a coincidence, as more elements are found matching the loop, the more convinced we get that there is a repeat : the reading example is quite good , even when just one word appears in the same sentence context (meaning rather than syntactical context), my mind is triggered and I go back and check if there is actually a loop or not. Thus to implement this property in the computer we would need a mechanism able to remember the path and check whether it has been followed already or not (and how far), at each step. Detection of repeats of logical rather than word for word sentences (or sets of ideas) is still left open. I think that the loop detection mechanism is part of the memorization process, which is an integral part of the reasoning engine and it is not sitting "on top" and monitoring the reasoning process from above. Rene ------------------------------ Date: 2 January 1984 14:40 EST From: Herb Lin Subject: stupid questions.... Speaking as an interested outsider to AI, I have a few questions that I hope someone can answer in non-jargon. Any help is greatly appreciated: 1. Just why is a language like LISP better for doing AI stuff than a language like PASCAL or ADA? In what sense is LISP "more natural" for simulating cognitive processes? Why can't you do this in more tightly structured languages like PASCAL? 2. What is the significance of not distinguishing between data and program in LISP? How does this help? 3. What is the difference between decisions made in a production system (as I understand it, a production is a construct of the form IF X is true, then do Y, where X is a condition and Y is a procedure), and decisions made in a PASCAL program (in which IF statements also have the same (superficial) form). many thanks. ------------------------------ Date: 1 Jan 84 1:01:50-PST (Sun) From: hplabs!hpda!fortune!rpw3 @ Ucb-Vax Subject: Re: Re: a trivial reasoning problem? - (nf) Article-I.D.: fortune.2135 Gee, and to a non-Prolog person (me) your problem seemed so simple (even given the no-exhaustive-search rule). Let's see, 1. At least one of A or B is on = (A v B) 2. If A is on, B is not = (A -> ~B) = (~A v (~B)) [def'n of ->] 3. A and B are binary conditions. >From #3, we are allowed to use first-order Boolean algebra (WFF'n'PROOF game). (That is, #3 is a meta-condition.) So, #1 and #2 together is just (#1) ^ (#2) [using caret ^ for disjunction] or, #1 ^ #2 = (A v B) ^ (~A v ~B) (distributivity) = (A ^ ~A) v (A ^ ~B) v (B ^ ~A) v (B ^ ~B) (from #3 and ^-axiom) = (A ^ ~B) v (B ^ ~A) (def'n of xor) = A xor B Hmmm... Maybe I am missing your original question altogether. Is your real question "How does one enumerate the elements of a state-space (powerset) for which a certain logical proposition is true without enumerating (examining) elements of the state-space for which the proposition is false?"? To me (an ignorant "non-ai" person), this seems excluded by a version of the First Law of Thermodynamics, namely, the Law of the Excluded Miraculous Sort (i.e. to tell which of two elements is bigger, you have to look at both). It seems to me that you must at least look at SOME of the states for which the proposition is false, or equivalently, you must use the structure of the formula itself to do the selection (say, while doing a tree-walk). The problem of the former approach is that the number of "bad" states should be kept small (for efficiency), leading to all kinds of pruning heuristics; while for the latter method the problem of elimination of duplicates (assuming parallel processing) leads to the former method! In either case, however, reasoning about the variables does not seem to solve the problem; one must reason about the formulae. If Prolog admits of constructing such meta-rules, you may have a chance. (I.e., "For all true formula 'X xor Y', only X need be considered when ~Y, & v-v.) In any event, I think your problem can be simplified to: 1'. A xor B 2'. A, B are binary variables. Rob Warnock UUCP: {sri-unix,amd70,hpda,harpo,ihnp4,allegra}!fortune!rpw3 DDD: (415)595-8444 USPS: Fortune Systems Corp, 101 Twin Dolphins Drive, Redwood City, CA 94065 ------------------------------ Date: 28 Dec 83 4:01:48-PST (Wed) From: hplabs!hpda!fortune!rpw3 @ Ucb-Vax Subject: Re: REFERENCES FOR SPECIALIZED CAMERA DE - (nf) Article-I.D.: fortune.2114 Please clarify what you mean by "get close to the focal point of the optical system". For any lens system I've used (both cameras and TVs), the imaging surface (the film or the sensor) already IS at the focal point. As I recall, the formula (for convex lenses) is: 1 1 1 --- = --- + --- f obj img where "f" is the focal length of the lens, "obj" the distance to the "object", and "img" the distance to the (real) image. Solving for minimum "obj + img", the closest you can get a focused image to the object (using a lens) is 4*f, with the lens midway between the object and the image (1/f = 1/2f + 1/2f). Not sure what a bundle of fibers would do for you, since without a lens each fiber picks up all the light around it within a cone of its numerical aperture (NA). Some imaging systems DO use fiber bundles directly in contact with film, but that's generally going the other way (from a CRT to film). I think Tektronix has a graphics output device like that. I suppose you could use it if the object were self-luminous... Rob Warnock UUCP: {sri-unix,amd70,hpda,harpo,ihnp4,allegra}!fortune!rpw3 DDD: (415)595-8444 USPS: Fortune Systems Corp, 101 Twin Dolphins Drive, Redwood City, CA 94065 ------------------------------ End of AIList Digest ******************** 4-Jan-84 16:42:36-PST,19474;000000000001 Mail-From: LAWS created at 4-Jan-84 16:38:41 Date: Wed 4 Jan 1984 16:31-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #2 To: AIList@SRI-AI AIList Digest Thursday, 5 Jan 1984 Volume 2 : Issue 2 Today's Topics: Hardware - High Resolution Video Projection, Programming Languages - LISP vs. Pascal, Net Course - AI and Mysticism ---------------------------------------------------------------------- Date: 04 Jan 84 1553 PST From: Fred Lakin Subject: High resolution video projection I want to buy a hi-resolution monochrome video projector suitable for use with generic LISP machine or Star-type terminals (ie approx 1000 x 1000 pixels). It would be nice if it cost less than $15K and didn't require expensive replacement parts (like light valves). Does anybody know of such currently on the market? I know, chances seem dim, so on to my second point: I have heard it would be possible to make a portable video projector that would cost $5K, weigh 25lb, and project using monochrome green phosphor. The problem is that industry does not feel the market demand would justify production at such a price ... Any ideas on how to find out the demand for such an item? Of course if all of you who might be interested in this kind of projector let me know your suggestions, that would be a good start. Thanks in advance for replies and/or notions, Fred Lakin ------------------------------ Date: Wed 4 Jan 84 10:25:56-PST From: Christopher Schmidt Subject: Re: stupid questions (i.e. Why Lisp?) You might want to read an article by Beau Sheil (Xerox PARC) in the February '83 issue of Datamation called "Power tools for programmers." It is mostly about the Interlisp-D programming environment, but might give you some insights about LISP in general. I'll offer three other reasons, though. Algol family languages lack the datatypes to conveniently implement a large number of knowledge representation schemes. Ditto wrt. rules. Try to imagine setting up a pascal record structure to embody the rules "If I have less than half of a tank of gas then I have as a goal stopping at a gas station" & "If I am carrying valuable goods, then I should avoid highway bandits." You could write pascal CODE that sort of implemented the above, but DATA would be extremely difficult. You would almost have to write a lisp interpreter in pascal to deal with it. And then, when you've done that, try writing a compiler that will take your pascal data structures and generate native code for the machine in question! Now, do it on the fly, as a knowledge engineer is augmenting the knowledge base! Algol languages have a tedious development cycle because they typically do not let a user load/link the same module many times as he debugs it. He typically has to relink the entire system after every edit. This prevents much in the way of incremental compilation, and makes such languages tedious to debug in. This is an argument against the languages in general, and doesn't apply to AI explicitly. The AI community feels this as a pressure more, though, perhaps because it tends to build such large systems. Furthermore, consider that most bugs in non-AI systems show up at compile time. If a flaw is in the KNOWLEDGE itself in an AI system, however, the flaws will only show up in the form of incorrect (unintelligent?) behavior. Typically only lisp-like languages provide the run-time tools to diagnose such problems. In Pascal, etc, the programmer would have to go back and explicitly put all sorts of debugging hooks into the system, which is both time consuming, and is not very clean. --Christopher ------------------------------ Date: 4 Jan 84 13:59:07 EST From: STEINBERG@RUTGERS.ARPA Subject: Re: Herb Lin's questons on LISP etc. Herb: Those are hardly stupid questions. Let me try to answer: 1. Just why is a language like LISP better for doing AI stuff than a language like PASCAL or ADA? There are two kinds of reasons. You could argue that LISP is more oriented towards "symbolic" processing than PASCAL. However, probably more important is the fact that LISP provides a truly outstanding environment for exploratory programming, that is, programming where you do not completely understand the problem or its solutions before you start programming. This is normally the case in AI programming - even if you think you understand things you normally find out there was at least something you were wrong about or had forgotten. That's one major reason for actually writing the programs. Note that I refer to the LISP environment, not just the language. The existence of good editors, debuggers, cross reference aids, etc. is at least as important as the language itself. A number of features of LISP make a good environment easy to provide for LISP. These include the compatible interpreter/compiler, the centrality of function calls, and the simplicity and accessibility of the internal representation of programs. For a very good introduction to the flavor of programming in LISP environments, see "Programming in an Interactive Environment, the LISP Experience", by Erik Sandewall, Computing Surveys, V. 10 #1, March 1978. 2. What is the significance of not distinguishing between data and program in LISP? How does this help? Actually, in ANY language, the program is also data for the interpreter or compiler. What is important about LISP is that the internal form used by the interpreter is simple and accessible. It is simple in that the the internal form is a structure of nested lists that captures most of both the syntactic and the semantic structure of the code. It is accessible in that this structure of nested lists is in fact a basic built in data structure supported by all the facilities of the system, and in that a program can access or set the definition of a function. Together these make it easy to write programs which operate on other programs. E.g. to add a trace feature to PASCAL you have to modify the compiler or interpreter. To add a trace feature to LISP you need not modify the interpreter at all. Furthermore, it turns out to be easy to use LISP to write interpreters for other languages, as long as the other languages use a similar internal form and have a similarly simple relation between form and semantics. Thus, a common way to solve a problem in LISP is to implement a language in which it is easy to express solutions to problems in a general class, and then use this language to solve your particular problem. See the Sandewall article mentioned above. 3. What is the difference between decisions made in a production system and decisions made in a PASCAL program (in which IF statements also have the same (superficial) form). Production Systems gain some advantages by restricting the languages for the IF and THEN parts. Also, in many production systems, all the IF parts are evaluated first, to see which are true, before any THEN part is done. If more than one IF part is true, some other mechanism decides which THEN part (or parts) to do. Finally, some production systems such as EMYCIN do "backward chaining", that is, one starts with a goal and asks which THEN parts, if they were done, would be useful in achieving the goal. One then looks to see if their corresponding IF parts are true, or can be made true by treating them as sub-goals and doing the same kind of reasoning on them. A very good introduction to production systems is "An Overview of Production Systems" by Randy Davis and Jonathan King, October 1975, Stanford AI Lab Memo AIM-271 and Stanford CS Dept. Report STAN-CS-75-524. It's probably available from the National Technical Information Service. ------------------------------ Date: 1 Jan 84 8:42:34-PST (Sun) From: harpo!floyd!clyde!akgua!psuvax!bobgian @ Ucb-Vax Subject: Netwide Course -- AI and Mysticism!! Article-I.D.: psuvax.395 ************************************************************************* * * * An Experiment in Teaching, an Experiment in AI * * Spring Term Artificial Intelligence Seminar Announcement * * * ************************************************************************* This Spring term Penn State inaugurates a new experimental course: "THE HUMAN CONDITION: PROBLEMS AND CREATIVE SOLUTIONS". This course explores all that makes the human condition so joyous and delightful: learning, creative expression, art, music, inspiration, consciousness, awareness, insight, sensation, planning, action, community. Where others study these DESCRIPTIVELY, we will do so CONSTRUCTIVELY. We will gain familiarity by direct human experience and by building artificial entities which manifest these wonders!! We will formulate and study models of the human condition -- an organism of bounded rationality confronting a bewilderingly complex environment. The human organism must fend for survival, but it is aided by some marvelous mechanisms: perception (vision, hearing), cognition (understanding, learning, language), and expression (motor skill, music, art). We can view these respectively as the input, processing, and output of symbolic information. These mechanisms somehow encode all that is uniquely human in our experience -- or do they?? Are these mechanisms universal among ALL sentient beings, be they built from doped silicon or neural jelly? Are these mechanisms really NECESSARY and SUFFICIENT for sentience? Not content with armchair philosophizing, we will push these models toward the concreteness needed for physical implementation. We will build the tools that will help us to understand and use the necessary representations and processes, and we will use these tools to explore the space of possible realizations of "artificial sentience". This will be no ordinary course. For one thing, it has no teacher. The course will consist of a group of highly energetic individuals engaged in seeking the secrets of life, motivated solely by the joy of the search itself. I will function as a "resource person" to the extent my background allows, but the real responsibility for the success of the expedition rests upon ALL of its members. My role is that of "encounter group facilitator": I jab when things lag. I provide a sheltered environment where the shy can "come out" without fear. I manipulate and connive to keep the discussions going at a fever pitch. I pick and poke, question and debunk, defend and propose, all to incite people to THINK and to EXPRESS. Several people who can't be at Penn State this Spring told me they wish they could participate -- so: I propose opening this course to the entire world, via the miracles of modern networks! We have arranged a local mailing list for sharing discussions, source-code, class-session summaries, and general flammage (with the chaff surely will be SOME wheat). I'm aware of three fora for sharing this: USENET's net.ai, Ken Laws' AIList, and MIT's SELF-ORG mailing list. PLEASE MAIL ME YOUR REACTIONS to using these resources: would YOU like to participate? would it be a productive use of the phone lines? would it be more appropriate to go to /dev/null? The goals of this course are deliberately ambitious. I seek participants who are DRIVEN to partake in this journey -- the best, brightest, most imaginative and highly motivated people the world has to offer. Course starts Monday, January 16. If response is positive, I'll post the network arrangements about that time. This course is dedicated to the proposition that the best way to secure for ourselves the blessings of life, liberty, and the pursuit of happiness is reverence for all that makes the human condition beautiful, and the best way to build that reverence is the scientific study and construction of the marvels that make us truly human. -- Bob Giansiracusa (Dept of Computer Science, Penn State Univ, 814-865-9507) Arpa: bobgian%psuvax1.bitnet@Berkeley Bitnet: bobgian@PSUVAX1.BITNET CSnet: bobgian@penn-state.csnet UUCP: allegra!psuvax!bobgian USnail: 333 Whitmore Lab, Penn State Univ, University Park, PA 16802 ------------------------------ Date: 1 Jan 84 8:46:31-PST (Sun) From: harpo!floyd!clyde!akgua!psuvax!bobgian @ Ucb-Vax Subject: Netwide AI Course -- Part 2 Article-I.D.: psuvax.396 ************************************************************************* * * * Spring Term Artificial Intelligence Seminar Syllabus * * * ************************************************************************* MODELS OF SENTIENCE Learning, Cognitive Model Formation, Insight, Discovery, Expression; "Subcognition as Computation", "Cognition as Subcomputation"; Physical, Cultural, and Intellectual Evolution. SYMBOLIC INPUT CHANNELS: PERCEPTION Vision, hearing, signal processing, the "signal/symbol interface". SYMBOLIC PROCESSING: COGNITION Language, Understanding, Goals, Knowledge, Reasoning. SYMBOLIC OUTPUT CHANNELS: EXPRESSION Motor skills, Artistic and Musical Creativity, Story Creation, Prose, Poetry, Persuasion, Beauty. CONSEQUENCES OF THESE MODELS Physical Symbol Systems and Godel's Incompleteness Theorems; The "Aha!!!" Phenomenon, Divine Inspiration, Extra-Sensory Perception, The Conscious/Unconscious Mind, The "Right-Brain/Left-Brain" Dichotomy; "Who Am I?", "On Having No Head"; The Nature and Texture of Reality; The Nature and Role of Humor; The Direct Experience of the Mystical. TECHNIQUES FOR DEVELOPING THESE ABILITIES IN HUMANS Meditation, Musical and Artistic Experience, Problem Solving, Games, Yoga, Zen, Haiku, Koans, "Calculus for Peak Experiences". TECHNIQUES FOR DEVELOPING THESE ABILITIES IN MACHINES REVIEW OF LISP PROGRAMMING AND FORMAL SYMBOL MANIPULATION: Construction and access of symbolic expressions, Evaluation and Quotation, Predicates, Function definition; Functional arguments and returned values; Binding strategies -- Local versus Global, Dynamic versus Lexical, Shallow versus Deep; Compilation of LISP. IMPLEMENTATION OF LISP: Storage Mapping and the Free List; The representation of Data: Typed Pointers, Dynamic Allocation; Symbols and the Symbol Table (Obarray); Garbage Collection (Sequential and Concurrent algorithms). REPRESENTATION OF PROCEDURE: Meta-circular definition of the evaluation process. "VALUES" AND THE OBJECT-ORIENTED VIEW OF PROGRAMMING: Data-Driven Programming, Message-Passing, Information Hiding; the MIT Lisp Machine "Flavor" system; Functional and Object-Oriented systems -- comparison with SMALLTALK. SPECIALIZED AI PROGRAMMING TECHNIQUES: Frames and other Knowledge Representation Languages, Discrimination Nets, Augmented Transition Networks; Pattern-Directed Inference Systems, Agendas, Chronological Backtracking, Dependency-Directed Backtracking, Data Dependencies, Non-Monotonic Logic, and Truth-Maintenance Systems. LISP AS THE "SYSTEMS SUBSTRATE" FOR HIGHER LEVEL ABSTRACTIONS: Frames and other Knowledge Representation Languages, Discrimination Nets, "Higher" High-Level Languages: PLANNER, CONNIVER, PROLOG. SCIENTIFIC AND ETHICAL CONSEQUENCES OF THESE ABILITIES IN HUMANS AND IN MACHINES The Search for Extra-Terrestrial Intelligence. (Would we recognize it if we found it? Would they recognize us?) The Search for Terrestrial Intelligence. Are We Unique? Are we worth saving? Can we save ourselves? Why are we here? Why is ANYTHING here? WHAT is here? Where ARE we? ARE we? Is ANYTHING? These topics form a cluster of related ideas which we will pursue more-or- less concurrently; the listing is not meant to imply a particular sequence. Various course members have expressed interest in the following software engineering projects. These (and possibly others yet to be suggested) will run concurrently throughout the course: LISP Implementations: For CMS, in PL/I and/or FORTRAN In PASCAL, optimized for personal computers (esp HP 9816) In Assembly, optimized for Z80 and MC68000 In 370 BAL, modifications of LISP 1.5 New "High-Level" Systems Languages: Flavor System (based on the MIT Zetalisp system) Prolog Interpreter (plus compiler?) Full Programming Environment (Enhancements to LISP): Compiler, Editor, Workspace Manager, File System, Debug Tools Architectures and Languages for Parallel {Sub-}Cognition: Software and Hardware Alternatives to the Von-Neuman Computer Concurrent Processing and Message Passing systems Machine Learning and Discovery Systems: Representation Language for Machine Learning Strategy Learning for various Games (GO, CHECKERS, CHESS, BACKGAMMON) Perception and Motor Control Systems: Vision (implementations of David Marr's theories) Robotic Welder control system Creativity Systems: Poetry Generators (Haiku) Short-Story Generators Expert Systems (traditional topic, but including novel features): Euclidean Plane Geometry Teaching and Theorem-Proving system Welding Advisor Meteorological Analysis Teaching system READINGS -- the following books will be very helpful: 1. ARTIFICIAL INTELLIGENCE, Patrick H. Winston; Addison Wesley, 1984. 2. THE HANDBOOK OF ARTIFICIAL INTELLIGENCE, Avron Barr, Paul Cohen, and Edward Feigenbaum; William Kaufman Press, 1981 and 1982. Vols 1, 2, 3. 3. MACHINE LEARNING, Michalski, Carbonell, and Mitchell; Tioga, 1983. 4. GODEL, ESCHER, BACH: AN ETERNAL GOLDEN BRAID, Douglas R. Hofstadter; Basic Books, 1979. 5. THE MIND'S I, Douglas R. Hofstadter and Daniel C. Dennett; Basic Books, 1981. 6. LISP, Patrick Winston and Berthold K. P. Horn; Addison Wesley, 1981. 7. ANATOMY OF LISP, John Allen; McGraw-Hill, 1978. 8. ARTIFICIAL INTELLIGENCE PROGRAMMING, Eugene Charniak, Christopher K. Riesbeck, and Drew V. McDermott; Lawrence Erlbaum Associates, 1980. -- Bob Giansiracusa (Dept of Computer Science, Penn State Univ, 814-865-9507) Arpa: bobgian%psuvax1.bitnet@Berkeley Bitnet: bobgian@PSUVAX1.BITNET CSnet: bobgian@penn-state.csnet UUCP: allegra!psuvax!bobgian USnail: 333 Whitmore Lab, Penn State Univ, University Park, PA 16802 ------------------------------ End of AIList Digest ******************** 5-Jan-84 11:20:47-PST,18365;000000000001 Mail-From: LAWS created at 4-Jan-84 17:33:00 Date: Wed 4 Jan 1984 17:23-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #3 To: AIList@SRI-AI AIList Digest Thursday, 5 Jan 1984 Volume 2 : Issue 3 Today's Topics: Course - Penn State's First Undergrad AI Course ---------------------------------------------------------------------- Date: 31 Dec 83 15:18:20-PST (Sat) From: harpo!floyd!clyde!akgua!psuvax!bobgian @ Ucb-Vax Subject: Penn State's First Undergrad AI Course Article-I.D.: psuvax.380 Last fall I taught Penn State's first ever undergrad AI course. It attracted 150 students, including about 20 faculty auditors. I've gotten requests from several people initiating AI courses elsewhere, and I'm posting this and the next 6 items in hopes they may help others. 1. General Information 2. Syllabus (slightly more detailed topic outline) 3. First exam 4. Second exam 5. Third exam 6. Overview of how it went. I'll be giving this course again, and I hate to do anything exactly the same twice. I welcome comments and suggestions from all net buddies! -- Bob [Due to the length of Bob's submission, I will send the three exams as a separate digest. Bob's proposal for a network AI course associated with his spring semester curriculum was published in the previous AIList issue; that was entirely separate from the following material. -- Ken Laws] -- Spoken: Bob Giansiracusa Bell: 814-865-9507 Bitnet: bobgian@PSUVAX1.BITNET Arpa: bobgian%psuvax1.bitnet@Berkeley CSnet: bobgian@penn-state.csnet UUCP: allegra!psuvax!bobgian USnail: Dept of Comp Sci, Penn State Univ, University Park, PA 16802 ------------------------------ Date: 31 Dec 83 15:19:52-PST (Sat) From: harpo!floyd!clyde!akgua!psuvax!bobgian @ Ucb-Vax Subject: PSU's first AI course, Part 1/6 Article-I.D.: psuvax.381 CMPSC 481: INTRODUCTION TO ARTIFICIAL INTELLIGENCE An introduction to the theory, research paradigms, implementation techniques, and philosopies of Artificial Intelligence considered both as a science of natural intelligence and as the engineering of mechanical intelligence. OBJECTIVES -- To provide: 1. An understanding of the principles of Artificial Intelligence; 2. An appreciation for the power and complexity of Natural Intelligence; 3. A viewpoint on programming different from and complementary to the viewpoints engendered by other languages in common use; 4. The motivation and tools for developing good programming style; 5. An appreciation for the power of abstraction at all levels of program design, especially via embedded compilers and interpreters; 6. A sense of the excitement at the forefront of AI research; and 7. An appreciation for the tremendous impact the field has had and will continue to have on our perception of our place in the Universe. TOPIC SUMMARY: INTRODUCTION: What is "Intelligence"? Computer modeling of "intelligent" human performance. The Turing Test. Brief history of AI. Relation of AI to psychology, computer science, management, engineering, mathematics. PRELUDE AND FUGUE ON THE "SECRET OF INTELLIGENCE": "What is a Brain that it may possess Intelligence, and Intelligence that it may inhabit a Brain?" Introduction to Formal Systems, Physical Symbol Systems, and Multilevel Interpreters. Necessity and Sufficiency of Physical Symbol Systems as the basis for intelligence. REPRESENTATION OF PROBLEMS, GOALS, ACTIONS, AND KNOWLEDGE: State Space, Predicate Calculus, Production Systems, Procedural Representations, Semantic Networks, Frames and Scripts. THE "PROBLEM-SOLVING" PARADIGM AND TECHNIQUES: Generate and Test, Heuristic Search (Search WITH Heuristics, Search FOR Heuristics), Game Trees, Minimax, Problem Decomposition, Means-Ends Analysis, The General Problem Solver (GPS). LISP PROGRAMMING: Symbolic Expressions and Symbol Manipulation, Data Structures, Evaluation and Quotation, Predicates, Input/Output, Recursion. Declarative and Procedural knowledge representation in LISP. LISP DETAILS: Storage Mapping, the Free List, and Garbage Collection, Binding strategies and the concept of the "Environment", Data-Driven Programming, Message-Passing, The MIT Lisp Machine "Flavor" system. LISP AS THE "SYSTEMS SUBSTRATE" FOR HIGHER LEVEL ABSTRACTIONS: Frames and other Knowledge Representation Languages, Discrimination Nets, "Higher" High-Level Languages: PLANNER, CONNIVER, PROLOG. LOGIC, RULE-BASED SYSTEMS, AND INFERENCE: Logic: Axioms, Rules of Inference, Theorems, Truth, Provability. Production Systems: Rule Interpreters, Forward/Backward Chaining. Expert Systems: Applied Knowledge Representation and Inference. Data Dependencies, Non-Monotonic Logic, and Truth-Maintenance Systems, Theorem Proving, Question Answering, and Planning systems. THE UNDERSTANDING OF NATURAL LANGUAGE: Formal Linguistics: Grammars and Machines, the Chomsky Hierarchy. Syntactic Representation: Augmented Transition Networks (ATNs). Semantic Representation: Conceptual Dependency, Story Understanding. Spoken Language Understanding. ROBOTICS: Machine Vision, Manipulator and Locomotion Control. MACHINE LEARNING: The Spectrum of Learning: Learning by Adaptation, Learning by Being Told, Learning from Examples, Learning by Analogy, Learning by Experimentation, Learning by Observation and Discovery. Model Induction via Generate-and-Test, Automatic Theory Formation. A Model for Intellectual Evolution. RECAPITULATION AND CODA: The knowledge representation and problem-solving paradigms of AI. The key ideas and viewpoints in the modeling and creation of intelligence. Is there more (or less) to Intelligence, Consciousness, the Soul? Prospectus for the future. Handouts for the course include: 1. Computer Science as Empirical Inquiry: Symbols and Search. 1975 Turing Award Lecture by Allen Newell and Herb Simon; Communications of the ACM, Vol. 19, No. 3, March 1976. 2. Steps Toward Artificial Intelligence. Marvin Minsky; Proceedings of the IRE, Jan. 1961. 3. Computing Machinery and Intelligence. Alan Turing; Mind (Turing's original proposal for the "Turing Test"). 4. Exploring the Labyrinth of the Mind. James Gleick; New York Times Magazine, August 21, 1983 (article about Doug Hofstadter's recent work). TEXTBOOKS: 1. ARTIFICIAL INTELLIGENCE, Patrick H. Winston; Addison Wesley, 1983. Will be available from publisher in early 1984. I will distribute a copy printed from Patrick's computer-typeset manuscript. 2. LISP, Patrick Winston and Berthold K. P. Horn; Addison Wesley, 1981. Excellent introductory programming text, illustrating many AI implementation techniques at a level accessible to novice programmers. 4. GODEL, ESCHER, BACH: AN ETERNAL GOLDEN BRAID, Douglas R. Hofstadter; Basic Books, 1979. One of the most entertaining books on the subject of AI, formal systems, and symbolic modeling of intelligence. 5. THE HANDBOOK OF ARTIFICIAL INTELLIGENCE, Avron Barr, Paul Cohen, and Edward Feigenbaum; William Kaufman Press, 1981 and 1982. Comes as a three volume set. Excellent (the best available), but the full set costs over $100. 6. ANATOMY OF LISP, John Allen; McGraw-Hill, 1978. Excellent text on the definition and implementation of LISP, sufficient to enable one to write a complete LISP interpreter. ------------------------------ Date: 31 Dec 83 15:21:46-PST (Sat) From: harpo!floyd!clyde!akgua!psuvax!bobgian @ Ucb-Vax Subject: PSU's first AI course -- part 2/6 (Topic Outline) Article-I.D.: psuvax.382 CMPSC 481: INTRODUCTION TO ARTIFICIAL INTELLIGENCE TOPIC OUTLINE: INTRODUCTION: What is "Intelligence"? Computer modeling of "intelligent" human performance. Turing Test. Brief history of AI. Examples of "intelligent" programs: Evan's Geometric Analogies, the Logic Theorist, General Problem Solver, Winograd's English language conversing blocks world program (SHRDLU), MACSYMA, MYCIN, DENDRAL. PRELUDE AND FUGUE ON THE "SECRET OF INTELLIGENCE": "What is a Brain that it may possess Intelligence, and Intelligence that it may inhabit a Brain?" Introduction to Formal Systems, Physical Symbol Systems, and Multilevel Interpreters. REPRESENTATION OF PROBLEMS, GOALS, ACTIONS, AND KNOWLEDGE: State Space problem formulations. Predicate Calculus. Semantic Networks. Production Systems. Frames and Scripts. SEARCH: Representation of problem-solving as graph search. "Blind" graph search: Depth-first, Breadth-first. Heuristic graph search: Best-first, Branch and Bound, Hill-Climbing. Representation of game-playing as tree search: Static Evaluation, Minimax, Alpha-Beta. Heuristic Search as a General Paradigm: Search WITH Heuristics, Search FOR Heuristics THE GENERAL PROBLEM SOLVER (GPS) AS A MODEL OF INTELLIGENCE: Goals and Subgoals -- problem decomposition Difference-Operator Tables -- the solution to subproblems Does the model fit? Does GPS work? EXPERT SYSTEMS AND KNOWLEDGE ENGINEERING: Representation of Knowledge: The "Production System" Movement The components: Knowledge Base Inference Engine Examples of famous systems: MYCIN, TEIRESIAS, DENDRAL, MACSYMA, PROSPECTOR INTRODUCTION TO LISP PROGRAMMING: Symbolic expressions and symbol manipulation: Basic data types Symbols The special symbols T and NIL Numbers Functions Assignment of Values to Symbols (SETQ) Objects constructed from basic types Constructor functions: CONS, LIST, and APPEND Accessor functions: CAR, CDR Evaluation and Quotation Predicates Definition of Functions (DEFUN) Flow of Control (COND, PROG, DO) Input and Output (READ, PRINT, TYI, TYO, and friends) REPRESENTATION OF DECLARATIVE KNOWLEDGE IN LISP: Built-in representation mechanisms Property lists Arrays User-definable data structures Data-structure generating macros (DEFSTRUCT) Manipulation of List Structure "Pure" operations (CONS, LIST, APPEND, REVERSE) "Impure" operations (RPLACA and RPLACD, NCONC, NREVERSE) Storage Mapping, the Free List, and Garbage Collection REPRESENTATION OF PROCEDURAL KNOWLEDGE IN LISP: Types of Functions Expr: Call by Value Fexpr: Call by Name Macros and macro-expansion Functions as Values APPLY, FUNCALL, LAMBDA expressions Mapping operators (MAPCAR and friends) Functional Arguments (FUNARGS) Functional Returned Values (FUNVALS) THE MEANING OF "VALUE": Assignment of values to symbols Binding of values to symbols "Local" vs "Global" variables "Dynamic" vs "Lexical" binding "Shallow" vs "Deep" binding The concept of the "Environment" "VALUES" AND THE OBJECT-CENTERED VIEW OF PROGRAMMING: Data-Driven programming Message-passing Information Hiding Safety through Modularity The MIT Lisp Machine "Flavor" system LISP'S TALENTS IN REPRESENTATION AND SEARCH: Representation of symbolic structures in LISP Predicate Calculus Rule-Based Expert Systems (the Knowledge Base examined) Frames Search Strategies in LISP Breadth-first, Depth-first, Best-first search Tree search and the simplicity of recursion Interpretation of symbolic structures in LISP Rule-Based Expert Systems (the Inference Engine examined) Symbolic Mathematical Manipulation Differentiation and Integration Symbolic Pattern Matching The DOCTOR program (ELIZA) LISP AS THE "SYSTEMS SUBSTRATE" FOR HIGHER LEVEL ABSTRACTIONS Frames and other Knowledge Representation Languages Discrimination Nets Augmented Transition Networks (ATNs) as a specification of English syntax Interpretation of ATNs Compilation of ATNs Alternative Control Structures Pattern-Directed Inference Systems (production system interpreters) Agendas (best-first search) Chronological Backtracking (depth-first search) Dependency-Directed Backtracking Data Dependencies, Non-Monotonic Logic, and Truth-Maintenance Systems "Higher" High-Level Languages: PLANNER, CONNIVER PROBLEM SOLVING AND PLANNING: Hierarchical models of planning GPS, STRIPS, ABSTRIPS Non-Hierarchical models of planning NOAH, MOLGEN THE UNDERSTANDING OF NATURAL LANGUAGE: The History of "Machine Translation" -- a seemingly simple task The Failure of "Machine Translation" -- the need for deeper understanding The Syntactic Approach Grammars and Machines -- the Chomsky Hierarchy RTNs, ATNs, and the work of Terry Winograd The Semantic Approach Conceptual Dependency and the work of Roger Schank Spoken Language Understanding HEARSAY HARPY ROBOTICS: Machine Vision Early visual processing (a signal processing approach) Scene Analysis and Image Understanding (a symbolic processing approach) Manipulator and Locomotion Control Statics, Dynamics, and Control issues Symbolic planning of movements MACHINE LEARNING: Rote Learning and Learning by Adaptation Samuel's Checker player Learning from Examples Winston's ARCH system Mitchell's Version Space approach Learning by Planning and Experimentation Samuel's program revisited Sussman's HACKER Mitchell's LEX Learning by Heuristically Guided Discovery Lenat's AM (Automated Mathematician) Extending the Heuristics: EURISKO Model Induction via Generate-and-Test The META-DENDRAL project Automatic Formation of Scientific Theories Langley's BACON project A Model for Intellectual Evolution (my own work) RECAP ON THE PRELUDE AND FUGUE: Formal Systems, Physical Symbol Systems, and Multilevel Interpreters revisited -- are they NECESSARY? are they SUFFICIENT? Is there more (or less) to Intelligence, Consciousness, the Soul? SUMMARY, CONCLUSIONS, AND FORECASTS: The representation of knowledge in Artificial Intelligence The problem-solving paradigms of Artificial Intelligence The key ideas and viewpoints in the modeling and creation of intelligence The results to date of the noble effort Prospectus for the future ------------------------------ Date: 31 Dec 83 15:28:32-PST (Sat) From: harpo!floyd!clyde!akgua!psuvax!bobgian @ Ucb-Vax Subject: PSU's first AI course -- part 6/6 (Overview) Article-I.D.: psuvax.386 A couple of notes about how the course went. Interest was high, but the main problem I found is that Penn State students are VERY strongly conditioned to work for grades and little else. Most teachers bore them, expect them to memorize lectures and regurgitate on exams, and students then get drunk (over 50 frats here) and promptly forget all. Initially I tried to teach, but I soon realized that PEOPLE CAN LEARN (if they really want to) BUT NOBODY CAN TEACH (students who don't want to learn). As the course evolved my role became less "information courier" and more "imagination provoker". I designed exams NOT to measure learning but to provoke thinking (and thereby learning). The first exam (on semantic nets) was given just BEFORE covering that topic in lecture -- students had a hell of a hard time on the exam, but they sure sat up and paid attention to the next week's lectures! For the second exam I announced that TWO exams were being given: an easy one (if they sat on one side of the room) and a hard one (on other side). Actually the exams were identical. (This explains the first question.) The winning question submitted from the audience related to the chapter in GODEL, ESCHER, BACH on the MU system: I gave a few axioms and inference rules and then asked whether a given wff was a theorem. The third exam was intended ENTIRELY to provoke discussion and NOT AT ALL to measure anything. It started with deadly seriousness, then (about 20 minutes into the exam) a few "audience plants" started acting out a prearranged script which included discussing some of the questions and writing some answers on the blackboard. The attempt was to puncture the "exam mentality" and generate some hot-blooded debate (you'll see what I mean when you see the questions). Even the Teaching Assistants were kept in the dark about this "script"! Overall, the attempt failed, but many people did at least tell me that taking the exams was the most fun part of the course! With this lead-in, you probably have a clearer picture of some of the motivations behind the spring term course. To put it bluntly: I CANNOT TEACH AI. I CAN ONLY HOPE TO INSPIRE INTERESTED STUDENTS TO WANT TO LEARN AI. I'LL DO ANYTHING I CAN THINK OF WHICH INCREASES THAT INSPIRATION. The motivational factors also explain my somewhat unusual grading system. I graded on creativity, imagination, inspiration, desire, energy, enthusiasm, and gusto. These were partly measured by the exams, partly by the energy expended on several optional projects (and term paper topics), and partly by my seat-of-the-pants estimate of how determined a student was to DO real AI. This school prefers strict objective measures of student performance. Tough. This may all be of absolutely no relevance to others teaching AI. Maybe I'm just weird. I try to cultivate that image, for it seems to attract the best and brightest students! -- Bob Giansiracusa ------------------------------ End of AIList Digest ******************** 5-Jan-84 11:33:19-PST,17105;000000000001 Mail-From: LAWS created at 5-Jan-84 11:30:29 Date: Thu 5 Jan 1984 11:16-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #4 To: AIList@SRI-AI AIList Digest Thursday, 5 Jan 1984 Volume 2 : Issue 4 Today's Topics: Course - PSU's First AI Course (continued) ---------------------------------------------------------------------- Date: 31 Dec 83 15:23:38-PST (Sat) From: harpo!floyd!clyde!akgua!psuvax!bobgian @ Ucb-Vax Subject: PSU's first AI course -- part 3/6 (First Exam) Article-I.D.: psuvax.383 [The intent and application of the following three exams was described in the previous digest issue. The exams were intended to look difficult but to be fun to take. -- KIL] ******** ARTIFICIAL INTELLIGENCE -- First Exam ******** The field of Artificial Intelligence studies the modeling of human intelligence in the hope of constructing artificial devices that display similar behavior. This exam is designed to study your ability to model artificial intelligence in the hope of improving natural devices that display similar behavior. Please read ALL the questions first, introspect on how an AI system might solve these problems, then simulate that system. (Please do all work on separate sheets of paper.) EASY PROBLEM: The rules for differentiating polynomials can be expressed as follows: IF the input is: (A * X ^ 3) + (B * X ^ 2) + (C * X ^ 1) + (D * X ^ 0) THEN the output is: (3 * A * X ^ 2) + (2 * B * X ^ 1) + (1 * C * X ^ 0) + (0 * D * X ^ -1) (where "*" indicates multiplication and "^" indicates exponentiation). Note that all letters here indicate SYMBOLIC VARIABLES (as in algebra), not NUMERICAL VALUES (as in FORTRAN). 1. Can you induce from this sample the general rule for polynomial differentiation? Express that rule in English or Mathematical notation. (The mathematicians in the group may have some difficulty here.) 2. Can you translate your "informal" specification of the differentiation rule into a precise statement of an inference rule in a Physical Symbol System? That is, define a set of objects and relations, a notation for expressing them (hint: it doesn't hurt for the notation to look somewhat like a familiar programming language which was invented to do mathematical notation), and a symbolic transformation rule that encodes the rule of inference representing differentiation. 3. Can you now IMPLEMENT your Physical Symbol System using some familiar programming language? That is, write a program which takes as input a data structure encoding your symbolic representation of a polynomial and returns a data structure encoding the representation of its derivative. (Hint as a check on infinite loops: this program can be done in six or fewer lines of code. Don't be afraid to define a utility function or two if it helps.) SLIGHTLY HARDER PROBLEM: Consider a world consisting of one block (a small wooden cubical block) standing on the floor in the middle of a room. A fly is perched on the South wall, looking North at the block. We want to represent the world as seen by the fly. In the fly's world the only thing that matters is the position of that block. Let's represent the world by a graph consisting of a single node and no links to any other nodes. Easy enough. 4. Now consider a more complicated world. There are TWO blocks, placed apart from each other along an East/West line. From the fly's point of view, Block A (the western block) is TO-THE-LEFT-OF Block B (the eastern block), and Block B has a similar relationship (TO-THE-RIGHT-OF) to Block A. Draw your symbolic representation of the situation as a graph with nodes for the blocks and labeled links for the two relationships which hold between the blocks. (Believe it or not, you have just invented the representation mechanism called a "semantic network".) 5. Now the fly moves to the northern wall, looking south. Draw the new semantic network which represents the way the blocks look to him from his new vantage point. 6. What you have diagrammed in the above two steps is a Physical Symbol System: a symbolic representation of a situation coupled with a process for making changes in the representation which correspond homomorphically with changes in the real world represented by the symbol system. Unfortunately, your symbol system does not yet have a concrete representation for this changing process. To make things more concrete, let's transform to another Physical Symbol System which can encode EXPLICITLY the representation both of the WORLD (as seen by the fly) and of HOW THE WORLD CHANGES when the fly moves. Invent a representation for your semantic network using some familiar programming language. Remember what is being modeled are OBJECTS (the blocks) and RELATIONS between the objects. Hint: you might like to use property lists, but please feel no obligations to do so. 7. Now the clincher which demonstrates the power of the idea that a physical symbol system can represent PROCESSES as well as OBJECTS and RELATIONS. Write a program which transforms the WORLD-DESCRIPTION for FLY-ON-SOUTH-WALL to WORLD-DESCRIPTION for FLY-ON-NORTH-WALL. The program should be a single function (with auxiliaries if you like) which takes two arguments, the symbol SOUTH for the initial wall and NORTH for target wall, uses a global symbol whose value is your semantic network representing the world seen from the south wall, and returns T if successful and NIL if not. As a side effect, the function should CHANGE the symbolic structure representing the world so that afterward it represents the blocks as seen by the fly from the north wall. You might care to do this in two steps: first describing in English or diagrams what is going on and then writing code to do it. 8. The world is getting slightly more complex. Now there are four blocks, A and B as before (spread apart along an East/West line), C which is ON-TOP-OF B, and D which is just to the north of (ie, in back of when seen from the south) B. Let's see your semantic network in both graphical and Lisp forms. The fly is on South wall, looking North. (Note that we mean "directly left-of" and so on. A is LEFT-OF B but has NO relation to D.) 9. Generalize the code you wrote for question 4 (if you haven't already) so that it correctly transforms the world seen by the fly from ANY of the four walls (NORTH, EAST, SOUTH, and WEST) to that seen from any other (including the same) wall. What I mean by "generalize" is don't write code that works only for the two-block or four-block worlds; code it so it will work for ANY semantic network representing a world consisting of ANY number of blocks with arbitrary relations between them chosen from the set {LEFT-OF, RIGHT-OF, IN-FRONT-OF, IN-BACK-OF, ON-TOP-OF, UNDER}. (Hint: if you are into group theory you might find a way to do this with only ONE canonical transformation; otherwise just try a few examples until you catch on.) 10. Up to now we have been assuming the fly is always right-side-up. Can you do question 6 under the assumption that the fly sometimes perches on the wall upside-down? Have your function take two extra arguments (whose values are RIGHT-SIDE-UP or UPSIDE-DOWN) to specify the fly's vertical orientation on the initial and final walls. 11. Up to now we have been modeling the WORLD AS SEEN BY THE FLY. If the fly moves, the world changes. Why is this approach no good when we allow more flies into the room and wish to model the situation from ANY of their perspectives? 12. What can be done to fix the problem you pointed out above? That is, redefine the "axioms" of your representation so it works in the "multiple conscious agent" case. (Hint: new axioms might include new names for the relations.) 13. In your new representation, the WORLD is a static object, while we have functions called "projectors" which given the WORLD and a vantage point (a symbol from the set {NORTH, EAST, SOUTH, WEST} and another from the set {RIGHT-SIDE-UP, UPSIDE-DOWN}) return a symbolic description (a "projection") of the world as seen from that vantage point. For the reasons you gave in answer to question 11, the projectors CANNOT HAVE SIDE EFFECTS. Write the projector function. 14. Now let's implement a perceptual cognitive model builder, a program that takes as input a sensory description (a symbolic structure which represents the world as seen from a particular vantage point) and a description of the vantage point and returns a "static world descriptor" which is invariant with respect to vantage point. Code up such a model builder, using for input a semantic network of the type you used in questions 6 through 10 and for output a semantic network of the type used in questions 12 and 13. (Note that this function in nothing more than the inverse of the projector from question 13.) ******** THAT'S IT !!! THAT'S IT !!! THAT'S IT !!! ******** SOME HELPFUL LISP FUNCTIONS You may use these plus anything else discussed in class. Function Argument description Return value Side effect PUTPROP ==> adds property GET ==> REMPROP ==> removes property *********************************************************************** -- Bob Giansiracusa ------------------------------ Date: 31 Dec 83 15:25:34-PST (Sat) From: harpo!floyd!clyde!akgua!psuvax!bobgian @ Ucb-Vax Subject: PSU's first AI course -- part 4/6 (Second Exam) Article-I.D.: psuvax.384 1. (20) Why are you now sitting on this side of the room? Can you cite an AI system which used a similar strategy in deciding what to do? 2. (10) Explain the difference between vs CHRONOLOGICAL and DEPENDENCY- DIRECTED backtracking. 3. (10) Compare and contrast PRODUCTION SYSTEMS and SEMANTIC NETWORKS as far as how they work, what they can represent, what type of problems are well-suited for solution using that type of knowledge representation. 4. (20) Describe the following searches in detail. In detail means: 1) How do they work?? 2) How are they related to each other?? 3) What are their advantages?? 4) What are their disadvantages?? Candidate methods: 1) Depth-first 2) Breadth-first 3) Hill-climbing 4) Beam search 5) Best-first 6) Branch-and-bound 7) Dynamic Programming 8) A* 5. (10) What are the characteristics of good generators for the GENERATE and TEST problem-solving method? 6. (10) Describe the ideas behind Mini-Max. Describe the ideas behind Alpha-Beta. How do you use the two of them together and why would you want to?? 7. (50) Godel's Incompleteness Theorem states that any consistent and sufficiently complex formal system MUST express truths which cannot be proved within the formal system. Assume that THIS theorem is true. 1. If UNPROVABLE, how did Godel prove it? 2. If PROVABLE, provide an example of a true but unprovable statement. 8. (40) Prove that this exam is unfinishable correctly; that is, prove that this question is unsolvable. 9. (50) Is human behavior governed by PREDESTINATION or FREE-WILL? How could you design a formal system to solve problems like that (that is, to reason about "non-logical" concepts)? 10. (40) Assume only ONE question on this exam were to be graded -- the question that is answered by the FEWEST number of people. How would you decide what to do? Show the productions such a system might use. 11. (100) You will be given extra credit (up to 100 points) if by 12:10 pm today you bring to the staff a question. If YOUR question is chosen, it will be asked and everybody else given 10 points for a correct answer. YOU will be given 100 points for a correct answer MINUS ONE POINT FOR EACH CORRECT ANSWER GIVEN BY ANOTHER CLASS MEMBER. What is your question? -- Bob Giansiracusa ------------------------------ Date: 31 Dec 83 15:27:19-PST (Sat) From: harpo!floyd!clyde!akgua!psuvax!bobgian @ Ucb-Vax Subject: PSU's first AI course -- part 5/6 (Third Exam) Article-I.D.: psuvax.385 1. What is the sum of the first N positive integers? That is, what is: [put here the sigma-sign notation for the sum] 2. Prove that the your answer works for any N > 0. 3. What is the sum of the squares of the first N positive integers: [put here the sigma-sign notation for the sum] 4. Again, prove it. 5. The proofs you gave (at least, if you are utilizing "traditional" mathematical background,) are based on "mathematical induction". Briefly state this principle and explain why it works. 6. If you are like most people, your definition will work only over the domain of NATURAL NUMBERS (positive integers). Can this definition be extended to work over ANY countable domain? 7. Consider the lattice of points in N-dimensional space having integer valued coordinates. Is this space countable? 8. Write a program (or express an algorithm in pseudocode) which returns the number of points in this space (the one in #7) inside an N-sphere of radius R (R is a real number > 0). 9. The domains you have considered so far are all countable. The problem solving methods you have used (if you're "normal") are based on mathematical induction. Is it possible to extend the principle of mathematical induction (and recursive programming) to NON-COUNTABLE domains? 10. If you answered #9 NO, why not? If you answered it YES, how? 11. Problems #1 and #3 require you to perform INDUCTIVE REASONING (a related but different use of the term "induction"). Discuss some of the issues involved in getting a computer to perform this process automatically. (I mean the process of generating a finite symbolic representation which when evaluated will return the partial sum for an infinite sequence.) 12. Consider the "sequence extrapolation" task: given a finite sequence of symbols, predict the next few terms of the sequence or give a rule which can generate ALL the terms of the sequence. Is this problem uniquely solvable? Why or why not? 13. If you answered #12 YES, how would you build a computer program to do so? 14. If you answered #12 NO, how could you constrain the problem to make it uniquely solvable? How would you build a program to solve the constrained problem? 15. Mankind is faced with the threat of nuclear annihilation. Is there anything the field of AI has to offer which might help avert that threat? (Don't just say "yes" or "no"; come up with something real.) 16. Assuming mankind survives the nuclear age, it is very likely that ethical issues relating to AI and the use of computers will have very much to do with the view the "person on the street" has of the human purpose and role in the Universe. In what way can AI researchers plan NOW so that these ethical issues are resolved to the benefit of the greatest number of people? 17. Could it be that our (humankind's) purpose on earth is to invent and build the species which will be the next in the evolutionary path? Should we do so? How? Why? Why not? 18. Suppose you have just discovered the "secret" of Artificial Intelligence; that is, you (working alone and in secret) have figured out a way (new hardware, new programming methodology, whatever) to build an artificial device which is MORE INTELLIGENT, BY ANY DEFINITION, BY ANY TEST WHATSOEVER, that any human being. What do you do with this knowledge? Explain the pros and cons of several choices. 19. Question #9 indicates that SO FAR all physical symbol systems have dealt ONLY with discrete domains. Is it possible to generalize the idea to continuous domains? Since many aspects of the human nervous system function on a continuous (as opposed to discrete) basis, is it possible that the invention of CONTINUOUS PHYSICAL SYMBOL SYSTEMS might provide part of the key to the "secret of intelligence"? 20. What grade do you feel you DESERVE in this course? Why? What grade do you WANT? Why? If the two differ, is there anything you want to do to reduce the difference? Why or Why Not? What is it? Why is it (or is it not) worth doing? -- Spoken: Bob Giansiracusa Bell: 814-865-9507 Bitnet: bobgian@PSUVAX1.BITNET Arpa: bobgian%psuvax1.bitnet@Berkeley CSnet: bobgian@penn-state.csnet UUCP: allegra!psuvax!bobgian USnail: Dept of Comp Sci, Penn State Univ, University Park, PA 16802 ------------------------------ End of AIList Digest ******************** 9-Jan-84 15:02:04-PST,8392;000000000001 Mail-From: LAWS created at 9-Jan-84 14:59:13 Date: Mon 9 Jan 1984 14:53-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #5 To: AIList@SRI-AI AIList Digest Tuesday, 10 Jan 1984 Volume 2 : Issue 5 Today's Topics: AI and Weather Forecasting - Request, Expert Systems - Request, Pattern Recognition & Cognition, Courses - Reaction to PSU's AI Course, Programming Lanuages - LISP Advantages ---------------------------------------------------------------------- Date: Mon 9 Jan 84 14:15:13-PST From: Ken Laws Subject: AI and Weather Forecasting I have been talking with people interested in AI techniques for weather prediction and meteorological analysis. I would appreciate pointers to any literature or current work on this subject, especially * knowledge representations for spatial/temporal reasoning; * symbolic description of weather patterns; * capture of forecasting expertise; * inference methods for estimating meteorological variables from (spatially and temporally) sparse data; * methods of interfacing symbolic knowledge and heuristic reasoning with numerical simulation models; * any weather-related expert systems. I am aware of some recent work by Gaffney and Racer (NBS Trends and Applications, 1983) and by Taniguchi et al. (6th Pat. Rec., 1982), but I have not been following this field. A bibliography or guide to relevant literature would be welcome. -- Ken Laws ------------------------------ Date: 5 January 1984 13:47 est From: RTaylor.5581i27TK at RADC-MULTICS Subject: Expert Systems Info Request Hi, y'all...I have the names (hopefully, correct) of four expert systems/tools/environments (?). I am interested in the "usual": that is, general info, who to contact, feedback from users, how to acquire (if we want it), etc. The four names I have are: RUS, ALX, FRL, and FRED. Thanks. Also, thanks to those who provided info previously...I have info (similar to that requested above) on about 15 other systems/tools/environments...some of the info is a little sketchy! Roz (aka: rtaylor at radc-multics) ------------------------------ Date: 3 Jan 84 20:38:52-PST (Tue) From: decvax!genrad!mit-eddie!rh @ Ucb-Vax Subject: Re: Loop detection and classical psychology Article-I.D.: mit-eddi.1114 One of the truly amazing things about the human brain is that its pattern recognition capabilities seem limitless (in extreme cases). We don't even have a satisfactory way to describe pattern recognition as it occurs in our brains. (Well, maybe we have something acceptable at a minimum level. I'm always impressed by how well dollar-bill changers seem to work.) As a friend of mine put it, "the brain immediately rejects an infinite number of wrong answers," when working on a problem. Randwulf (Randy Haskins); Path= genrad!mit-eddie!rh ------------------------------ Date: Fri 6 Jan 84 10:11:01-PST From: Ron Brachman Subject: PSU's First AI Course Wow! I actually think it's kind of neat (but, of course, very wacko). I particularly like making people think about the ethical and philosphical considerations at the same time as their thinking about minimax, etc. ------------------------------ Date: Wed 4 Jan 84 17:23:38-PST From: Richard Treitel Subject: Re: AIList Digest V2 #1 [in response to Herb Lin's questions] Well, 2 more or less answers 1. One of the main reasons why Lisp and not C is the language of many people's choice for AI work is that you can easily cons up at run time a piece of data which "is" the next action you are going to take. In most languages you are restricted to choosing from pre-written actions, unless you include some kind of interpreter right there in your AI program. Another reason is that Lisp has all sorts of extensibility. As for 3, the obvious response is that in Pascal control has to be routed to an IF statement before it can do any good, whereas in a production system, control automatically "goes" to any production that is applicable. This is highly over-simplified and may not be the answer you were looking for. - Richard ------------------------------ Date: Friday, 6 Jan 1984 13:10-PST From: narain@rand-unix Subject: Reply to Herb Lin: Why is Lisp good for AI? A central issue in AI is knowledge representation. Experimentation with a new KR scheme often involves defining a new language. Often, definitions and meanings of new languages are conceived of naturally in terms of recursive (hierarchical) structures. For instance, many grammars of English- like frontends are recursive, so are production system definitions, so are theorem provers. The abstract machinery underlying Lisp, the Lambda Calculus, is also inherently recursive, yet very simple and powerful. It involves the notion of function application to symbolic expressions. Functions can themselves be symbolic expressions. Symbolic expressions provide a basis for SIMPLE implementation and manipulation of complex data/knowledge/program structures. It is therefore possible to easily interpret new language primitives in terms of Lisp's already very high level primitives. Thus, Lisp is a great "machine language" for AI. The usefulness of a well understood, powerful, abstract machinery of the implementation language is probably more obvious when we consider Prolog. The logical interpretation of Prolog programs helps considerably in their development and verification. Logic is a convenient specification language for a lot of AI, and it is far easier to 'compile' those specifications into a logic language like Prolog than into Pascal. For instance, take natural language front ends implemented in DCGs or database/expert-system integrity and redundancy constraints. The fact that programs can be considered as data is not true only of Lisp. Even in Pascal you can analyze a Pascal program. The nice thing in Lisp, however, is that because of its few (but very powerful) primitives, programs tend to be simply structured and concise (cf. claims in recent issues of this bulletin that Lisp programs were much shorter than Pascal programs). So naturally it is simpler to analyze Lisp programs in Lisp than it is to analyze Pascal programs in Pascal. Of course, Lisp environments have evolved for over two decades and contribute no less to its desirability for AI. Some of the nice features include screen-oriented editors, interactiveness, debugging facilities, and an extremely simple syntax. I would greatly appreciate any comments on the above. Sanjai Narain Rand. ------------------------------ Date: 6 Jan 84 13:20:29-PST (Fri) From: ihnp4!mit-eddie!rh @ Ucb-Vax Subject: Re: Herb Lin's questons on LISP etc. Article-I.D.: mit-eddi.1129 One of the problems with LISP, however, is it does not force one to subscribe the code of good programming practices. I've found that the things I have written for my bridge-playing program (over the last 18 months or so) have gotten incredibly crufty, with some real brain-damaged patches. Yeah, I realize it's my fault; I'm not complaining about it because I love LISP, I just wanted to mention some of the pitfalls for people to think about. Right now, I'm in the process of weeding out the cruft, trying to make it more clearly modular, decrease the number of similar functions and so on. Sigh. Randwulf (Randy Haskins); Path= genrad!mit-eddie!rh ------------------------------ Date: 7 January 1984 15:08 EST From: Herb Lin Subject: my questions of last Digest on differences between PASCAL and LISP So many people replied that I send my thanks to all via the list. I very much appreciate the time and effort people put into their comments. ------------------------------ End of AIList Digest ******************** 10-Jan-84 10:13:10-PST,13872;000000000001 Mail-From: LAWS created at 10-Jan-84 10:10:10 Date: Tue 10 Jan 1984 09:48-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #6 To: AIList@SRI-AI AIList Digest Tuesday, 10 Jan 1984 Volume 2 : Issue 6 Today's Topics: Humor, Seminars - Programming Styles & ALICE & 5th Generation, Courses - Geometric Data Structures & Programming Techniques & Linguistics ---------------------------------------------------------------------- Date: Mon, 9 Jan 84 08:45 EST From: MJackson.Wbst@PARC-MAXC.ARPA Subject: An AI Joke Last week a cartoon appeared in our local (Rochester NY) paper. It was by a fellow named Toles, a really excellent editorial cartoonist who works out of, of all places, Buffalo: Panel 1: [medium view of the Duckburg Computer School building. A word balloon extends from one of the windows] "A lot of you wonder why we have to spend so much time studying these things." Panel 2: [same as panel 1] "It so happens that they represent a lot of power. And if we want to understand and control that power, we have to study them." Panel 3: [interior view of a classroom full of personal computers. At right, several persons are entering. At left, a PC speaks] ". . .so work hard and no talking. Here they come." Tickler (a mini-cartoon down in the corner): [a lone PC speaks to the cartoonist] "But I just HATE it when they touch me like that. . ." Mark ------------------------------ Date: Sat, 7 Jan 84 20:02 PST From: Vaughan Pratt Subject: Imminent garbage collection of Peter Coutts. :=) [Here's another one, reprinted from the SU-SCORE bboard. -- KIL] Les Goldschlager is visiting us on sabbatical from Sydney University, and stayed with us while looking for a place to stay. We belatedly pointed him at Peter Coutts, which he immediately investigated and found a place to stay right away. His comment was that no pointer to Peter Coutts existed in any of the housing assistance services provided by Stanford, and that therefore it seemed likely that it would be garbage collected soon. -v ------------------------------ Date: 6 January 1984 23:48 EST From: Steven A. Swernofsky Subject: Seminar on Programming Styles in AI DATE: Thursday, January 12, 1984 TIME: 3.45 p.m. Refreshments 4.00 p.m. Lecture PLACE: NE43-8th Floor, AI Playroom PROGRAMMING STYLES IN ARTIFICIAL INTELLIGENCE Herbert Stoyan University of Erlangen, West Germany ABSTRACT Not much is clear about the scientific methods used in AI research. Scientific methods are sets of rules used to collect knowledge about the subject being researched. AI is an experimental branch of computer science which does not seem to use established programming methods. In several works on AI we can find the following method: 1. develop a new convenient programming style 2. invent a new programming language which supports the new style (or embed some appropriate elements into an existing AI language, such as LISP) 3. implement the language (interpretation as a first step is typically less efficient than compilation) 4. use the new programming style to make things easier. A programming style is a way of programming guided by a speculative view of a machine which works according to the programs. A programming style is not a programming method. It may be detected by analyzing the text of a completed program. In general, it is possible to program in one programming language according to the principles of various styles. This is true in spite of the fact that programming languages are usually designed with some machine model (and therefore with some programming style) in mind. We discuss some of the AI programming styles. These include operator-oriented, logic-oriented, function-oriented, rule- oriented, goal-oriented, event-oriented, state-oriented, constraint- oriented, and object-oriented. (We shall not however discuss the common instruction-oriented programming style). We shall also give a more detailed discussion of how an object-oriented programming style may be used in conventional programming languages. HOST: Professor Ramesh Patil ------------------------------ Date: Mon 9 Jan 84 14:09:07-PST From: Laws@SRI-AI Subject: SRI Talk on ALICE, 1/23, 4:30pm, EK242 ALICE: A parallel graph-reduction machine for declarative and other languages. SPEAKER - John Darlington, Department of Computing, Imperial College, London WHEN - Monday, January 23, 4:30pm WHERE - AIC Conference Room, EK242 [This is an SRI AI Center talk. Contact Margaret Olender at MOLENDER@SRI-AI or 859-5923 if you would like to attend. -- KIL] ABSTRACT Alice is a highly parallel-graph reduction machine being designed and built at Imperial College. Although designed for the efficient execution of declarative languages, such as functional or logic languages, ALICE is general purpose and can execute sequential languages also. This talk will describe the general model of computation, extended graph reduction, that ALICE executes, outline how different languages can be supported by this model, and describe the concrete architecture being constructed. A 24-processor prototype is planned for early 1985. This will give a two-orders-of-magnitude improvement over a VAX 11/750 for derclarative languages. ALICE is being constructed out of two building blocks, a custom-designed switching chip and the INMOS transputer. So far, compilers for a functional language, several logic languages, and LISP have been constructed. ------------------------------ Date: 9 Jan 1984 1556-PST From: OAKLEY at SRI-CSL Subject: SRI 5th Generation Talk Japan's 5th Generation Computer Project: Past, Present, and Future -- personal observations by a researcher of ETL (ElectroTechnical Laboratory) Kokichi FUTATSUGI Senior Research Scientist, ETL International Fellow, SRI-CSL Talk on January 24, l984, in conference room EL369 at 10:00am. [This is an SRI Computer Science Laboratory talk. Contact Mary Oakley at OAKLEY@SRI-AI or 859-5924 if you would like to attend. -- KIL] 1 Introduction * general overview of Japan's research activities in computer science and technology * a personal view 2 Past -- pre-history of ICOT (the Institute of New Generation ComputerTechnology) * ETL's PIPS project * preliminary research and study activities * the establishment of ICOT 3 Present -- present activities * the organization of ICOT * research activities inside ICOT * research activities outside ICOT 4 Future -- ICOT's plans and general overview * ICOT's plans * relations to other research activities * some comments ------------------------------ Date: Thu 5 Jan 84 16:41:57-PST From: Martti Mantyla Subject: Data Structures & Algorithms for Geometric Problems [Reprinted from the SU-SCORE bboard.] NEW COURSE: EE392 DATA STRUCTURES AND ALGORITHMS FOR GEOMETRIC PROBLEMS Many problems arising in science and engineering deal with geometric information. Engineering design is most often spatial activity, where a physical shape with certain desired properties must be created. Engineering analysis also uses heavily information on the geometric form of the object. The seminar Data Structures and Algorithms for Geometric Problems deals with problems related to representing and processing data on the geometric shape of an object in a computer. It will concentrate on practically interesting solutions to tasks such as - representation of digital images, - representation of line figures, - representation of three-dimensional solid objects, and - representation of VLSI circuits. The point of view taken is hence slightly different from a "hard-core" Computational Geometry view that puts emphasis on asymptotic computational complexity. In practice, one needs solutions that can be implemented in a reasonable time, are efficient and robust enough, and can support an interesting scope of applications. Of growing importance is to find representations and algorithms for geometry that are appropriate for implementation in special hardware and VLSI in particular. The seminar will be headed by Dr. Martti Mantyla (MaM) Visiting Scholar CSL/ERL 405 7-9310 MANTYLA@SU-SIERRA.ARPA who will give intruductory talks. Guest speakers of the seminar include well-known scientists and practitioners of the field such as Dr. Leo Guibas and Dr. John Ousterhout. Classes are held on Tuesdays, 2:30 - 3:30 in ERL 126 First class will be on 1/10. The seminar should be of interest to CS/EE graduate students with research interests in computer graphics, computational geometry, or computer applications in engineering. ------------------------------ Date: 6 Jan 1984 1350-EST From: KANT at CMU-CS-C.ARPA Subject: AI Programming Techniques Course [Reprinted from the CMUC bboard.] Announcing another action-packed AI mini-course! Starting soon in the 5409 near you. This course covers a variety of AI programming techniques and languages. The lectures will assume a background equivalent to an introductory AI course (such as the undergraduate course 15-380/381 or the graduate core course 15-780.) They also assume that you have had at least a brief introduction to LISP and a production-system language such as OPS5. 15-880 A, Artificial Intelligence Programming Techniques MW 2:30-3:50, WeH 5409 T Jan 10 (Brief organizational meeting only) W Jan 11 LISP: Basic Pattern Matching (Carbonell) M Jan 16 LISP: Deductive Data Bases (Steele) W Jan 18 LISP: Basic Control: backtracking, demons (Steele) M Jan 23 LISP: Non-Standard Control Mechanisms (Carbonell) W Jan 25 LISP: Semantic Grammar Interpreter (Carbonell) M Jan 30 LISP: Case-Frame interpreter (Hayes) W Feb 1 PROLOG I (Steele) M Feb 6 PROLOG II (Steele) W Feb 8 Reason Maintenance and Comparison with PROLOG (Steele) M Feb 13 AI Programming Environments and Hardware I (Fahlman) W Feb 15 AI Programming Environments and Hardware II (Fahlman) M Feb 20 Schema Representation Languages I (Fox) W Feb 22 Schema Representation Languages II (Fox) W Feb 29 User-Interface Issues in AI (Hayes) M Mar 5 Efficient Game Playing and Searching (Berliner) W Mar 7 Production Systems: Basic Programming Techniques (Kant) M Mar 12 Production Systems: OPS5 Programming (Kant) W Mar 14 Efficiency and Measurement in Production Systems (Forgy) M Mar 16 Implementing Diagnostic Systems as Production Systems (Kahn) M Mar 26 Intelligent Tutoring Systems: GRAPES and ACT Implementations (Anderson) W Mar 28 Explanation and Knowledge Acquisition in Expert Systems (McDermott) M Apr 2 A Production System for Problem Solving: SOAR2 (Laird) W Apr 4 Integrating Expert-System Tools with SRL (KAS, PSRL, PDS) (Rychener) M Apr 9 Additional Expert System Tools: EMYCIN, HEARSAY-III, ROSIE, LOOPS, KEE (Rosenbloom) W Apr 11 A Modifiable Production-System Architecture: PRISM (Langley) M Apr 16 (additional topics open to negotiation) ------------------------------ Date: 9 Jan 1984 1238:48-EST From: Lori Levin Subject: Linguistics Course [Reprinted from the CMUC bboard.] NATURAL LANGUAGE SYNTAX FOR COMPUTER SCIENTISTS FRIDAYS 10:00 AM - 12:00 4605 Wean Hall Lori Levin Richmond Thomason Department of Linguistics University of Pittsburgh This is an introduction to recent work in generative syntax. The course will deal with the formalism of some of the leading syntactic theories as well as with methodological issues. Computer scientists find the formalism used by syntacticians easy to learn, and so the course will begin at a fairly advanced level, though no special knowledge of syntax will be presupposed. We will begin with a sketch of the "Standard Theory," Chomsky's approach of the mid-60's from which most of the current theories have evolved. Then we will examine Government-Binding Theory, the transformational approach now favored at M.I.T. Finally, we will discuss in more detail two nontransformational theories that are more computationally tractable and have figured in joint research projects involving linguists, psychologists, and computer scientists: Lexical-Functional Grammar and Generalized Context-Free Phrase Structure Grammar. ------------------------------ End of AIList Digest ******************** 16-Jan-84 22:15:32-PST,15986;000000000001 Mail-From: LAWS created at 16-Jan-84 22:13:48 Date: Mon 16 Jan 1984 21:55-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #7 To: AIList@SRI-AI AIList Digest Tuesday, 17 Jan 1984 Volume 2 : Issue 7 Today's Topics: Production Systems - Requests, Expert Systems - Software Debugging Aid, Logic Programming - Prolog Textbooks & Disjunction Problem, Alert - Fermat's Last Theorem Proven?, Seminars - Mulitprocessing Lisp & Lisp History, Conferences - Logic Programming Discount & POPL'84, Courses - PSU's First AI Course & Net AI Course ---------------------------------------------------------------------- Date: 11 Jan 1984 1151-PST From: Jay Subject: Request for production systems I would like pointers to free or public domain production systems (running on Tops-20, Vax-Unix, or Vax-Vms) both interpreters (such as ross) and systems built up on them (such as emycin). I am especially interested in Rosie, Ross, Ops5, and Emycin. Please reply directly to me. j' ARPA: jay@eclc ------------------------------ Date: Thu 12 Jan 84 12:13:20-MST From: Stanley T. Shebs Subject: Taxonomy of Production Systems I'm looking for info on a formal taxonomy of production rule systems, sufficiently precise that it can distinguish OPS5 from YAPS, but also say that they're more similar than either of them is to Prolog. The only relevant material I've seen is the paper by Davis & King in MI 8, which characterizes PSs in terms of syntax, complexity of LHS and RHS, control structure, and "programmability" (seems to mean meta-rules). This is a start, but too vague to be implemented. A formal taxonomy should indicate where "holes" exist, that is, strange designs that nobody has built. Also, how would Georgeff's (Stanford STAN-CS-79-716) notion of "controlled production systems" fit in? He showed that CPSs are more general than PSs, but then one can also show that any CPS can be represented by some ordinary PS. I'm particularly interested in formalization of the different control strategies - are text order selection (as in Prolog) and conflict resolution (as in OPS5) mutually exclusive, or can they be intermixed (perhaps using text order to find 5 potential rules, then conflict resolution to choose among the 5). Presumably a sufficiently precise taxonomy could answer these sorts of questions. Has anyone looked at these questions? stan shebs ------------------------------ Date: 16 Jan 84 19:13:21 PST (Monday) From: Ron Newman Subject: Expert systems for software debugging? Debugging is a black art, not at all algorithmic, but almost totally heuristic. There is a lot of expert knowledge around about how to debug faulty programs, but it is rarely written down or systemetized. Usually it seems to reside solely in the minds of a few "debugging whizzes". Does anyone know of an expert system that assists in software debugging? Or any attempts (now or in the past) to produce such an expert? /Ron ------------------------------ Date: 12 Jan 84 20:43:31-PST (Thu) From: harpo!floyd!clyde!akgua!sb1!mb2c!uofm-cv!lah @ Ucb-Vax Subject: prolog reference Article-I.D.: uofm-cv.457 Could anybody give some references to good introductory book on prolog? ------------------------------ Date: 14 Jan 84 14:50:57-PST (Sat) From: decvax!duke!mcnc!unc!bts @ Ucb-Vax Subject: Re: prolog reference Article-I.D.: unc.6594 There's only one introductory book I know of, that's Clocksin and Mellish's "Programming in Prolog", Springer-Verlag, 1981. It's a silver paperback, probably still under $20.00. For more information on the language, try Clark and Tarnlund's "Logic Programming", Academic Press, 1982. It's a white hard- back, with an elephant on the cover. The papers by Bruynooghe and by Mellish tell a lot about Prolog inplementation. Bruce Smith, UNC-Chapel Hill decvax!duke!unc!bts (USENET) bts.unc@CSnet-Relay (lesser NETworks) ------------------------------ Date: 13 Jan 84 8:11:49-PST (Fri) From: hplabs!hao!seismo!philabs!sbcs!debray @ Ucb-Vax Subject: re: trivial reasoning problem? Article-I.D.: sbcs.572 Re: Marcel Schoppers' problem: given two lamps A and B, such that: condition 1) at least one of them is on at any time; and condition 2) if A is on then B id off, we are to enumerate the possible configurations without an exhaustive generate-and-test strategy. The following "pure" Prolog program that will generate the various configurations without exhaustively generating all possible combinations: config(A, B) :- cond1(A, B), cond2(A, B). /* both conditions must hold */ cond1(1, _). /* at least one is on an any time ... condition 1 above */ cond1(_, 1). cond2(1, 0). /* if A is on then B is off */ cond2(0, _). /* if A is off, B's value is a don't care */ executing Prolog gives: | ?- config(A, B). A = 1 B = 0 ; A = 0 B = 1 ; no | ?- halt. [ Prolog execution halted ] Tracing the program shows that the configuration "A=0, B=0" is not generated. This satisfies the "no-exhaustive-listing" criterion. Note that attempting to encode the second condition above using "not" will be both (1) not pure Horn Clause, and (2) using exhaustive generation and filtering. Saumya Debray Dept. of Computer Science SUNY at Stony Brook {floyd, bunker, cbosgd, mcvax, cmcl2}!philabs! \ Usenet: sbcs!debray / {allegra, teklabs, hp-pcd, metheus}!ogcvax! CSNet: debray@suny-sbcs@CSNet-Relay [Several other messages discussing this problem and suggesting Prolog code were printed in the Prolog Digest. Different writers suggested very different ways of structuring the problem. -- KIL] ------------------------------ Date: Fri 13 Jan 84 11:16:21-CST From: Clive Dawson Subject: Fermat's Last Theorem Proven? [Reprinted from the UTEXAS-20 bboard.] There was a report last night on National Public Radio's All Things Considered about a British mathematician named Arnold Arnold who claims to have developed a new technique for dealing with multi-variable, high-dimensional spaces. The method apparently makes generation of large prime numbers very easy, and has applications in genetics, the many-body problem, orbital mechanics, etc. Oh yeah, the proof to Fermat's Last Theorem falls out of this as well! The guy apparently has no academic credentials, and refuses to publish in the journals because he's interested in selling his technique. There was another mathematician named Jeffrey Colby who had been allowed to examine Arnold's work on the condition he didn't disclose anything. He claims the technique is all it's claimed to be, and shows what can be done when somebody starts from pure ignorance not clouded with some of the preconceptions of a formal mathematical education. If anybody hears more about this, please pass it along. Clive ------------------------------ Date: 12 Jan 84 2350 PST From: Rod Brooks Subject: Next week's CSD Colloquium. [Reprinted from the SU-SCORE bboard.] Dr. Richard P. Gabriel, Stanford CSD ``Queue-based Multi-processing Lisp'' 4:30pm Terman Auditorium, Jan 17th. As the need for high-speed computers increases, the need for multi-processors will be become more apparent. One of the major stumbling blocks to the development of useful multi-processors has been the lack of a good multi-processing language---one which is both powerful and understandable to programmers. Among the most compute-intensive programs are artificial intelligence (AI) programs, and researchers hope that the potential degree of parallelism in AI programs is higher than in many other applications. In this talk I will propose a version of Lisp which is multi-processed. Unlike other proposed multi-processing Lisps, this one will provide only a few very powerful and intuitive primitives rather than a number of parallel variants of familiar constructs. The talk will introduce the language informally, and many examples along with performance results will be shown. ------------------------------ Date: 13 January 1984 07:36 EST From: Kent M Pitman Subject: What is Lisp today and how did it get that way? [Reprinted from the MIT-MC bboard.] Modern Day Lisp Time: 3:00pm Date: Wednesdays and Fridays, 18-27 January Place: 8th Floor Playroom The Lisp language has changed significantly in the past 5 years. Modern Lisp dialects bear only a superficial resemblance to each other and to their common parent dialects. Why did these changes come about? Has progress been made? What have we learned in 5 hectic years of rapid change? Where is Lisp going? In a series of four lectures, we'll be surveying a number of the key features that characterize modern day Lisps. The current plan is to touch on at least the following topics: Scoping. The move away from dynamic scoping. Namespaces. Closures, Locales, Obarrays, Packages. Objects. Actors, Capsules, Flavors, and Structures. Signals. Errors and other unusual conditions. Input/Output. From streams to window systems. The discussions will be more philosophical than technical. We'll be looking at several Lisp dialects, not just one. These lectures are not just something for hackers. They're aimed at just about anyone who uses Lisp and wants an enhanced appreciation of the issues that have shaped its design and evolution. As it stands now, I'll be giving all of these talks, though there is some chance there will be some guest lecturers on selected topics. If you have questions or suggestions about the topics to be discussed, feel free to contact me about them. Kent Pitman (KMP@MC) NE43-826, x5953 ------------------------------ Date: Wed 11 Jan 84 16:55:02-PST From: PEREIRA@SRI-AI.ARPA Subject: IEEE Logic Programming Symposium (update) 1984 International Symposium on Logic Programming Student Registration Rates In our original symposium announcements, we failed to offer a student registration rate. We would like to correct that situation now. Officially enrolled students may attend the symposium for the reduced rate of $75.00. This rate includes the symposium itself (all three days) and one copy of the symposium proceedings. It does not include the tutorial, the banquet, or cocktail parties. It does however, include the Casino entertainment show. Questions and requests for registration forms by US mail to: Doug DeGroot Fernando Pereira Program Chairman SRI International IBM Research or 333 Ravenswood Ave. P.O. Box 218 Menlo Park, CA 94025 Yorktown Heights, NY 10598 (415) 859-5494 (914) 945-3497 or by net mail to: PEREIRA@SRI-AI (ARPANET) ...!ucbvax!PEREIRA@SRI-AI (UUCP) ------------------------------ Date: Tue 10 Jan 84 15:54:09-MST From: Subra Subject: *** P O P L 1984 --- Announcement *** ******************************* POPL 1984 ********************************* ELEVENTH ANNUAL ACM SIGACT/SIGPLAN SYMPOSIUM ON PRINCIPLES OF PROGRAMMING LANGUAGES *** POPL 1984 will be held in Salt Lake City, Utah January 15-18. **** (The skiing is excellent, and the technical program threatens to match it!) For additional details, please contact Prof. P. A. Subrahmanyam Department of Computer Science University of Utah Salt Lake City, Utah 84112. Phone: (801)-581-8224 ARPANET: Subrahmanyam@UTAH-20 (or Subra@UTAH-20) ------------------------------ Date: 12 Jan 84 4:51:51-PST (Thu) From: Subject: Re: PSU's First AI Course - Comment Article-I.D.: sjuvax.108 I would rather NOT get into social issues of AI: there are millions of forums for that (and I myself have all kinds of feelings and reservations on the issue, including Vedantic interpretations), so let us keep this one technical, please. ------------------------------ Date: 13 Jan 84 11:42:21-PST (Fri) From: Subject: Net AI course -- the communications channel Article-I.D.: psuvax.413 Responses so far have strongly favored my creating a moderated newsgroup as a sub to net.ai for this course. Most were along these lines: From: ukc!srlm (S.R.L.Meira) I think you should act as the moderator, otherwise there would be too much noise - in the sense of unordered information and discussions - and it could finish looking like just another AI newsgroup argument. Anybody is of course free to post whatever they want if they feel the thing is not coming out like they want. Also, if the course leads to large volume, many net.ai readers (busy AI professionals rather than students) might drop out of net.ai. For a contrasting position: From: cornell!nbires!stcvax!lat I think the course should be kept as a newsgroup. I don't think it will increase the nation-wide phone bills appreciably beyond what already occurs due to net.politics, net.flame, net.religion and net.jokes. So HERE's how I'll try to keep EVERYBODY happy ... :-) ... a "three-level" communication channel. 1: a "free-for-all" via mail (or possibly another newsgroup), 2: a moderated newsgroup sub to net.ai, 3: occasional abstracts, summaries, pointers posted to net.ai and AIList. People can then choose the extent of their involvement and set their own "bull-rejection threshold". (1) allows extensive involvement and flaming, (2) would be the equivalent of attending a class, and (3) makes whatever "good stuff" evolves from the course available to all others. The only remaining question: should (1) be done via a newsgroup or mail? Please send in your votes -- I'll make the final decision next week. Now down to the REALLY BIG decisions: names. I suggest "net.ai.cse" for level (2). The "cse" can EITHER mean "Computer Science Education" or abbreviate "course". For level (1), how about "net.ai.ffa" for "free-for-all", or .raw, or .disc, or .bull, or whatever. Whatever I create gets zapped at end of course (June), unless by then it has taken on a life of its own. -- Bob [PS to those NOT ON USENET: please mail me your address for private mailings -- and indicate which of the three "participation levels" best suits your tastes.] Bob Giansiracusa (Dept of Computer Science, Penn State Univ, 814-865-9507) UUCP: bobgian@psuvax.UUCP -or- allegra!psuvax!bobgian Arpa: bobgian@PSUVAX1 -or- bobgian%psuvax1.bitnet@Berkeley Bitnet: bobgian@PSUVAX1.BITNET CSnet: bobgian@penn-state.csnet USnail: 333 Whitmore Lab, Penn State Univ, University Park, PA 16802 ------------------------------ End of AIList Digest ******************** 17-Jan-84 22:57:32-PST,14885;000000000001 Mail-From: LAWS created at 17-Jan-84 22:52:40 Date: Tue 17 Jan 1984 22:43-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #8 To: AIList@SRI-AI AIList Digest Wednesday, 18 Jan 1984 Volume 2 : Issue 8 Today's Topics: Programming Languages - Lisp for IBM, Intelligence - Subcognition, Seminar - Knowledge-Based Design Environment ---------------------------------------------------------------------- Date: Thu 12 Jan 84 15:07:55-PST From: Jeffrey Mogul Subject: Re: lisp for IBM [Reprinted from the SU-SCORE bboard.] Does anyone know of LISP implementations for IBM 370--3033--308x? Reminds me of an old joke: How many IBM machines does it take to run LISP? Answer: two -- one to send the input to the PDP-10, one to get the output back. ------------------------------ Date: Thursday, 12 Jan 1984 21:28-PST From: Steven Tepper Subject: Re: lisp for IBM [Reprinted from the SU-SCORE bboard.] Well, I used Lisp on a 360 once, but I certainly wouldn't recommend that version (I don't remember where it came from anyway -- the authors were probably so embarrassed they wanted to remain anonymous). It was, of course, a batch system, and its only output mode was "uglyprint" -- no matter what the input looked like, the output would just be printed 120 columns to a line. ------------------------------ Date: Fri 13 Jan 84 06:55:00-PST From: Ethan Bradford Subject: LISP (INTERLISP) for IBM [Reprinted from the SU-SCORE bboard.] Chris Ryland (CPR@MIT-XX) sent out a query on this before and he got back many good responses (he gave me copies). The main thing most people said is that a version was developed at Uppsula in Sweden in the 70's. One person gave an address to write to, which I transcribe here with no gua- rantees of currentness: Klaus Appel UDAC Box 2103 750 02 Uppsala Sweden Phone: 018-11 13 30 ------------------------------ Date: 13 Jan 84 0922 PST From: Jussi Ketonen Subject: Lisp for IBM machines [Reprinted from the SU-SCORE bboard.] Standard Lisp runs quite well on the IBM machines. The folks over at IMSSS on campus know all about it -- they have written several large theorem proving/CAI programs for that environment. ------------------------------ Date: 11 January 1984 06:27 EST From: Jerry E. Pournelle Subject: intelligence and genius I should have thought that if you can make a machine more or less intelligent; and make another machine ABLE TO RECOGNIZE GENIUS (it need not itself be able to "be" or "have" genius) then the "genius machine " problem is probably solved: have the somewhat intelligent one generate lots of ideas, with random factors thrown in, and have the second "recognizing" machine judge the products. Obviously they could be combined into one machine. ------------------------------ Date: Sunday, 15 January 1984, 00:18-EST From: Marek W. Lugowski Subject: Adrressing DRogers' questions (at last) + on subcogniton DROGERS (c. November '84): I have a few questions I would like to ask, some (perhaps most) essentially unanswerable at this time. Appologies in advance for rashly attempting to answer at this time. - Should the initially constructed subcognitive systems be "learning" systems, or should they be "knowledge-rich" systems? That is, are the subcognitive structures implanted with their knowledge of the domain by the programmer, or is the domain presented to the system in some "pure" initial state? Is the approach to subcognitive systems without learning advisable, or even possible? I would go off on a limb and claim that attempting wholesale "learning" first (whatever that means these days) is silly. I would think one would first want to spike the system with hell of a lot of knowledge (e.g., Dughof's "Slipnet" of related concepts whose links are subject to cummulative, partial activation which eventually makes the nodes so connected highly relevant and therefore taken into consideration by the system). To repeat Minsky (and probably, most of the AI folk: one can only learn if one already almost knows it). - Assuming human brains are embodiments of subcognitive systems, then we know how they were constructed: a very specific DNA blueprint controlling the paths of development possible at various times, with large assumptions as to the state of the intellectual environment. This grand process was created by trial-and-error through the process of evolution, that is, essentially random chance. How much (if any) of the subcognitive system must be created essentially by random processes? If essentially all, then there are strict limits as to how the problem should be approached. This is an empirical question. If my now-attempted implementation of the Copycat Project (which uses the Slipnet described above) [forthcoming MIT AIM #755 by Doug Hofstadter] will converge nicely, with trivial tweaking, I'll be inclined to hold that random processes can indeed do most of the work. Such is my current, unfounded, belief. On the other hand, a failure will not debunk my position--I could always have messed up implementationally and made bad guesses which "threw" the system out of its potential convergence. - Which processes of the human brain are essentially subcognitive in construction, and which use other techniques? Is this balance optimal? Which structures in a computational intelligence would be best approached subcognitively, and which by other methods? Won't even touch the "optimal" question. I would guess any process involving a great deal of fan-in would need to be subcognitive in nature. This is argued from efficiency. For now, and for want of better theories, I'd approach ALL brain functions using subcognitive models. The alternative to this at present means von Neumannizing the brain, an altogether quaint thing to do... - How are we to judge the success of a subcognitive system? The problems inherent in judging the "ability" of the so-called expert systems will be many times worse in this area. Without specific goal criteria, any results will be unsatisfying and potentially illusory to the watching world. Performance and plausibility (in that order) ought to be our criteria. Judging performance accurately, however, will continue to be difficult as long as we are forced to use current computer architectures. Still, if a subcognitive system converges at all on a LispM, there's no reason to damn its performance. Plausibility is easier to demonstrate; one needs to keep in touch with the neurosciences to do that. - Where will thinking systems REALLY be more useful than (much refined) expert systems? I would guess that for many (most?) applications, expertise might be preferable to intelligence. Any suggestions about fields for which intelligent systems would have a real edge over (much improved) expert systems? It's too early (or, too late?!) to draw such clean lines. Perhaps REAL thinking and expertise are much more intertwined than is currently thought. Anyway, there is nothing to be gained by pursuing that line of questioning before WE learn how to explicitly organize knowledge better. Over all, I defend pursuing things subcognitively for these reasons: -- Not expecting thinking to be a cleanly organized, top-down driven activity is minimizing one's expectations. Compare thinking with such activities as cellular automata (e.g., The Game of Life) or The Iterated Pairwise Prisoner's Dilemma Game to convince yourself of the futility of top-down modeling where local rules and their iterated interactions are very successful at concisely describing the problem at hand. No reason to expect the brain's top-level behavior to be any easier to explain away. -- AI has been spending a lot of itself on forcing a von Neumannian interpretation on the mind. At CMU they have it down to an art, with Simon's "symbolic information processing" the nowadays proverbial Holy Grail. With all due respect, I'd like to see more research devoted to modeling various alleged brain activities with high degree of parallelism and probabilistic interaction, systems where "symbols" are not givens but intricately invovled intermediates of computation. -- It has not been done carefully before and I want at least a thesis out of it. -- Marek ------------------------------ Date: Mon, 16 Jan 1984 12:40 EST From: GLD%MIT-OZ@MIT-MC.ARPA Subject: minority report From: MAREK To repeat Minsky (and probably, most of the AI folk: one can only learn if one already almost knows it). By "can only learn if..." do you mean "can't >soon< learn unless...", or do you mean "can't >ever< learn unless..."? If you mean "can't ever learn unless...", then the statement has the Platonic implication that a person at infancy must "already almost know" everything she is ever to learn. This can't be true for any reasonable sense of "almost know". If you mean "can't soon learn unless...", then by "almost knows X", do you intend: o a narrow interpretation, by which a person almost knows X only if she already has knowledge which is a good approximation to understanding X-- eg, she can already answer simpler questions about X, or can answer questions about X, but with some confusion and error; or o a broader interpretation, which, in addition to the above, counts as "almost knowing X" a situation where a person might be completely in the dark about X-- say, unable to answer any questions about X-- but is on the verge of becoming an instant expert on X, say by discovering (or by being told of) some easy-to-perform mapping which reduces X to some other, already-well-understood domain. If you intend the narrow interpretation, then the claim is false, since people can (sometimes) soon learn X in the manner described in the broad- interpretation example. But if you intend the broad interpretation, then the statement expands to "one can't soon learn X unless one's current knowledge state is quickly transformable to include X"-- which is just a tautology. So, if this analysis is right, the statement is either false, or empty. ------------------------------ Date: Mon, 16 Jan 1984 20:09 EST From: MAREK%MIT-OZ@MIT-MC.ARPA Subject: minority report From: MAREK To repeat Minsky (and probably, most of the AI folk): one can only learn if one already almost knows it. From: GLD By "can only learn if..." do you mean..."can't >ever< learn unless..."? If you mean "can't ever learn unless...", then the statement has the Platonic implication that a person at infancy must "already almost know" everything she is ever to learn. This can't be true for any reasonable sense of "almost know". I suppose I DO mean "can't ever learn unless". However, I disagree with your analysis. The "Platonic implication" need not be what you stated it to be if one cares to observe that some of the things an entity can learn are...how to learn better and how to learn more. My original statement presupposes an existence of a category system--a capacity to pigeonhole, if you will. Surely you won't take issue with the hypothesis that an infant's category system is lesser than that of an adult. Yet, faced with the fact that many infants do become adults, we have to explain how the category system can muster to grow up, as well. In order to do so, I propose to think that the human learning is a process where, say, in order to assimilate a chunk of information one has to have a hundred-, nay, a thousand-fold store of SIMILAR chunks. This is by direct analogy with physical growing up--it happens very slowly, gradually, incrementally--and yet it happens. If you recall, my original statement was made against attempting "wholesale learning" as opposed to "knowledge-rich" systems when building subcognitive sytems. Admittedly, the complexity of a human being is many an order of magnitude beyond that what AI will attempt for decades to come, yet by observing the physical development of a child we can arrive at some sobbering tips for how to successfully build complex systems. Abandoning the utopia of having complex systems just "self-organize" and pop out of simple interactions of a few even simplier pieces is one such tip. -- Marek ------------------------------ Date: Tue 17 Jan 84 11:56:01-PST From: Juanita Mullen Subject: SIGLUNCH ANNOUNCEMENT- JANUARY 20, l984 [Reprinted from the Stanford SIGLUNCH distribution.] Friday, January 20, 1984 12:05 LOCATION: Chemistry Gazebo, between Physical & Organic Chemistry SPEAKER: Harold Brown Stanford University TOPIC: Palladio: An Exploratory Environment for Circuit Design Palladio is an environment for experimenting with design methodologies and knowledge-based design aids. It provides the means for constructing, testing and incrementally modifying design tools and languages. Palladio is a testbed for investigationg elements of design including specification, simulation, refinement and use of previous designs. For the designer, Palladio supports the construction of new specification languages particular to the design task at hand and augmentation of the system's expert knowledge to reflect current design goals and constraints. For the design environment builder, Palladio provides several programming paradigms: rule based, object oriented, data oriented and logical reasoning based. These capabilities are largely provided by two of the programming systems in which Palladio is implemented: LOOPS and MRS. In this talk, we will describe the basic design concepts on which Palladio is based, give examples of knowledge-based design aids developed within the environment, and describe Palladio's implementation. ------------------------------ End of AIList Digest ******************** 22-Jan-84 15:29:15-PST,15132;000000000001 Mail-From: LAWS created at 22-Jan-84 15:25:44 Date: Sun 22 Jan 1984 15:15-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #9 To: AIList@SRI-AI AIList Digest Monday, 23 Jan 1984 Volume 2 : Issue 9 Today's Topics: AI Culture - Survey Results Available, Digests - Vision-List Request, Expert Systems - Software Debugging, Seminars - Logic Programming & Bagel Architecture, Conferences - Principles of Distributed Computing ---------------------------------------------------------------------- Date: 18 Jan 84 14:50:21 EST From: Smadar Subject: How AI People Think - Cultural Premises of the AI Community... [Reprinted from the Rutgers bboard.] How AI People Think - Cultural Premises of the AI Community... is the name of a report by sociologists at the University of Genoa, Italy, based on a survey of AI researchers attending the International AI conference (IJCAI-8) this past summer. [...] Smadar. ------------------------------ Date: Wed, 18 Jan 84 13:08:34 PST From: Philip Kahn Subject: TO THOSE INTERESTED IN COMPUTER VISION, IMAGE PROCESSING, ETC This is the second notice directed to all of those interested in IMAGE PROCESSING, COMPUTER VISION, etc. There has been a great need, and interest, in compiling a VISION list that caters to the specialized needs and interests of those involved in image/vision processing/theory/ implementation. I broadcast a message to this effect over this BBOARD about three weeks ago asking for all those that are interested to respond. Again, I reiterate the substance of that message: 1) If you are interested in participating in a VISION list, and have not already expressed your interest to me, please do so! NOW is the time to express that interest, since NOW is when the need for such a list is being evaluated. 2) I cannot moderate the list (due to a lack of the proper type of resources to deal with the increased mail traffic). A moderator is DESPERATELY NEEDED! I will assist you in establishing the list, and I am presently in contact with the moderator of AILIST (Ken LAWS@SRI-AI) to establish what needs to be done. The job of moderator involves the following: i) All mail for the list is sent to you ii) You screen (perhaps, format or edit, depending upon the time and effort you wish to expend) all incoming messages, then redistribute them to the participants on the list at regular intervals. iii) You maintain/update the distribution list. Needless to say, the job of moderator is extremely rewarding and involves a great deal of high visibility. In addition, you get to GREATLY AID in the dissemination and sharing of ideas and information in this growing field. Enough said... 3) If you know of ANYONE that might be interested in such a list, PLEASE LET THEM KNOW and have them express that interest to me by sending mail to KAHN@UCLA-CS.ARPA Now's the time to let me know! Philip Kahn send mail to: KAHN@UCLA-CS.ARPA ------------------------------ Date: 19 Jan 84 15:14:04 EST From: Lou Subject: Re: Expert systems for software debugging I don't know of any serious work in AI on software debugging since HACKER. HACKER was a part of the planning work done at MIT some years ago - it was an approach to planning/automatic programming where planning was done with a simple planner that, e.g., ignored interactions between plan steps. Then HACKER ran the plan/program and had a bunch of mini-experts that detected various kinds of bugs. See Sussman, A Computer Model of Skill Acquisition, MIT Press, 1975. Also, there is some related work in hardware debugging. Are you aware of the work by Randy Davis at MIT and by Mike Genesereth at Stanford on hardware trouble shooting? This is the problem where you have a piece of hardware (e.g. a VAX) that used to work but is now broken, and you want to isolate the component (board, chip, etc.) that needs to be replaced. Of course this is a bit different from program debugging, since you are looking for a broken component rather than a mis-design. E.g. for trouble shooting you can usually assume a single thing is broken, but you often have multiple bugs in a program. Here at Rutgers, we're working on an aid for design debugging for VLSI. Design debugging is much more like software debugging. Our basic approach is to use a signal constraint propagation method to generate a set of possible places where the bug might be, and then use various sorts of heuristics to prune the set (e.g. a sub-circuit that's been used often before is less likely to have a bug than a brand new one). ------------------------------ Date: Fri, 20 Jan 84 8:39:38 EST From: Paul Broome Subject: Re: Expert systems for software debugging? Debugging is a black art, not at all algorithmic, but almost totally heuristic. There is a lot of expert knowledge around about how to debug faulty programs, but it is rarely written down or systemetized. Usually it seems to reside solely in the minds of a few "debugging whizzes". Does anyone know of an expert system that assists in software debugging? Or any attempts (now or in the past) to produce such an expert? There are some good ideas and a Prolog implementation in Ehud Shapiro's Algorithmic Program Debugging, which is published as an ACM distinguished dissertation by MIT Press, 1983. One of his ideas is "divide-and-query: a query-optimal diagnosis algorithm," which is essentially a simple binary bug search. If the program is incorrect on some input then the program is divided into two roughly equal subtrees and the computation backtracks to the midpoint. If this intermediate result is correct then the first subtree is ignored and the bug search is repeated on the second subtree. If the intermediate result is incorrect then the search continues instead on the first subtree. ------------------------------ Date: 20 Jan 84 19:25:30-PST (Fri) From: pur-ee!uiucdcs!nielsen @ Ucb-Vax Subject: Re: Expert systems for software debuggin - (nf) Article-I.D.: uiucdcs.4980 The Knowledge Based Programming Assistant Project here at the University of Illinois was founded as a result of a very similar proposal. A thesis you may be interested in which explains some of our work is "GPSI : An Expert System to Aid in Program Debugging" by Andrew Laursen which should be available through the university. I would be very interested in corresponding with anyone who is considering the use of expert systems in program debugging. Paul Nielsen {pur-ee, ihnp4}!uiucdcs!nielsen nielsen@uiucdcs ------------------------------ Date: 01/19/84 22:25:55 From: PLUKEL Subject: January Monthly Meeting, Greater Boston Chapter/ACM [Forwarded from MIT by SASW@MIT-MC.] On behalf of GBC/ACM, J. Elliott Smith, the Lecture Chairman, is pleased to present a discussion on the topic of LOGIC PROGRAMMING Henryk Jan Komorowski Division of Applied Sciences Harvard University Cambridge, Massachusetts Dr. Komorowski is an Assistant Professor of Computer Science, who received his MS from Warsaw University and his Phd from Linkoeping University, Linkoeping, Sweden, in 1981. His current research interests include applications of logic programming to: rapid prototyping, programming/specification development envir- onments, expert systems, and databases. Dr. Komorowski's articles have appeared in proceedings of the IXth POPL, the 1980 Logic Programming Workshop (Debrecen, Hungary), and the book "Logic Programming", edited by Clark and Taernlund. He acted as Program Chairman for the recent IEEE Prolog tutorial at Brandies University, is serving on the Program Committee of the 1984 Logic Programming Symposium (Atlantic City), and is a member of the Editorial Board of THE JOURNAL OF LOGIC PROGRAMMING. Prolog has been selected as the programming language of the Japanese Fifth Generation Computer Project. It is the first realization of logic programming ideas, and implements a theorem prover based on a design attributed to J.A. Robinson, which limits resolution to a Horn clause subset of assertions. A Prolog program is a collection of true statements in the form of RULES. A computation is a proof from these assertions. Numerous implementations of Prolog have elaborated Alain Colmerauer's original, including Dr. Komorowski's own Qlog, which operates in LISP environments. Dr. Komorowski will present an introduction to elementary logic programming concepts and an overview of more advanced topics, including metalevel inference, expert systems programming, databases, and natural language processing. DATE: Thursday, 26 January 1984 TIME: 8:00 PM PLACE: Intermetrics Atrium 733 Concord Avenue Cambridge, MA (near Fresh Pond Circle) COMPUTER MOVIE and REFRESHMENTS before the talk. Lecture dinner at 6pm open to all GBC members. Call (617) 444-5222 for additional details. ------------------------------ Date: 20 Jan 84 1006 PST From: Rod Brooks Subject: Shaprio Seminars at Stanford and Berkeley [Adapted from the SU-SCORE bboard and the Prolog Digest.] Ehud Shapiro, The Weizmann Institute of Science The Bagel: A Systolic Concurrent Prolog Machine 4:30pm, Terman Auditorium, Tues, Jan 24th, Stanford CSD Colloq. 1:30pm, Evans 597, Wed., Jan 2th, Berkeley Prolog Seminar It is argued that explicit mapping of processes to processors is essential to effectively program a general-purpose parallel computer, and, as a consequence, that the kernel language of such a computer should include a process-to-processor mapping notation. The Bagel is a parallel architecture that combines concepts of dataflow, graph-reduction and systolic arrays. The Bagel's kernel language is Concurrent Prolog, augmented with Turtle programs as a mapping notation. Concurrent Prolog, combined with Turtle programs, can easily implement systolic systems on the Bagel. Several systolic process structures are explored via programming examples, including linear pipes (sieve of Erasthotenes, merge sort, natural-language interface to a database), rectangular arrays (rectangular matrix multiplication, band-matrix multiplication, dynamic programming, array relaxation), static and dynamic H-trees (divide-and-conquer, distributed database), and chaotic structures (a herd of Turtles). All programs shown have been debugged using the Turtle graphics Bagel simulator, which is implemented in Prolog. ------------------------------ Date: Fri 20 Jan 84 14:56:58-PST From: Jayadev Misra Subject: call for Papers- Principles of Distributed Computing CALL FOR PAPERS 3rd ACM SIGACT-SIGOPS Symposium on Principles of Distributed Computing (PODC) Vancouver, Canada August 27 - 29, 1984 This conference will address fundamental issues in the theory and practice of concurrent and distributed systems. Original research papers describing theoretical or practical aspects of specification. design or implementation of such systems are sought. Topics of interest include, but are not limited to, the following aspects of concurrent and distributed systems. . Algorithms . Formal models of computations . Methodologies for program development . Issues in specifications, semantics and verifications . Complexity results . Languages . Fundamental results in application areas such as distributed databases, communication protocols, distributed operating systems, distributed transaction processing systems, real time systems. Please send eleven copies of a detailed abstract (not a complete paper) not exceeding 10 double spaced typewritten pages, by MARCH 8, 1984, to the Program Chairman: Prof. J. Misra Computer Science Department University of Texas Austin, Texas 78712 The abstract must include a clear description of the problem be- ing addressed, comparisons with extant work and a section on ma- jor original contributions of this work. The abstract must pro- vide sufficient detail for the program committee to make a deci- sion. Papers will be chosen on the basis of scientific merit, originality, clarity and appropriateness for this conference. Authors will be notified of acceptance by April 30, 1984. Ac- cepted papers, typed on special forms, are due at the above ad- dress by June 1, 1984. Authors of accepted papers will be asked to sign ACM Copyright forms. The Conference Chairman is Professor Tiko Kameda (Simon Fraser University). The Publicity Chairman is Professor Nicola Santoro (Carleton University). The Local Arrangement Chiarman is Profes- sor Joseph Peters (Simon Fraser University). The Program Commit- tee consists of Ed Clarke (C.M.U.), Greg N. Frederickson (Pur- due), Simon Lam (U of Texas, Austin), Leslie Lamport (SRI Inter- national), Michael Malcom (U of Waterloo), J. Misra, Program Chairman (U of Texas, Austin), Hector G. Molina (Princeton), Su- san Owicki (Stanford), Fred Schneider (Cornell), H. Ray Strong (I.B.M. San Jose), and Howard Sturgis (Xerox Parc). ------------------------------ End of AIList Digest ******************** 26-Jan-84 14:42:19-PST,17480;000000000001 Mail-From: LAWS created at 26-Jan-84 14:40:48 Date: Thu 26 Jan 1984 14:23-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #10 To: AIList@SRI-AI AIList Digest Friday, 27 Jan 1984 Volume 2 : Issue 10 Today's Topics: AI Culture - IJCAI Survey, Cognition - Parallel Processing Query, Programming Languages - Symbolics Support & PROLOG/ZOG Request, AI Software - KEE Knowledge Representation System, Review - Rivest Forsythe Lecture on Learning, Seminars - Learning with Constraints & Semantics of PROLOG, Courses - CMU Graduate Program in Human-Computer Interaction ---------------------------------------------------------------------- Date: 24 Jan 84 12:19:21 EST From: Smadar Subject: Report on "How AI People Think..." I received a free copy because I attended IJCAI. I have an address here, but I don't know if it is the appropriate one for ordering this report: Re: the report "How AI People Think - Cultural Premises of the AI community" Commission of the European Communities Rue de la Loi, 200 B-1049 Brussels, Belgium (The report was compiled by Massimo Negrotti, Chair of Sociology of Knowledge, University of Genoa, Italy) Smadar (KEDAR-CABELLI@RUTGERS). ------------------------------ Date: Wed 18 Jan 84 11:05:26-PST From: Rene Bach Subject: brain, a parallel processor ? What are the evidences that the brain is a parallel processor? My own introspection seem to indicate that mine is doing time-sharing. That is I can follow only one idea at a time, but with a lot of switching between reasoning paths (often more non directed than controlled switching). Have different people different processors ? Or is the brain able to function in more than one way (parallel, serial, time-sharing) ?? Rene (bach@sumex) ------------------------------ Date: Wed, 25 Jan 84 15:37:39 CST From: Mike Caplinger Subject: Symbolics support for non-Lisp languages [This is neither an AI nor a graphics question per se, but I thought these lists had the best chance of reaching Symbolics users...] What kind of support do the Symbolics machines provide for languages other than Lisp? Specifically, are there interactive debugging facilities for Fortran, Pascal, etc.? It's my understanding that the compilers generate Lisp output. Is this true, and if so, is the interactive nature of Lisp exploited, or are the languages just provided as batch compilers? Finally, does anyone have anything to say about efficiency? Answers to me, and I'll summarize if there's any interest. Thanks. ------------------------------ Date: Wed 25 Jan 84 09:38:25-PST From: Ken Laws Reply-to: AIList-Request@SRI-AI Subject: KEE Representation System The Jan. issue of IEEE Computer Graphics reports the following: Intelligenetics has introduced the Knowledge Engineering Environment AI software development system for AI professionals, computer scientists, and domain specialists. The database management program development system is graphics oriented and interactive, permitting use of a mouse, keyboard, command-option menus, display-screen windows, and graphic symbols. KEE is a frame-based representation system that provides support for descriptive and procedural knowledge representation, and a declarative, extendable formalism for controlling inheritance of attributes and attribute values between related units of knowledge. The system provides support for multiple inheritance hierarchies; the use of user-extendable data types to promote knowledge-base integrity; object-oriented programming; multiple- inference engines/rule systems; and a modular system design through multiple knowledge bases. The first copy of KEE sells for $60,000; the second for $20,000. Twenty copies cost $5000 each. ------------------------------ Date: 01/24/84 12:08:36 From: JAWS@MIT-MC Subject: PROLOG and/or ZOG for TOPS-10 Does anyone out there know where I can get a version of prolog and/or ZOG to that will run on a DEC-10 (7.01)? The installation is owned by the US government, albeit beneign (DOT). THANX JAWS@MC ------------------------------ Date: Tue 24 Jan 84 11:26:14-PST From: Armar Archbold Subject: Rivest Forsythe Lecture on Learning [The following is a review of a Stanford talk, "Reflections on AI", by Dr. Ron Rivest of MIT. I have edited the original slightly after getting Armar's permission to pass it along. -- KIL] Dr. Rivest's talk emphasized the interest of small-scale studies of learning through experience (a "critter" with a few sensing and effecting operations building up a world model of a blocks environment). He stressed such familiar themes as - "the evolutionary function and value of world models is predicting the future, and consequently knowledge is composed principally of expectations, possibilities, hypotheses - testable action-sensation sequences, at the lowest level of sophistication", - "the field of AI has focussed more on 'backdoor AI', where you directly program in data structures representing high-level knowledge, than on 'front-door' AI, which studies how knowledge is built up from non-verbal experience, or 'side door AI', which studies how knowledge might be gained through teaching and instruction using language; - such a study of simple learning systems in a simple environment -- in which an agent with a given vocabulary but little or no initial knowledge ("tabula rasa") investigates the world (either through active experiementation or through changes imposed by perturbations in the surroundings) and attempts to construct a useful body of knowledge through recognition of identities, equivalences, symmetries, homomorphisms, etc., and eventually metapatterns, in action-sensation chains (represented perhaps in dynamic logic) -- is of considerable interest. Such concepts are not new. There have been many mathematical studies, psychological similations, and AI explorations along the lines since the 50s. At SRI, Stan Rosenschein was playing around with a simplified learning critter about a year ago; Peter Cheeseman shares Rivest's interest in Jaynes' use of entropy calculations to induce safe hypotheses in an overwhelmingly profuse space of possibilities. Even so, these concerns were worth having reactivated by a talk. The issues raised by some of the questions from the audience were also intesting, albeit familiar: - The critter which starts out with a tabula rasa will only make it through the enormous space of possible patterns induceable from experience if it initially "knows" an awful lot about how to learn, at whatever level of procedural abstraction and/or "primitive" feature selection (such as that done at the level of the eye itself). - Do we call intelligence the procedures that permit one to gain useful knowledge (rapidly), or the knowledge thus gained, or what mixture of both? - In addition, there is the question of what motivational structure best furthers the critter's education. If the critter attaches value to minimum surprise (various statistical/entropy measures thereof), it can sit in a corner and do nothing, in which case it may one day suddenly be very surprised and very dead. If it attaches tremendous value to surprise, it could just flip a coin and always be somewhat surprised. The mix between repetition (non-surprise/confirmatory testing) and exploration which produces the best cognitive system is a fundamental problem. And there is the notion of "best" - "best" given the critter's values other than curiosity, or "best" in terms of survivability, or "best" in a kind of Occam's razor sense vis-a-vis truth (here it was commented you could rank Carnapian world models based on the simple primitive predicates using Kolmogorov complexity measures, if one could only calculate the latter...) - The success or failure of the critter to acquire useful knowledge depends very much on the particular world it is placed in. Certain sequences of stimuli will produce learning and others won't, with a reasonable, simple learning procedure. In simple artificial worlds, it is possible to form some kind of measure of the complexity of the environment by seeing what the minimum length action-sensation chains are which are true regularities. Here there is another traditional but fascinating question: what are the best worlds for learning with respect to critters of a given type - if the world is very stochastic, nothing can be learned in time; if the world is almost unchanging, there is little motivation to learn and precious little data about regular covariances to learn from. Indeed, in psychological studies, there are certain sequences which will bolster reliance on certain conclusions to such an extent that those conclusions become (illegitimately) protected from disconfirmation. Could one recreate this phenomenon with a simple learning critter with a certain motivational structure in a certain kind of world? Although these issues seemed familiar, the talk certainly could stimulate the general public. Cheers - Armar ------------------------------ Date: Tue 24 Jan 84 15:45:06-PST From: Juanita Mullen Subject: SIGLUNCH ANNOUNCEMENT - FRIDAY, January 27, 1984 [Reprinted from the Stanford SIGLUNCH distribution.] Friday, January 27, 1984 Chemistry Gazebo, between Physical & Organic Chemistry 12:05 SPEAKER: Tom Dietterich, HPP Stanford University TOPIC: Learning with Constraints In attempting to construct a program that can learn the semantics of UNIX commands, several shortcomings of existing AI learning techniques have been uncovered. Virtually all existing learning systems are unable to (a) perform data interpretation in a principled way, (b) form theories about systems that contain substantial amounts of state information, (c) learn from partial data, and (d) learn in a highly incremental fashion. This talk will describe these shortcomings and present techniques for overcoming them. The basic approach is to employ a vocabulary of constraints to represent partial knowledge and to apply constraint-propagation techniques to draw inferences from this partial knowledge. These techniques are being implemented in a system called, EG, whose task is to learn the semantics of 13 UNIX commands (ls, cp, mv, ln, rm, cd, pwd, chmod, umask, type, create, mkdir, rmdir) by watching "over-the-shoulder" of a teacher. ------------------------------ Date: 01/25/84 17:07:14 From: AH Subject: Theory of Computation Seminar [Forwarded from MIT-MC by SASW.] DATE: February 2nd, 1984 TIME: 3:45PM Refreshments 4:00PM Lecture PLACE: NE43-512A "OPERATIONAL AND DENOTATIONAL SEMANTICS FOR P R O L O G" by Neil D. Jones Datalogisk Institut Copenhagen University Abstract A PROLOG program can go into an infinite loop even when there exists a refutation of its clauses by resolution theorem proving methods. Conseguently one can not identify resolution of Horn clauses in first-order logic with PROLOG as it is actually used, namely, as a deterministic programming language. In this talk two "computational" semantics of PROLOG will be given. One is operational and is expressed as an SECD-style interpreter which is suitable for computer implementation. The other is a Scott-Strachey style denotational semantics. Both were developed from the SLD-refutation procedure of Kowalski and APT and van Embden, and both handle "cut". HOST: Professor Albert R. Meyer ------------------------------ Date: Wednesday, 25 Jan 84 23:47:29 EST From: reiser (brian reiser) @ cmu-psy-a Reply-to: Subject: Human-Computer Interaction Program at CMU ***** ANNOUNCEMENT ***** Graduate Program in Human-Computer Interaction at Carnegie-Mellon University The field of human-computer interaction brings to bear theories and methodologies from cognitive psychology and computer science to the design of computer systems, to instruction about computers, and to computer-assisted instruction. The new Human-Computer Interaction program at CMU is geared toward the development of cognitive models of the complex interaction between learning, memory, and language mechanisms involved in using computers. Students in the program apply their psychology and computer science training to research in both academic and industry settings. Students in the Human-Computer Interaction program design their educational curricula with the advice of three faculty members who serve as the student's committee. The intent of the program is to guarantee that students have the right combination of basic and applied research experience and coursework so that they can do leading research in the rapidly developing field of human-computer interaction. Students typically take one psychology course and one computer science course each semester for the first two years. In addition, students participate in a seminar on human-computer interaction held during the summer of the first year in which leading industry researchers are invited to describe their current projects. Students are also actively involved in research throughout their graduate career. Research training begins with a collaborative and apprentice relationship with a faculty member in laboratory research for the first one or two years of the program. Such involvement allows the student several repeated exposures to the whole sequence of research in cognitive psychology and computer science, including conceptualization of a problem, design and execution of experiments, analyzing data, design and implementation of computer systems, and writing scientific reports. In the second half of their graduate career, students participate in seminars, teaching, and an extensive research project culminating in a dissertation. In addition, an important component of students' training involves an internship working on an applied project outside the academic setting. Students and faculty in the Human-Computer Interaction program are currently studying many different cognitive tasks involving computers, including: construction of algorithms, design of instruction for computer users, design of user-friendly systems, and the application of theories of learning and problem solving to the design of systems for computer-assisted instruction. Carnegie-Mellon University is exceptionally well suited for a program in human-computer interaction. It combines a strong computer science department with a strong psychology department and has many lines of communication between them. There are many shared seminars and research projects. They also share in a computational community defined by a large network of computers. In addition, CMU and IBM have committed to a major effort to integrate personal computers into college education. By 1986, every student on campus will have a powerful state-of-the-art personal computer. It is anticipated that members of the Human-Computer Interaction program will be involved in various aspects of this effort. The following faculty from the CMU Psychology and Computer Science departments are participating in the Human-Computer Interaction Program: John R. Anderson, Jaime G. Carbonell, John R. Hayes, Elaine Kant, David Klahr, Jill H. Larkin, Philip L. Miller, Alan Newell, Lynne M. Reder, and Brian J. Reiser. Our deadline for receiving applications, including letters of recommendation, is March 1st. Further information about our program and application materials may be obtained from: John R. Anderson Department of Psychology Carnegie-Mellon University Pittsburgh, PA 15213 ------------------------------ End of AIList Digest ******************** 31-Jan-84 10:19:37-PST,15850;000000000001 Mail-From: LAWS created at 31-Jan-84 10:14:56 Date: Tue 31 Jan 1984 10:05-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #11 To: AIList@SRI-AI AIList Digest Tuesday, 31 Jan 1984 Volume 2 : Issue 11 Today's Topics: Techniques - Beam Search Request, Expert Systems - Expert Debuggers, Mathematics - Arnold Arnold Story, Courses - PSU Spring AI Mailing Lists, Awards - Fredkin Prize for Computer Math Discovery, Brain Theory - Parallel Processing, Intelligence - Psychological Definition, Seminars - Self-Organizing Knowledge Base, Learning, Task Models ---------------------------------------------------------------------- Date: 26 Jan 1984 21:44:11-EST From: Peng.Si.Ow@CMU-RI-ISL1 Subject: Beam Search I would be most grateful for any information/references to studies and/or applications of Beam Search, the search procedure used in HARPY. Peng Si Ow pso@CMU-RI-ISL1 ------------------------------ Date: 25 Jan 84 7:51:06-PST (Wed) From: harpo!eagle!mhuxl!ulysses!unc!mcnc!ncsu!uvacs!erh @ Ucb-Vax Subject: Expert debuggers Article-I.D.: uvacs.1148 See also "Sniffer: a system that understands bugs", Daniel G. Shapiro, MIT AI Lab Memo AIM-638, June 1981 (The debugging knowledge of Sniffer is organized as a bunch of tiny experts, each understanding a specific type of error. The program has an in- depth understanding of a (very) limited class of errors. It consists of a cliche-finder and a "time rover". Master's thesis.) ------------------------------ Date: Thursday, 26-Jan-84 19:11:37-GMT From: BILL (on ERCC DEC-10) Reply-to: Clocksin Subject: AIList entry In reference to a previous AIList correspondent wishing to know more about Arnold Arnold's "proof" of Fermat's Last Theorem, last week's issue of New Scientist explains all. The "proof" is faulty, as expected. Mr Arnold is a self-styled "cybernetician" who has a history of grabbing headlines with announcements of revolutionary results which are later proven faulty on trivial grounds. I suppose A.I. has to put up with its share of circle squarers and angle trisecters. ------------------------------ Date: 28 Jan 84 18:23:09-PST (Sat) From: ihnp4!houxm!hocda!hou3c!burl!clyde!akgua!sb1!sb6!bpa!burdvax!psu vax!bobgian@Ucb-Vax Subject: PSU Spring AI mailing lists Article-I.D.: psuvax.433 I will be using net.ai for occasionally reporting "interesting" items relating to the PSU Spring AI course. If anybody would also like "administrivia" mailings (which could get humorous at times!), please let me know. Also, if you want to be included on the "free-for-all" discussion list, which will include flames and other assorted idiocies, let me know that too. Otherwise you'll get only "important" items. The "official Netwide course" (ie, net.ai.cse) will start up in a month or so. Meanwhile, you are welcome to join the fun via mail! Bob Bob Giansiracusa (Dept of Computer Science, Penn State Univ, 814-865-9507) UUCP: bobgian@psuvax.UUCP -or- allegra!psuvax!bobgian Arpa: bobgian@PSUVAX1 -or- bobgian%psuvax1.bitnet@Berkeley Bitnet: bobgian@PSUVAX1.BITNET CSnet: bobgian@penn-state.csnet USnail: 333 Whitmore Lab, Penn State Univ, University Park, PA 16802 ------------------------------ Date: 26 Jan 84 19:39:53 EST From: AMAREL@RUTGERS.ARPA Subject: Fredkin Prize for Computer Math Discovery [Reprinted from the RUTGERS bboard.] Fredkin Prize to be Awarded for Computer Math Discovery LOUISVILLE, Ky.--The Fredkin Foundation will award a $100,000 prize for the first computer to make a major mathematical discovery, it was announced today (Jan. 26). Carnegie-Mellon University has been named trustee of the "Fredkin Prize for Computer Discovery in Mathematics", according to Raj Reddy, director of the university's Robotics Institute, and a trustee of IJCAI (International Joint Council on Artificial Intelligence) responsible for AI prizes. Reddy said the prize will be awarded "for a mathematical work of distinction in which some of the pivotal ideas have been found automatically by a computer program in which they were not initially implicit." "The criteria for awarding this prize will be widely publicized and reviewed by the artificial intelligence and mathematics communities to determine their adequacy," Reddy said. Dr. Woody Bledsoe of the University of Texas at Austin will head a committee of experts who will define the rules of the competition. Bledsoe is president-elect of the American Association for Artificial Intelligence. "It is hoped," said Bledsoe, "that this prize will stimulate the use of computers in mathematical research and have a good long-range effect on all of science." The committee of mathematicians and computer scientists which will define the rules of the competition includes: William Eaton of the University of Texas at Austin, Daniel Gorenstein of Rutgers University, Paul Halmos of Indiana University, Ken Kunen of the University of Wisconsin, Dan Mauldin of North Texas State University and John McCarthy of Stanford University. Also, Hugh Montgomery of the University of Michigan, Jack Schwartz of New York University, Michael Starbird of the University of Texas at Austin, Ken Stolarsky of the University of Illinois and Francois Treves of Rutgers University. The Fredkin Foundation has a similar prize for a world champion computer chess system. Recently, $5,000 was awarded to Ken Thompson and Joseph Condon, Bell Laboratories researchers who developed the first computer system to achieve a Master rating in tournament chess. ------------------------------ Date: 26 Jan 84 15:34:50 PST (Thu) From: Mike Brzustowicz Subject: Re: Rene Bach's query on parallel processing in the brain What happens when something is "on the tip of your tounge" but is beyond recall. Often (for me at least) if the effort to recall is displaced by some other cognitive activity, the searched-for information "pops-up" at a later time. To me, this suggests at least one background process. -Mike (mab@AIDS-UNIX) ------------------------------ Date: Thu, 26 Jan 84 17:19:30 PST From: Charlie Crummer Subject: How my brain works I find that most of what my brain does is pattern interpretation. I receive various sensory input in the form of various kinds of vibrations (i.e. eletromagnetic and acoustic) and my brain perceives patterns in this muck. Then it attaches meanings to the patterns. Within limits, I can attach these meanings at will. The process of logical deduction a la Socrates takes up a negligible time-slice in the CPU. --Charlie ------------------------------ Date: Fri, 27 Jan 84 15:35:21 PST From: Charlie Crummer Subject: Re: How my brain works I see what you mean about the question as to whether the brain is a parallel processor in consious reasoning or not. I also feel like a little daemon that sits and pays attention to different lines of thought at different times. An interesting counterexample is the aha! phenomenon. The mathematician Henri Poincare, among others, has written an essay about his experience of being interrupted from his conscious attention somehow and becoming instantly aware of the solution to a problem he had "given up" on some days before. It was as though some part of his brain had been working on the problem all along even though he had not been aware of it. When it had gotten the solution an interrupt occurred and his conscious mind was triggered into the awareness of the solution. --Charlie ------------------------------ Date: Mon 30 Jan 84 09:47:49-EST From: Alexander Sen Yeh Subject: Request for Information I am getting started on a project which combines symbolic artificial intelligence and image enhancement techniques. Any leads on past and present attempts at doing this (or at combining symbolic a.i. with signal processing or even numerical methods in general) would be greatly appreciated. I will send a summary of replies to AILIST and VISION LIST in the future. Thanks. --Alex Yeh --electronic mail: AY@MIT-XX.ARPA --US mail: Rm. 222, 545 Technology Square, Cambridge, MA 02139 ------------------------------ Date: 30 January 1984 1554-est From: RTaylor.5581i27TK @ RADC-MULTICS Subject: RE: brain, a parallel processor ? I agree that based on my own observations, my brain appears to be working more like a time-sharing unit...complete with slow downs, crashes, etc., due to overloading the inputs by fatigue, poor maintenance, and numerous inputs coming too fast to be covered by the time-sharing/switching mechanism! Roz ------------------------------ Date: Monday, 30 Jan 84 14:33:07 EST From: shrager (jeff shrager) @ cmu-psy-a Subject: Psychological Definition of (human) Intelligence Recommended reading for persons interested in a psychological view of (human) intelligence: Sternberg, R.J. (1983) "What should intelligence tests test? Implications of a triarchic theory of intelligence for intelligence testing." in Educational Researcher, Jan 1984. Vol. 13 #1. This easily read article (written for educational researchers) reviews Sternberg's current view of what makes intelligent persons intelligent: "The triarchic theory accounts for why IQ tests work as well as they do and suggests ways in which they might be improved...." Although the readership of this list are probably not interested in IQ tests per se, Sternberg is the foremost cognitive psychologist concerned directly with intelligence so his view of "What is intelligence?" will be of interest. This is reviewed quite nicely in the cited paper: "The triachric theory of human intelligence comprises three subtheories. The first relates intelligence to the internal world of the individual, specifying the mental mechanisms that lead to more and less intelligent behavior. This subtheory specifies three kinds of information processing components that are instrumental in (a) learning how to do things, (b) planning what to do and how to do them, and in (c) actually doing them. ... The second subtheory specifies those points along the continuum of one's experience with tasks or situations that most critically involve the use of intelligence. In particular, the account emphasizes the roles of novelty (...) and of automatization (...) in intelligence. The third subtheory relates intelligence to the external world of the individual, specifying three classes of acts -- environmental adaptation, selection, and shaping -- that characterize intelligent behavior in the everyday world." There is more detail in the cited article. (Robert J. Sternberg is professor of Psychology at Yale University. See also, his paper in Behavior and Flame Sciences (1980, 3, 573-584): "Sketch of a componential subtheory of human intelligence." and his book (in press with Cambridge Univ. Press): "Beyond IQ: A triarchic theory of human intelligence.") ------------------------------ Date: Thu 26 Jan 84 14:11:55-CST From: CS.BUCKLEY@UTEXAS-20.ARPA Subject: Database Seminar [Reprinted from the UTEXAS-20 bboard.] 4-5 Wed afternoon in Pai 5.60 [...] Mail-From: CS.LEVINSON created at 23-Jan-84 15:47:25 I am developing a system which will serve as a self-organizing knowledge base for an expert system. The knowledge base is currently being developed to store and retrieve Organic Chemical reactions. As the fundamental structures of the system are merely graphs and sets, I am interested in finding other domains is which the system could be used. Expert systems require a large amount of knowledge in order to perform their tasks successfully. In order for knowledge to be useful for the expert task it must be characterized accurately. Data characterization is usually the responsibility of the system designer and the consulting experts. It is my belief that the computer itself can be used to help characterize and classify its knowledge. The system's design is based on the assumption that the key to knowledge characterization is pattern recognition. ------------------------------ Date: 28 Jan 84 21:25:17 EST From: MSIMS@RUTGERS.ARPA Subject: Machine Learning Seminar Talk by R. Banerji [Reprinted from the RUTGERS bboard.] MACHINE LEARNING SEMINAR Speaker: Ranan Banerji St. Joseph's University, Philadelphia, Pa. 19130 Subject: An explanation of 'The Induction of Theories from Facts' and its relation to LEX and MARVIN In Ehud Shapiro's Yale thesis work he presented a framework for inductive inference in logic, called the incremental inductive inference algorithm. His Model Inference System was able to infer axiomatizations of concrete models from a small number of facts in a practical amount of time. Dr. Banerji will relate Shapiro's work to the kind of inductive work going on with the LEX project using the version space concept of Tom Mitchell, and the positive focusing work represented by Claude Sammut's MARVIN. Date: Monday, January 30, 1984 Time: 2:00-3:30 Place: Hill 7th floor lounge (alcove) ------------------------------ Date: 30 Jan 84 1653 PST From: Terry Winograd Subject: Talkware seminar Mon Feb 6, Tom Moran (PARC) [Reprinted from the SU-SCORE bboard.] Talkware Seminar (CS 377) Date: Feb 6 Speaker: Thomas P. Moran, Xerox PARC Topic: Command Language Systems, Conceptual Models, and Tasks Time: 2:15-4 Place: 200-205 Perhaps the most important property for the usability of command language systems is consistency. This notion usually refers to the internal (self-) consistency of the language. But I would like to reorient the notion of consistency to focus on the task domain for which the system is designed. I will introduce a task analysis technique, called External-Internal Task (ETIT) analysis. It is based on the idea that tasks in the external world must be reformulated in to the internal concepts of a computer system before the system can be used. The analysis is in the form of a mapping between sets of external tasks and internal tasks. The mapping can be either direct (in the form of rules) or "mediated" by a conceptual model of how the system works. The direct mapping shows how a user can appear to understand a system, yet have no idea how it "really" works. Example analyses of several text editing systems and, for contrast, copiers will be presented; and various properties of the systems will be derived from the analysis. Further, it is shown how this analysis can be used to assess the potential transfer of knowledge from one system to another, i.e., how much knowing one system helps with learning another. Exploration of this kind of analysis is preliminary, and several issues will be raised for discussion. ------------------------------ End of AIList Digest ******************** 3-Feb-84 23:16:27-PST,12398;000000000001 Mail-From: LAWS created at 3-Feb-84 23:15:01 Date: Fri 3 Feb 1984 22:50-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #12 To: AIList@SRI-AI AIList Digest Saturday, 4 Feb 1984 Volume 2 : Issue 12 Today's Topics: Hardware - Lisp Machine Benchmark Request, Machine Translation - Request, Mathematics - Fermat's Last Theorem & Four Color Request, Alert - AI Handbooks & Constraint Theory Book, Expert Systems - Software Debugging Correction, Course - PSU's Netwide AI Course, Conferences - LISP Conference Deadline & Cybernetics Congress ---------------------------------------------------------------------- Date: Wed, 1 Feb 84 16:37:00 cst From: dyer@wisc-ai (Chuck Dyer) Subject: Lisp Machines Does anyone have any reliable benchmarks comparing Lisp machines, including Symbolics, Dandelion, Dolphin, Dorado, LMI, VAX 780, etc? Other features for comparison are also of interest. In particular, what capabilities are available for integrating a color display (at least 8 bits/pixel)? ------------------------------ Date: Thu 2 Feb 84 01:54:07-EST From: Andrew Y. Chu Subject: language translator [Forwarded by SASW@MIT-ML.] Hi, I am looking for some information on language translation (No, not fortran->pascal, like english->french). Does anyone in MIT works on this field? If not, anyone in other schools? Someone from industry ? Commercial product ? Pointer to articles, magazines, journals etc. will be greatly appreciated. Please reply to aychu@mit-xx. I want this message to reach as many people as possible, are there other bboards I can send to? Thanx. ------------------------------ Date: Thu, 2 Feb 84 09:48:48 PST From: Charlie Crummer Subject: Fermat's Last Theorem Fortunately (or unfortunately) puzzles like Fermat's Last Theorem, Goldbach's conjecture, the 4-color theorem, and others are not in the same class as the geometric trisection of an angle or the squaring of a circle. The former class may be undecidable propositions (a la Goedel) and the latter are merely impossible. Since one of the annoying things about undecidable propositions is that it cannot be decided whether or not they are decidable, (Where are you, Doug Hofstader, now that we need you?) people seriously interested in these candidates for undecidablilty should not dismiss so-called theorem provers like A. Arnold without looking at their work. I have heard that the ugly computer proof(?) of the 4-color theorem that appeared in Scientific American is incorrect, i.e. not a proof. I also have heard that one G. Spencer-Brown has proved the 4-color theorem. I do not know whether either of these things is true and it's bugging me! Is the 4-color theorem undecidable or not? --Charlie ------------------------------ Date: 30 Jan 84 19:48:36-PST (Mon) From: pur-ee!uiucdcs!uicsl!keller @ Ucb-Vax Subject: AI Handbooks only .95 Article-I.D.: uiucdcs.5251 Several people here have joined "The Library of Computer and Information Sciences Book Club" because they have an offer of the complete AI Handbook set (3 vols) for $3.95 instead of the normal $100.00. I got mine and they are the same production as non book club versions. You must buy three more books during the comming year and it will probably be easy to find ones that you want. Here's the details: Send to: The Library of Computer and Information Sciences Riverside NJ 08075 Copy of Ad: Please accept my application for trial membership in the Library of Computer and Information Sciences and send me the 3-volume HANDBOOK OF ARTIFICIAL INTELLIGENCE (10079) billing me only $3.95. I agree to purchase at least three additional Selections or Alternates over the next 12 months. Savings may range up to 30% and occasionally even more. My membership is cancelable any time after I buy these three books. A shipping and handling charge is added to all shipments. No-Risk Guarantee: If you are not satisfied--for any reason--you may return the HANDBOOK OF ARTIFICIAL INTELLIGENCE within 10 days and your membership will be canceled and you will owe nothing. Name ________ Name of Firm ____ (if you want subscription to your office) Address _____________ City ________ State _______ Zip ______ (Offer good in Continental U.S. and Canada only. Prices slightly higher in Canada.) Scientific American 8/83 7-BV8 -Shaun ...uiucdcs!uicsl!keller [I have been a member for several years, and have found this club's service satisfactory (and improving). The selection leans towards data processing and networking, but there have been a fair number of books on AI, graphics and vision, robotics, etc. After buying several books you get enough bonus points for a very substantial discount on a selection of books that you passed up when they were first offered. I do get tired, though, of the monthly brochures that use the phrase "For every computer professional, ..." in the blurb for nearly every book. If you aren't interested in the AI Handbook, find a current club member for a list of other books you can get when you enroll. The current member will also get a book for signing you up. -- KIL] ------------------------------ Date: 31 Jan 84 19:55:24-PST (Tue) From: pur-ee!uiucdcs!ccvaxa!lipp @ Ucb-Vax Subject: Constraint Theory - (nf) Article-I.D.: uiucdcs.5285 *********************BOOK ANNOUNCEMENT******************************* CONSTRAINT THEORY An Approach to Policy-Level Modelling by Laurence D. Richards The cybernetic concepts of variety, constraint, circularity, and process provide the foundations for a theoretical framework for the design of policy support systems. The theoretical framework consists of a modelling language and a modelling mathematics. An approach to building models for policy support sys- tems is detailed; two case studies that demonstrate the approach are described. The modelling approach focuses on the structure of mental models and the subjec- tivity of knowledge. Consideration is given to ideas immanent in second-order cybernetics, including paradox, self-reference, and autonomy. Central themes of the book are "complexity", "negative reasoning", and "robust" or "value-rich" policy. 424 pages; 23 tables; 56 illustrations Hardback: ISBN 0-8191-3512-7 $28.75 Paperback:ISBN 0-8191-3513-5 $16.75 order from: University Press of America 4720 Boston Way Lanham, Maryland 20706 USA ------------------------------ Date: 28 Jan 84 0:25:20-PST (Sat) From: pur-ee!uiucdcs!renner @ Ucb-Vax Subject: Re: Expert systems for software debugging Article-I.D.: uiucdcs.5217 Ehud Shapiro's error diagnosis system is not an expert system. It doesn't depend on a heuristic approach at all. Shapiro tries to find the faulty part of a bad program by executing part of the program, then asking an "oracle" to decide if that part worked correctly. I am very impressed with Shapiro's work, but it doesn't have anything to do with "expert knowledge." Scott Renner {ihnp4,pur-ee}!uiucdcs!renner ------------------------------ Date: 28 Jan 84 12:25:56-PST (Sat) From: ihnp4!houxm!hocda!hou3c!burl!clyde!akgua!sb1!sb6!bpa!burdvax!psuvax!bobgian @ Ucb-Vax Subject: PSU's Netwide AI course Article-I.D.: psuvax.432 The PSU ("in person") component of the course has started up, but things are a bit slow and confused regarding the "netwide" component. For one thing, I am too busy finishing a thesis and teaching full-time to handle the administrative duties, and we don't (yet, at least) have the resources to hire others to do it. For another, my plans presupposed a level of intellectual maturity and drive that is VERY rare in Penn State students. I believe the BEST that PSU can offer are in my course right now, but only 30 percent of them are ready for what I wanted to do (and most of THEM are FACULTY!!). I'm forced to backtrack and run a slightly more traditional "mini" course to build a common foundation. That course essentially will read STRUCTURE AND INTERPRETATION OF COMPUTER PROGRAMS by Hal Abelson and Gerry Sussman. [This book was developed for the freshman CS course (6.001) at MIT and will be published in April. It is now available as an MIT LCS tech report by writing Abelson at 545 Technology Square, Cambridge, MA 02139.] The "netwide" version of the course WILL continue in SOME (albeit perhaps delayed) form. My "mini" course should take about 6 weeks. After that the "AI and Mysticism" course can be restarted. For now, I won't create net.ai.cse but rather will use net.ai for occasional announcements. I'll also keep addresses of all who wrote expressing interest (and lack of a USENET connection). Course distributions will go (low volume) to that list and to net.ai until things start to pick up. When it becomes necessary we will "fork off" into a net.ai subgroup. So keep the faith, all you excited people! This course is yet to be!! Bob Bob Giansiracusa (Dept of Computer Science, Penn State Univ, 814-865-9507) UUCP: bobgian@psuvax.UUCP -or- allegra!psuvax!bobgian Arpa: bobgian@PSUVAX1 -or- bobgian%psuvax1.bitnet@Berkeley Bitnet: bobgian@PSUVAX1.BITNET CSnet: bobgian@penn-state.csnet USnail: 333 Whitmore Lab, Penn State Univ, University Park, PA 16802 ------------------------------ Date: Fri 3 Feb 84 00:24:28-EST From: STEELE%TARTAN@CMU-CS-C.ARPA Subject: 1984 LISP Conference submissions deadline moved back Because of delays that occurred in getting out the call for papers, the deadline for submissions to the 1984 ACM Symposium on LISP and Functional Programming (to be held August 5-8, 1984) has been moved back from February 6 to February 15. The date for notification of acceptance or rejection of papers is now March 20 (was March 12). The date for return of camera-ready copy is now May 20 (was May 15). Please forward this message to anyone who may find it of interest. --Thanks, Guy L. Steele Jr. Program Chairman, 1984 ACM S. on L. and F.P. Tartan Laboratories Incorporated 477 Melwood Avenue Pittsburgh, Pennsylvania 15213 (412)621-2210 ------------------------------ Date: 31 Jan 84 19:54:56-PST (Tue) From: pur-ee!uiucdcs!ccvaxa!lipp @ Ucb-Vax Subject: Cybernetics Congress - (nf) Article-I.D.: uiucdcs.5284 6th International Congress of the World Organisation of General Systems and Cybernetics 10--14 September 1984 Paris, France This transdisciplinary congress will present the contemporary aspects of cybernetics and of systems, and examine their different currents. The proposed topics include both methods and domains of cybernetics and systems: 1) foundations, epistemology, analogy, modelisation, general methods of systems, history of cybernetics and systems science ideas. 2) information, organisation, morphogenesis, self-reference, autonomy. 3) dynamic systems, complex systems, fuzzy systems. 4) physico-chemical systems. 5) technical systems: automatics, simulation, robotics, artificial intelligence, learning. 6) biological systems: ontogenesis, physiology, systemic therapy, neurocybernetics, ethology, ecology. 7) human and social systems: economics, development, anthropology, management, education, planification. For further information: WOGSC Comite de lecture AFCET 156, Bld. Pereire F 75017 Paris, France Those who want to attend the congress are urged to register by writing to AFCET, at the above address, as soon as possible. ------------------------------ End of AIList Digest ******************** 4-Feb-84 23:14:11-PST,10353;000000000001 Mail-From: LAWS created at 4-Feb-84 23:12:52 Date: Sat 4 Feb 1984 23:06-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #13 To: AIList@SRI-AI AIList Digest Sunday, 5 Feb 1984 Volume 2 : Issue 13 Today's Topics: Brain Theory - Parallelism, Seminars - Feb 7th CSD Colloquium [STORY: Neural networks] ---------------------------------------------------------------------- Date: 31 Jan 84 09:15:02 EST (Tue) From: Dana S. Nau Subject: parallel processing in the brain From: Rene Bach What are the evidences that the brain is a parallel processor? My own introspection seem to indicate that mine is doing time-sharing. That is I can follow only one idea at a time, but with a lot of switching between reasoning paths (often more non directed than controlled switching). Does that mean you hold your breath and stop thinking while you're walking, and stop walking in order to breathe or think? More pointedly, I think it's incorrect to consider only consciously-controlled processes when we talk about whether or not the brain is doing parallel processing. Perhaps the conscious part of your mind can keep track of only one thing at a time, but most (probably >90%) of the processing done by the brain is subconscious. For example, most of us have to think a LOT about what we're doing when we're first learning to drive. But after a while, it becomes largely automatic, and the conscious part of our mind is freed to think about other things while we're driving. As another example, have you ever had the experience of trying unsuccessfully to remember something, and later remembering whatever-it-was while you were thinking about something else? SOME kind of processing was going on in the interim, or you wouldn't have remembered whatever-it-was. ------------------------------ Date: 30 Jan 84 20:18:33-PST (Mon) From: pur-ee!uiucdcs!parsec!ctvax!uokvax!andree @ Ucb-Vax Subject: Re: intelligence and genius - (nf) Article-I.D.: uiucdcs.5259 Sorry, js@psuvax, but I DO know something about what I spoke, even if I do have trouble typing. I am aware that theorom-proving machines are impossible. It's also fairly obvious that they would use lots of time and space. However, I didn't even MENTION them. I talked about two flavors of machine. One generated well-formed strings, and the other said whether they were true or not. I didn't say either machine proved them. My point was that the second of these machines is also impossible, and is closely related to Jerry's genius finding machines. [I assume that any statement containing genius is true.] Down with replying without reading! Subject: Brain Processing The Feb Scientific American has an article entitled "The Skill of Typing" which can help one to form insights into mechanisms of the brains processing. richard ------------------------------ Date: Thu, 2 Feb 84 08:24:35 PST From: Charlie Crummer Subject: AIList Digest V2 #10 Re: Parallel Processing in the Brain There are several instances of people experiencing what can most easily be explained as "tasking" in the brain. (an essay by Henri Poincare in "The World of Mathematics", "The Seamless Web" by Stanley Burnshaw) It appears that the conscious mind is rather clumsy at creative work and in large measure assigns tasks (in parallel) to the subconscious mind which operates in the background. When the background task is finished, an interrupt is generated and the conscious mind becomes aware of the solution without knowing how the problem was solved. --Charlie ------------------------------ Date: Thu 2 Feb 84 10:17:08-PST From: Kenji Sugiyama Subject: Re: Parallel brain? I had a strange experience when I had practiced abacus in Japan. An abacus is used for adding, subtracting, multipling, and dividing numbers. The practice consisted of a set of calculations in a definite amount of time, say, 15 minutes. During that time, I began to think of something other than the problem at hand. Then I noticed that fact ("Aha, I thought of this and that!"), and grinned at myself in my mind. In spite of these detours, I continued my calculations without an interruption. This kind of experience repeated several times. It seems to me that my brain might be parallel, at least, in simple tasks. ------------------------------ Date: 2 Feb 1984 8:16-PST From: fc%USC-CSE@ECLA.ECLnet Subject: Re: AIList Digest V2 #10 parallelism in the brain: Can you walk and chew gum at the same time? Fred ------------------------------ Date: Sat, 4 Feb 84 15:06:09 PST From: Philip Kahn Subject: The brain is parallel, yet data flow can be serial... In response to Rene Bach's question whether "the brain is a parallel processor." There is no other response other than an emphatic YES! The brain is comprised of about 10E9 neurons. Each one of those neurons is making locally autonomous calculations; it's hard to get more parallel than that! The lower brain functions (e.g., sensory preprocessing, lower motor control, etc.) are highly distributed and locally autonomous processors (i.e., pure parallel data flow). At the higher thought processing levels, however, it has been shown (can't cite anything, but I can get sources if someone wants me to dig them out) that logic tends to run in a serial fashion. That is, the brain is parallel (a hardware structure), yet higher logic processes apply the timing of thought in a serial nature (a "software" structure). It is generally agreed that the brain is an associational machine; it processes based upon the timing of diffuse stimuli and the resulting changes in the "action potential" of its member neurons. "Context" helps to define the strength and structure of those associational links. Higher thinking is generally a cognitive process where the context of situations is manipulated. Changing context (and some associational links) will often result in a "conclusion" significantly different than previously arrived upon. Higher thought may be thought as a three process cycle: decision (evaluation of an associational network), reasonability testing (i.e., is the present decision using a new "context" no different from the decision arrived upon utilizing the previous "context"?), and context alteration (i.e., "if my 'decision' is not 'reasonable' what 'contextual association' may be omitted or in error?"). This cycle is continued until the second step -- 'reasonability testing' -- has concluded that the result of this 'thinking' process is at least plausible. Although the implementation (assuming the trichotomy is correct) in the brain is via parallel neural structures, the movement of information through those structures is serial in nature. An interesting note on the above trichotomy; note what occurs when the input to the associational network is changed. If the new input is not consistent with the previously existing 'context' then the 'reasonability tester' will cause an automatic readjustment of the 'context'. Needless to say, this is not a rigorously proven theory of mine, but I feel it is quite plausible and that there are profuse psychophysical and phychological studies that reinforce the above model. As of now, I use it as a general guiding light in my work with vision systems, but it seems equally appplicable to general AI. Philip Kahn KAHN@UCLA-CS.ARPA ------------------------------ Date: 02/01/84 16:09:21 From: STORY at MIT-MC Re: Neural networks [Forwarded by SASW@MIT-ML.] DATE: Friday, February 3, 1984 TITLE: "NEURAL NETWORKS: A DISCUSSION OF VARIOUS MATHEMATICAL MODELS" SPEAKER: Margaret Lepley, MIT Neural networks are of interest to researchers in artificial intelligence, neurobiology, and even statistical mechanics. Because of their random parallel structure it is difficult to study the transient behavior of the networks. We will discuss various mathematical models for neural networks and show how the behaviors of these models differ. In particular we will investigate asynchronous vs. synchronous models with undirected vs. directed edges of various weights. HOST: Professor Silvio Micali ------------------------------ Date: 01 Feb 84 1832 PST From: Rod Brooks Subject: Feb 7th CSD Colloquium - Stanford [Reprinted from the SU-SCORE bboard.] A Perspective on Automatic Programming David R. Barstow Schlumberger-Doll Research 4:30pm, Terman Aud., Tues Feb 7th Most work in automatic programming has focused primarily on the roles of deduction and programming knowledge. However, the role played by knowledge of the task domain seems to be at least as important, both for the usability of an automatic programming system and for the feasibility of building one which works on non-trivial problems. This perspective has evolved during the course of a variety of studies over the last several years, including detailed examination of existing software for a particular domain (quantitaive interpretation of oil well logs) and the implementation of an experimental automatic programming system for that domain. The importance of domain knowledge has two importatnt implications: a primary goal of automatic programming research should be to characterize the programming process for specific domains; and a crucial issue to be addressed in these characterizations is the interaction of domain and programming knowledge during program synthesis. ------------------------------ End of AIList Digest ******************** 10-Feb-84 22:33:57-PST,18414;000000000001 Date: Fri 10 Feb 1984 22:16-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #14 To: AIList@SRI-AI AIList Digest Saturday, 11 Feb 1984 Volume 2 : Issue 14 Today's Topics: Requests - SHRDLU & Spencer-Brown & Programming Tests & UNITS, Replys - R1/XCON & AI Text & Lisp Machine Comparisons, Seminars - Symbolic Supercomputer & Expert Systems & Multiagent Planning ---------------------------------------------------------------------- Date: Sun, 29 Jan 84 16:30:36 PST From: Rutenberg.pa@PARC-MAXC.ARPA Reply-to: Rutenberg.pa@PARC-MAXC.ARPA Subject: does anyone have SHRDLU? I'm looking for a copy of SHRDLU, ideally in machine readable form although a listing would also be fine. If you have a copy or know of somebody who does, please send me a message! Thanks, Mike ------------------------------ Date: Mon, 6 Feb 84 14:48:37 PST From: Charlie Crummer Subject: Re: AIList Digest V2 #12 I would dearly like to get in contact with G. Spencer-Brown. Can anyone give me any kind of lead? I have tried his publisher, Bantam, and got no results. Thanks. --Charlie ------------------------------ Date: Wed, 8 Feb 84 19:26:38 CST From: Stan Barber Subject: Testing Programming Aptitude or Compentence I am interested in information on the following tests that have been or are currently administered to determine Programming Aptitude or Compentence. 1. Aptitude Assessment Battery:Programming (AABP) created by Jack M. Wolfe and made available to employers only from Programming Specialists, Inc. Brooklyn NY. 2. Programmer Aptitude/Compentence Test System sold by Haverly Systems, Inc. (Introduced in 1970) 3. Computer Programmer Aptitude Battery by SRA (Science Research Associates), Inc. (Examined in by F.L. Schmidt et.al. in Journal of Applied Psychology, Volume 65 [1980] p 643-661) 4. CLEP Exam on Computers and Data Processing. The College Board and the Educational Testing Service. 5. Graudate Record Exam Advanced Test in Computer Science by the Education Testing Service. Please send the answers to the following questions if you have taken or had experience with any of these tests: 1. How many scores and what titles did they used for the version of the exam that you took? 2. Did you feel the test actually measured your ability to learn to program or your current programming competence (that is, did you feel it asked relevant questions)? 3. What are your general impressions about testing and more specifically about testing special abilities or skills (like programming, writing, etc.) I will package up the results and send them to Human-nets. My thanks. Stan Barber Department of Psychology Rice University Houston TX 77251 sob@rice (arapnet,csnet) sob.rice@rand-relay (broken arpa mailers) ...!{parsec,lbl-csam}!rice!sob (uucp) (713) 660-9252 (bulletin board) ------------------------------ Date: 6 Feb 84 8:10:41-PST (Mon) From: decvax!linus!vaxine!chb @ Ucb-Vax Subject: UNITS request: Second Posting Article-I.D.: vaxine.182 Good morning! I am looking for a pointer to someone (or something) who is knowledgeable about the features and the workings of the UNITS package, developed at Stanford HPP. If you know something, or someone, and could drop me a note (through mail) I would greatly appreciate it. Thanks in advance. Charlie Berg ...allegra!linus!vaxine!chb ------------------------------ Date: 5 Feb 84 20:28:09-PST (Sun) From: hplabs!hpda!fortune!amd70!decwrl!daemon @ Ucb-Vax Subject: DEC's expert system for configuring VAXen Article-I.D.: decwrl.5447 [This is in response to an unpublished request about R1. -- KIL] Just for the record - we changed the name from "R1" to "XCON" about a year ago I think. It's a very useful system and is part of a family of expert systems which assist us in the operation of various corporate divisions (sales, service, manufacturing, installation). Mark Palmer Digital (UUCP) {decvax, ucbvax, allegra}!decwrl!rhea!nacho!mpalmer (ARPA) decwrl!rhea!nacho!mpalmer@Berkeley decwrl!rhea!nacho!mpalmer@SU-Shasta ------------------------------ Date: 6 Feb 84 7:15:33-PST (Mon) From: harpo!utah-cs!hansen @ Ucb-Vax Subject: Re: AI made easy?? Article-I.D.: utah-cs.2473 I'd try Artificial Intelligence by Elaine Rich (McGraw-Hill). It's easy reading, not too technical but gives a good overview to the novice. Chuck Hansen {...!utah-cs} ------------------------------ Date: 5 Feb 84 8:48:26-PST (Sun) From: hplabs!sdcrdcf!darrelj @ Ucb-Vax Subject: Re: Lisp Machines Article-I.D.: sdcrdcf.813 There really no such things as reasonable benchmarks for systems as different as various Lisp machines and VAXen are. Each machine has different strengths and weaknesses. Here is a rough ranking of machines: VAX 780 running Fortran/C standalone Dorado (5 to 10X dolphin) LMI Lambda, Symbolics 3600, KL-10 Maclisp (2 to 3X dolphin) Dolphin, dandelion, 780 VAX Interlisp, KL-10 Interlisp Relative speeds are very rough, and dependent on application. Notes: Dandelion and Dolphin have 16-bit ALUs, as a result most arithmetic is pretty slow (and things like trancendental functions are even worse because there's no way to floating arithmetic without boxing each intermediate result). There is quite a wide range of I/O bandwidth among these machines -- up to 530 Mbits/sec on a Dorado, 130M on a dolphin). Strong points of various systems: Xerox: a family of machines fully compatible at the core-image level, spanning a wide range of price and performance (as low as $26k for a minumum dandelion, to $150k for a heavily expanded Dorado). Further, with the exception of some of the networking and all the graphics, it is very highly compatible with both Interlisp-10 and Interlisp-VAX (it's reasonable to have a single set of sources with just a bit of conditional compilation). Because of the use of a relatively old dialect, they have a large and well debugged manual as well. LMI and Symbolics (these are really fairly similar as both are licensed from the MIT lisp machine work, and the principals are rival factions of the MIT group that developed it) these have fairly large microcode stores, and as a result more things are fast (e.g. much of graphics primitives are microcoded, so these are probably the machines for moby amounts of image processing and graphics. There are also tools for compiling directly to microcode for extra speed. These machines also contain a secondary bus such as Unibus or Multibus, so there is considerable flexibility in attaching exotic hardware. Weak points: Xerox machines have a proprietary bus, so there are very few options (philosphy is hook it to something else on the Ethernet). MIT machines speak a new dialect of lisp with only partial compatible with MACLISP (though this did allow adding many nice features), and their cost is too high to give everyone a machine. The news item to which this is a response also asked about color displays. Dolphin: 480x640x4 bits. The 4 bits go thru a color map to 24 bits. Dorado: 480x640x(4 or 8 or 24 bits). The 4 or 8 bits go thru a color map to 24 bits. Lisp software does not currently support the 24 bit mode. 3600: they have one or two (the LM-2 had 512x512x?) around 1Kx1Kx(8 or 16 or 24) with a color map to 30 bits. Dandelion: probably too little I/O bandwidth Lambda: current brochure makes passing mention of optional standard and high-res color displays. Disclaimer: I probably have some bias toward Xerox, as SDC has several of their machines (in part because we already had an application in Interlisp. Darrel J. Van Buer, PhD System Development Corp. 2500 Colorado Ave Santa Monica, CA 90406 (213)820-4111 x5449 ...{allegra,burdvax,cbosgd,hplabs,ihnp4,sdccsu3,trw-unix}!sdcrdcf!darrelj VANBUER@USC-ECL.ARPA ------------------------------ Date: 6 Feb 84 16:40 PDT From: Kandt.pasa@PARC-MAXC.ARPA Subject: Lisp Machines I have seen several benchmarks as a former Symbolics and current Xerox employee. These benchmarks have typically compared the LM-2 with the 1100; they have even included actual or estimated(?) 3600, 1108, or 1132 performances. These benchmarks, however, have seldom been very informative because the actual test code is not provided or a detailed discussion of the implementation. For example, is the test on the Symbolics machine coded in Zetalisp or with the Interlisp compatibility package? Or, in Interlisp, were fast functions used (FRPLACA vs. RPLACA)? (Zetalisp's RPLACA is equivalent to Interlisp's FRPLACA so that if this transformation was not performed the benchmark would favor the Symbolics machine.) What about efficiency issues such as block compiling, compiler optimizers, or explicitily declaring variables? There are also many other issues such as what happens when the data set gets very large in a real application instead of a toy benchmark or, in Zetalisp, should you turn the garbage collector on (its not normally on) and when you do what impact does it have on performance. In summary, be cautious about claims without thorough supportive evidence. Also realize that each machine has its own strengths and weaknesses; there is no definitive answer. Caveat emptor! ------------------------------ Date: Sat, 4 Feb 84 19:24 EST From: Thomas Knight Subject: Concurrent Symbolic Supercomputer [Forwarded by SASW@MIT-MC] FAIM-1 Fairchild AI Machine #1 An Ultra-Concurrent Symbolic Supercomputer by Dr. A. L. Davis Fairchild Laboratory for Artificial Intelligence Research Friday, February 10, 1984 Presently AI researchers are being hampered in the development of large scale symbolic applications such as expert systems, by the lack of sufficient machine horsepower to execute the application programs at a sufficiently rapid rate to make the application viable. The intent of the FAIM-1 machine is to provide a machine capable of 3 or 4 orders of magnitude performance improvement over that currently available on today's large main-frame machines. The main source of performance increase is in the exploitation of concurrency at the program, system, and architectural levels. In addition to the normal ancillary support activities, the work is being carried on in 3 areas: 1. Language Design - a frame based, object oriented language is being designed which allows the programmer to express highly concurrent symbolic algorithms. The mechanism permits both logical and procedural programming styles in a unified message based semantics fashion. In addition, the programmer may provide strategic information which aids the system in managing the concurrency structure on the physical resource components of the machine. 2. Machine Architecture - the machine derives its power from the homogeneous replication of a medium grain processor element. The element consists of a processor, message delivery subsystem, and a parallel pattern based memory subsystem known as the CxAM (Context Adressable Memory). 2 variants of a CxAM design are being done at this time and are targeted for fabrication on a sub 2 micron CMOS line. The connection topology for the replicated elements is a 3 axis, single twist, Hex plane which has the advantages of planar wiring, easy extensibility, variable off surface bandwidth, and permits a variety of fault tolerant designs. The Hex plane topology also permits nice hierarchical process growth without creating excess communication congestion which would cause false synchronization in otherwise concurrent activities. In addition the machine is being designed in hopes of an eventual wafer-scale integrated implementation. 3. Resource Allocation - with any concurrent system which does not require machine dependent programming styles, there is a generic problem in mapping the concurrent activities extant in the program efficiently onto the multi-resource ensemble. The strategy employed in the FAIM-1 system is to analyze the static structure of the source program, transform it into a graph, and then via a series of function preserving graph transforms produce a loadable version of the program which attempts to minimize communication cost while preserving the inherent concurrency structure. A certain level of dynamic compensation is guided by programmer supplied strategy information. The talk will present an overview of the work we have done in these areas. Host: Prof. Thomas Knight ------------------------------ Date: 8 Feb 84 15:59:49 EST From: Smadar Subject: III Seminar on Expert Systems this coming Tuesday... [Reprinted from the Rutgers bboard.] I I I SEMINAR Title: Automation of Modeling, Simulation and Experimental Design - An Expert System in Enzyme Kinetics Speaker: Von-Wun Soo Date: Tuesday, February 14,1983, 1:30-2:30 PM Location: Hill Center, Seventh floor lounge Von-Wun Soo, a Ph.D. student in our department, will give an informal talk on the thesis research he is proposing. This is his abstract: We are proposing to develop a general knowledge engineering tool to aid biomedical researchers in developing biological models and running simulation experiments. Without such powerful tools, these tasks can be tedious and costly. Our aim is to integrate these techniques used in modeling, simulation, optimization, and experimental design by using an expert system approach. In addition we propose to carry out experiments on the processes of theory formation used by the scientists. Enzyme kinetics is the domain where we are concentrating our efforts. However, our research goal is not restricted to this particular domain. We will attempt to demonstrate with this special case, how several new ideas in expert problem solving including automation of theory formation, scientific discovery, experimental design, and knowledge acquisition can be further developed. Four modules have been designed in parallel: PROKINAL, EPX, CED, DISC. PROKINAL is a model generator which simulates the qualitative reasoning of the kineticists who conceptualize and postulate a reaction mechanism for a set of experimental data. By using a general procedure known as the King-Altman procedure to convert a mechanism topology into a rate law function, and symbolic manipulation techniques to factor rate constant terms to kinetic constant term, PROKINAL yields a corresponding FORTRAN function which computes the reaction rate. EPX is a model simulation aid which is designed by combining EXPERT and PENNZYME. It is supposed to guide the novice user in using simulation tools and interpreting the results. It will take the data and the candidate model that has been generated from PROKINAL and estimate the parameters by a nonlinear least square fit. CED is a experimental design consultant which uses EXPERT to guide the computation of experimental conditions. Knowledge of optimal design from the statistical analysis has been taken into consideration by EXPERT in order to give advice on the appropriate measurements and reduce the cost of experimentation. DISC is a discovery module which is now at the stage of theoretical development. We wish to explore and simulate the behavior of scientific discovery in enzyme kinetics research and use the results in automating theory formation tasks. ------------------------------ Date: 09 Feb 84 2146 PST From: Rod Brooks Subject: CSD Colloquium [Reprinted from the Stanford bboard.] CSD Colloquium Tuesday 14th, 4:30pm Terman Aud Michael P. Georgeff, SRI International "Synthesizing Plans for Co-operating Agents" Intelligent agents need to be able to plan their activities so that they can assist one another with some tasks and avoid harmful interactions on others. In most cases, this is best achieved by communication between agents at execution time. This talk will discuss a method for synthesizing a synchronized multi-agent plan to achieve such cooperation between agents. The idea is first to form independent plans for each individual agent, and then to insert communication acts into these plans to synchronize the activities of the agents. Conditions for freedom from interference and cooperative behaviour are established. An efficient method of interaction and safety analysis is then developed and used to identify critical regions and points of synchronization in the plans. Finally, communication primitives are inserted into the plans and a supervisor process created to handle synchronization. ------------------------------ End of AIList Digest ******************** ------- 10-Feb-84 22:57:33-PST,13563;000000000001 Mail-From: LAWS created at 10-Feb-84 22:56:02 Date: Fri 10 Feb 1984 22:49-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #15 To: AIList@SRI-AI AIList Digest Saturday, 11 Feb 1984 Volume 2 : Issue 15 Today's Topics: Proofs - Fermat's Theorem & 4-Color Theorem, Brain Theory - Parallelism ---------------------------------------------------------------------- Date: 04 Feb 84 0927 PST From: Jussi Ketonen Subject: Fermat and decidability From the logical point of view, Fermat's last theorem is a Pi-1 statement. It follows that it is decidable. Whether it is valid or not is another matter. ------------------------------ Date: Sat 4 Feb 84 13:13:14-PST From: Richard Treitel Subject: Re: Spencer-Brown's Proof I don't know anything about the current status of the computer proof of the 4-colour theorem, though the last I heard (five years ago) was that it was "probably OK". That's why I use the word "theorem". However, I can shed some light on Spencer-Brown's alleged proof -- I was present at a lecture in Cambridge where he supposedly gave the outline of the proof, and I applauded politely, but was later fairly authoritatively informed that it disintegrated under closer scrutiny. This doesn't *necessarily* mean that the man is a total flake, since other such proofs by highly reputable mathematicians have done the same (we are told that one proof was believed for twelve whole years, late in the 19th century, before its flaw was discovered). - Richard ------------------------------ Date: Mon, 6 Feb 84 14:46:43 PST From: Charlie Crummer Subject: Scientific Method Isn't it interesting that most of what we think about proofs is belief! I guess until one actually retraces the steps of a proof and their justifications one can only express his belief in its truth or falsness. --Charlie ------------------------------ Date: 3 Feb 84 8:48:01-PST (Fri) From: harpo!eagle!allegra!alan @ Ucb-Vax Subject: Re: brain, a parallel processor ? Article-I.D.: allegra.2254 I've been reading things like: My own introspection seem to indicate that ... I find, upon introspection, that ... I find that most of what my brain does is ... I also feel like ... I agree that based on my own observations, my brain appears to be ... Is this what passes for scientific method in AI these days? Alan S. Driscoll AT&T Bell Laboratories ------------------------------ Date: 2 Feb 84 14:40:23-PST (Thu) From: decvax!genrad!grkermit!masscomp!clyde!floyd!cmcl2!rocky2!cucard! aecom!alex @ Ucb-Vax Subject: Re: brain, a parallel processor ? Article-I.D.: aecom.358 If the brain was a serial processor, the limiting processing speed would be the speed that neurons conduct signals. Humans, however, do very complex processing in real time! The other possibility is that the data structures of the brain are HIGHLY optimized. Alex S. Fuss {philabs, esquire, cucard}!aecom!alex ------------------------------ Date: Tue, 7 Feb 84 13:09:25 PST From: Adolfo Di-Mare Subject: I can think in parale||, but most of time I'm ---sequential. For example, a lot of * I can talk with (:-{) and at the same time I can be thinking on s.m.t.i.g else. I also do this when ai-list gets too boring: I keep browsing until I find something intere sting, and then I do read, with a better level of under-standing. In the u-time, I can daydream... However, If I really want to get s.m.t.i.g done, then I cannot think on anything else! In this cases, I just have one main-stream idea in my mind. When I'm looking for a solution, I seldom use depth first, or bread first search. Most of the time I use a convynatium of all these tricks I know to search, until one 'works'. To + up, I think we @|-< can do lots of things in lots of ways. And until we furnish computers with all this tools, they won't be able to be as intelligent as us. Just parale|| is not the ?^-1. Adolfo /// ------------------------------ Date: 7 Feb 1984 1433-PST From: EISELT%UCI-20A@Rand-Relay Subject: More on Philip Kahn's reply to Rene Bach I recently asked Philip Kahn (via personal net mail) to elaborate on his three cycle model of thought, which he described briefly in his reply to Rene Bach's question. Here is my request, and his reply: ------------------------- In your recent submission to AIList, you describe a three-process cycle model of higher-level brain function. Your model has some similarities to a model of text understanding we are working on here at UC Irvine. You say, though, that there are "profuse psychophysical and psychological studies that reinforce the ... model." I haven't seen any of these studies and would be very interested in reading them. Could you possibly send me references to these studies? Thank you very much. Kurt Eiselt eiselt@uci-20a ------------------------ Kurt, I said "profuse" because I have come across many psychological and physiological studies that have reinforced my belief. Unfortunately, I have very few specific references on this, but I'll tell you as much as I can.... I claim there are three stages: associational, reasonability, and context. I'll tell you what I've found to support each. Associational nets, also called "computational" or "parameter" nets, have been getting a lot of attention lately. Especially interesting are the papers coming out of Rochester (in New York state). I suggest the paper by Feldman called "Parameter Nets." Also, McCullough in "Embodiments of Mind" introduced a logical calculus that he proposes neural mechanisms use to form assocational networks. Since then, a considerable amount of work has been done on logical calculus, and these works are directly applicable to the analysis of associational networks. One definitive "associational network" found in nature that has been exhaustively defined by Ratliff is the lateral inhibition that occurs in the linear image sensor of the Limulus crab. Each element of the network inhibits its neighbors based upon its value, and the result is the second spatial derivative of the image brightness. Most of the works you will find to support associational nets are directly culled from neurophysiological studies. Yet, classical conditioning psychology defines the effects of association in its studies on forward and backward conditioning. Personally, I feel the biological proof of associational nets is more concrete. The support for a "reasonability" level of processing has more psychological support, because it is generally a cognitive process. For example, learning is facilitated by subject matter that is most consistent with past knowledge; that is, knowledge is most facilitated by a subject that is most "reasonable" in light of past knowledge. Some studies have shown, though I can't cite them, that the less "reasonable" a learning task, the lesser is the learned performance. I remember having seen at least a paper (I believe it was by a natural language processing researcher) that claimed that the facility of language is a metaphorical process. By definition, a metaphor is the comparison of alike traits in dissimilar things; it seems to me this is a very good way to look at the question of reasonability. Again, though, no specific references. In neurophysiology there are found "feedback loops" that may be considered "reasonability" testers in so far that they take action only when certain conditions are not met. You might want to look at work done on the cerebellum to document this. "Context" has been getting a lot of attention lately. Again, psychology is the major source of supporting evidence, yet neurophysiology has its examples also. Hormones are a prime example of "contextual" determinants. Their presence or absence affects the processing that occurs in the neurons that are exposed to them. But on a more AI level, the importance of context has been repeatedly demonstrated by psychologists. I believe that context is a learned phenomena. Children have no construct of context, and thus, they are often able to make conclusions that may be associationally feasible, yet clearly contrary to the context of presentation. Context in developmental psychology has been approached from a more motivational point of view. Maslowe's hierarchies and the extensive work into "values" are all defining different levels of context. Whereas an associational network may (at least in my book) involve excitatory nodal influences, context involves inhibitory control over the nodes in the associational network. In my view, associational networks only know (always associated), (often associated), and (weak association). (Never associated) dictates that no association exists by default. A contextual network knows only that the following states can occur between concepts: (never can occur) and (rarely occurs). These can be defined using logical calculus and learning theory. The associational links are solely determined by event pairing and is a more dynamic event. Contextual networks are more stable and can be the result of learning as well as by introspective analysis of the associational links. As you can see, I have few specific references on "context," and rely upon my own theory of context. I hope I've been of some help, and I would like to be kept apprised of your work. I suggest that if you want research evidence of some of the above, that you examine indices on the subjects I mentioned. Again, Good luck, Philip Kahn ------------------------------ Date: 6 Feb 84 7:18:25-PST (Mon) From: harpo!ulysses!mhuxl!eagle!hou5h!hou5a!hou5d!mat @ Ucb-Vax Subject: Re: brain, a parallel processor ? Article-I.D.: hou5d.809 See the Feb. Scientific American for an article on typists and speed. There is indeed evidence for a high degree of parallelism even in SIMILAR tasks. Mark Terribile ------------------------------ Date: Wed, 8 Feb 84 18:19:09 CST From: Doug Monk Subject: Re: AIList Digest V2 #11 Subject : Mike Brzustowicz's 'tip of the tongue' as parallel process Rather than being an example of parallel processing, the 'tip of the tongue' phenomenon is probably more an example of context switch, where the attempt to recall the information displaces it temporarily, due to too much pressure being brought to bear. ( Perhaps a form of performance anxiety ? ) Later, when the pressure is off, and the processor has a spare moment, a smaller recall routine can be used without displacing the information. This model assumes that concentrating on the problem causes more of the physical brain to be involved in the effort, thus perhaps 'overlaying' the data desired. Once a smaller recall routine is used, the recall can actually be performed. Doug Monk ( bro.rice@RAND-RELAY ) ------------------------------ Date: 6 Feb 84 19:58:33-PST (Mon) From: ihnp4!ihopa!dap @ Ucb-Vax Subject: Re: parallel processing in the brain Article-I.D.: ihopa.153 If you consider pattern recognition in humans when constrained to strictly sequential processing, I think we are MUCH slower than computers. In other words, how long do you think it would take a person to recognize a letter if he could only inquire as to the grayness levels in different pixels? Of course, he would not be allowed to "fill in" a grid and then recognize the letter on the grid. Only a strictly algorithmic process would be allowed. The difference here, as I see it, is that the human mind DOES work in parallel. If we were forced to think sequentially about each pixel in our field of vision, we would become hopelessly bogged down. It seems to me that the most likely way to simulate such a process is to have a HUGE number of VERY dumb processors in a heirarchy of "meshes" such that some small number of processors in common localities in a low level mesh would report their findings to a single processor in the next higher level mesh. This processor would do some very quick, very simple calculations and pass its findings on the the next higher level mesh. At the top level, the accumulated information would serve to recognize the pattern. I'm really speaking off the top of my head since I'm no AI expert. Does anybody know if such a thing exists or am I way off? Darrell Plank BTL-IH ihopa!dap [Researchers at the University of Maryland and at the University of Massachusetts, among others, have done considerable work on "pyramid" and "processing cone" vision models. The multilayer approach was also common in perceptron-based pattern recognition, although very little could be proven about multilayer networks. -- KIL] ------------------------------ End of AIList Digest ******************** 11-Feb-84 01:04:13-PST,21483;000000000001 Mail-From: LAWS created at 10-Feb-84 23:11:50 Date: Fri 10 Feb 1984 23:05-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #16 To: AIList@SRI-AI AIList Digest Saturday, 11 Feb 1984 Volume 2 : Issue 16 Today's Topics: Lab Description - New UCLA AI Lab, Report - National Computing Environment for Academic Research, AI Journal - New Developments in the Assoc. for Computational Linguistics, Course - Organization Design, Conference - Natural Language and Logic Programming & Systems Science ---------------------------------------------------------------------- Date: Fri, 3 Feb 84 22:57:55 PST From: Michael Dyer Subject: New UCLA AI Lab Announcing the creation of a new Lab for Artificial Intelligence Research at UCLA. Just recently, the UCLA CS department received a private foundation grant of $450,000 with $250,000 matching funds from the School of Engineering and Applied Sciences to create a Laboratory for Artificial Intelligence Research. The departmental chairman as well as the dean strongly support this effort and are both committed to the growth of AI at UCLA. In addition, UCLA has been chosen as the site of the next International Joint Conference on Artificial Intelligence (IJCAI-85) in August, 1985. UCLA is second in the nation among public research universities and in the top six overall in quality of faculty, according to a new national survey of 5,000 faculty and 228 universities. In a two year study (conducted by the Conference Board of the Associated Research Councils, consisting of the American Council of Learned Societies, the American Council on Education, the National Research Council and the Social Science Research Council) the UCLA Computer Science Dept. tied for sixth place with U. of Ill., after Stanford, MIT, CMU, UC Berkeley, and Cornell. The UCLA CS department is the recipient (in 1982) of a $3.6 million five-year NSF Coordinated Experimental Research grant, augmented by a $1.5 million award from DARPA. Right now the AI lab includes a dozen Apollo DN300 workstations on an Apollo Domain ring network. This ring is attached via an ethernet gate to the CS department LOCUS network of 20 Vax 750s and a 780. UCLA is on the Arpanet and CSNet. Languages include Prolog and T (a Scheme-based dialect of lisp). A number of DN320s, DN460s and a color Apollo (DN660) are on order and will be housed in a new area being reserved for graduate AI research. One Vax 750 on the LOCUS net and 10 Apollos will be reserved for graduate AI instruction. Robotics and vision equipment is also being acquired. The CS dept is seeking an assist. prof. (tenure track) in the area of AI, with preference for vision, robotics, problem-solving, expert systems, learning, and simulation of cognitive processes. The new AI faculty member will be able to direct expenditure of a portion of available funds. (Interested AI PhDs, reply to Michael Dyer, CS dept, UCLA, Los Angeles, CA 90024. Arpanet: dyer@ucla-cs). Our AI effort is new, but growing, and includes the following faculty: Michael Dyer: natural language processing, cognitive modeling. Margot Flowers: reasoning, argumentation, belief systems. Judea Pearl: theory of heuristics, search, expert systems. Alan Klinger: signal processing, pattern recognition, vision. Michel Melkanoff: CAD/CAM, robotics. Stott Parker: logic programming, databases. ------------------------------ Date: 26 Jan 84 14:22:30-EDT (Thu) From: Kent Curtis Subject: A National Computing Environment for Academic Research The National Science Foundation has released a report entitled "A National Computing Environment for Academic Research" prepared by an NSF Working Group on Computers for Research, Kent Curtis, Chairman. The table of contents is: Executive Summary I. The Role of Modern Computing in Scientific and Engineering Research with Special Concern for Large Scale Computation Background A. Summary of Current Uses and Support of Large Scale Computing for Research B. Critique of Current Facilities and Support Programs C. Unfilled Needs for Computer Support of Research II. The Role and Responsibilities of NSF with Respect to Modern Scientific Computing III. A Plan of Action for the NSF: Recommendations IV. A Plan of Action for the NSF: Funding Implications Bibliography Appendix Large-scale Computing Facilities If you are interested in receiving a copy of this report contact Kent Curtis, (202) 357-9747; curtis.nsf-cs@csnet-relay; or write Kent K. Curtis Div. of Computer Research NSF Washington, D.C. 20550 ------------------------------ Date: 10 Feb 84 09:35:51 EST (Fri) From: Journal Duties Subject: ~New Developments in the Assoc. for Computational Linguistics The AMERICAN JOURNAL OF COMPUTATIONAL LINGUISTICS -- Some New Developments The AMERICAN JOURNAL OF COMPUTATIONAL LINGUISTICS is the major international journal devoted entirely to computational approaches to natural language research. With the 1984 volume, its name is being changed to COMPUTATIONAL LINGUISTICS to reflect its growing international coverage. There is now a European chapter of the ASSOCIATION FOR COMPUTATIONAL LINGUISTICS and a growing interest in forming one in Asia. The journal also has many new people on its Editorial Staff. James Allen, of the University of Rochester, has taken over as Editor. The FINITE STRING Editor is now Ralph Weischedel of the University of Delaware. Lyn Bates of Bolt Beranek and Newman is the Book Review Editor. Michael McCord, now at IBM, remains as Associate Editor. With these major changes in editorial staffing, the journal has fallen behind schedule. In order to catch up this year, we will be publishing close to double the regular number of issues. The first issue for 1983, which was just mailed out, contains papers on "Paraphrasing Questions Using Given and New Information" by Kathleen McKeown and "Denotational Semantics for 'Natural' Language Question-Answering Programs" by Michael Main and David Benson. There is a lengthy review of Winograd's new book by Sergei Nirenburg and a comprehensive description of the new Center for the Study of Language and Information at Stanford University. Highlights of the forthcoming 1983 AJCL issues: - Volume 9, No. 2 (expected March '84) will contain, in addition to papers on "Natural Language Access to Databases: Interpreting Update Requests" by Jim Davidson and Jerry Kaplan and "Treating Coordination in Logic Grammars" by Veronica Dahl and Michael McCord, will be accompanied by a supplement: a Directory of Graduate Programs in Computational Linguistics. The directory is the result of two years of surveys, and provides a fairly complete listing of programs available internationally. - Volume 9, Nos. 3 and 4 (expected June '84) will be a special double issue on Ill-Formed Input. The issue will cover many aspects of processing ill-formed sentences from syntactic ungrammaticality to dealing with inaccurate reference. It will contain papers from many of the research groups that are working on such problems. We will begin publishing Volume 10 later in the summer. In addition to the regular contributions, we are planning a special issue on the mathematical properties of grammatical formalisms. Ray Perrault (now at SRI) will be guest editor for the issue, which will contain papers addressing most of the recent developments in grammatical formalisms (e.g., GPSG, Lexical-Function Grammars, etc). Also in the planning stage is a special issue on Machine Translation that Jonathan Slocum is guest editing. With its increased publication activity in 1984, COMPUTATIONAL LINGUISTICS can provide authors with an unusual opportunity to have their results published in the international community with very little delay. A paper submitted now (early spring '84) could actually be in print by the end of the year, provided that major revisions need not be made. Five copies of submissions should be sent to: James Allen, CL Editor Dept. of Computer Science The University of Rochester Rochester, NY 14627, USA Subscriptions to COMPUTATIONAL LINGUISTICS come with membership in the ASSOCIATION FOR COMPUTATIONAL LINGUISTICS, which still is only $15 per year. As a special bonus to new members, those who join the ACL for 1984 before August will receive the special issue on Ill-Formed Input, even though it is formally part of the volume for 1983. To become a member, simply send your name, address and a check made out to the Association for Computational Linguistics to: Don Walker, ACL membership SRI International 333 Ravenswood Avenue Menlo Park, CA 94025, USA People in Europe or with Swiss accounts can pay an equivalent value in Swiss francs, by personal check in their own currency, or by a banker's draft that credits account number 141.880.LAV at the Union Bank of Switzerland, 8 rue de Rhone, CH-1211 Geneva 11, SWITZERLAND; send the statement with payment or with a copy of the bank draft to: Mike Rosner, ACL ISSCO 54, route des Acacias CH-1227 Geneva, SWITZERLAND ------------------------------ Date: Wednesday, 8 February 1984, 14:28-EST From: Gerald R. Barber Subject: Course Announcement: Organization Design [Forwarded by SASW@MIT-MC.] The following is an announcement for a course that Tom Malone and I are organizing for this spring term. Anyone who is interested can come to the course or contact: Tom Malone Malone@XX E53-307, x6843, or Jerry Barber Jerryb@OZ NE43-809, x5871 Course Announcement 15.963 Oranization Design Wednesdays, 2:30 - 5:30 p.m, E51-016 Prof. Thomas Malone In this graduate seminar we will review research from a number of fields, identifying general principles of organization design that apply to many kinds of information processing systems, including human organizations and computer systems. This novel approach will integrate examples and theories from computer science, artificial intelligence, organization theory and economics. The seminar will also include discussion of several special issues that arise when these general principles are applied to designing organizations that include both people and computers. A partial list of topics includes: I. Introduction A. What is an organization? Scott, March & Simon, Etzioni, etc B. What is design? Simon: Science of Design II. Alternative Organizational Designs A. Markets Computer Systems: Contract Nets, Enterprise Organizational Theories: Simon, Arrow, Hurwicz B. Hierachies Computer Systems: Structured programming, inheritance hierarchies Organizational Theories: Simon, March, Cyert, Galbraith, Williamson C. Cooperating experts (or teams) Computer Systems: Hearsay, Ether, Actors, Smalltalk, Omega Organizational Theories: Marschak & Radner, Minsky & Papert III. Integrating Computer Systems and Human Organizations A. Techniques for analyzing organizational needs Office Analysis Methodology, Critical Success Factors, Information Control Networks, Sociotechnical systems B. Possible technologies for supporting organizational problem-solving Computer conferencing, Knowledge-based systems ------------------------------ Date: Thu 2 Feb 84 20:35:47-PST From: Pereira@SRI-AI Subject: Natural Language and Logic Programming Call for Papers International Workshop On Natural Lanugage Understanding and Logic Programming Rennes, France - September 18-20, 1984 The workshop will consider fundamental principles and important innovations in the design, definition, uses and extensions of logic programming for natural language understanding and, conversely, the adequacy of logic programming to express natural language grammar formalisms. The topics of interest are: * Formal representations of natural language * Logic grammar formalisms * Linguistic aspects (anaphora, coordination,...) * Analysis methods * Natural language generation * Uses of techniques for logic grammars (unification) in other grammar formalisms * Compilers and interpreters for grammar formalisms * Text comprehension * Applications: natural-language front ends (database interrogation, dialogues with expert systems...) Conference Chairperson Veronica Dahl Simon Fraser University, Burnaby B.C. V5A 1S6 Canada Program Committee H. Abrahamson (UBC, Canada) F. Pereira (SRI, USA) A. Colmerauer (GIA, France) L. Pereira (UNL, Portugal) V. Dahl (Simon Fraser U., Canada) P. Sabatier (CNRS, France) P. Deransart (INRIA, France) P. Saint-Dizier (IRISA, France) M. Gross (LADL, France) C. Sedogbo (Bull, France) M. McCord (IBM, USA) Sponsored by: IRISA, Groupe BULL, INRIA Deadlines: April 15: Submission of papers in final form June 10: Notification of acceptance to authors July 10: Registration in the Workshop Submission of papers: Papers should contain the following items: abstract and title of paper, author name, country, affiliation, mailing address and phone (or telex) number, one program area and the following signed statement: ``The paper will be presented at the Workshop by one of the authors''. Summaries should explain what is new or interesting abount the work and what has been accomplished. Papers must report recent and not yet published work. Please send 7 copies of a 5 to 10 page single spaced manuscript, including a 150 to 200 word abstract to: -- Patrick Saint-Dizier Local Organizing Committee IRISA - Campus de Beaulieu F-35042 Rennes CEDEX - France Tel: (99)362000 Telex: 950473 F ------------------------------ Date: Sat, 4 Feb 84 10:18 cst From: Bruce Shriver Subject: call for papers announcement Eighteenth Annual HAWAII INTERNATIONAL CONFERENCE ON SYSTEM SCIENCES JANUARY 2-4, 1985 / HONOLULU, HAWAII This is the eighteenth in a series of conferences devoted to advances in information and system sciences. The conference will encompass developments in theory or practice in the areas of COMPUTER HARDWARE and SOFTWARE, and advanced computer systems applications in selected areas. Special emphasis will be devoted to MEDICAL INFORMATION PROCESSING, computer-based DECISION SUPPORT SYSTEMS for upper-level managers in organizations, and KNOWLEDGE-BASED SYSTEMS. CALL FOR PAPERS Papers are invited in the preceeding and related areas and may be theoretical, conceptual, tutorial or descriptive in nature. The papers submitted will be refereed and those selected for conference presentation will be printed in the CONFERENCE PROCEEDINGS; therefore, papers submitted for presentation must not have been previously presented or published. Authors of selected papers are expected to attend the conference to present and discuss the papers with attendees. Relevant topics include: Deadlines HARDWARE * Abstracts may be submitted to track * Distributed Processing chairpersons for guidance and indication * Mini-Micro Systems of appropriate content by MAY 1, 1984. * Interactive Systems (Abstract is required for Medical * Personal Computing Information Processing Track.) * Data Communication * Full papers must be mailed to appropriate * Graphics track chairperson by JULY 6, 1984. * User-Interface Technologies * Notification of Accepted papers will be mailed to the author on or before SOFTWARE SEPTEMBER 7, 1984. * Software Design Tools & * Final papers in camera-ready form will Techniques be due by OCTOBER 19, 1984. * Specification Techniques * Testing and Validation * Performance Measurement & Instructions for Submitting Papers Modeling 1. Submit three copies of the full paper, * Formal Verification not to exceed 20 double-spaced pages, * Management of Software including diagrams, directly to the Development appropriate track chairperson listed below, or if in doubt, to the conference APPLICATIONS co-chairpersons. * Medical Information 2. Each paper should have a title page Processing Systems which includes the title of the paper, * Computer-Based Decision full name of its author(s), affiliat- Support Systems ation(s), complete address(es), and * Management Information Systems telephone number(s). * Data-Base Systems for 3. The first page should include the Decision Support title and a 200-word abstract of the * Knowledge-Based Systems paper. SPONSORS The Eighteenth Annual Hawaii International Conference on System Science is sponsored by the University of Hawaii and the University of Southwestern Louisiana, in cooperation with the ACM and the IEEE Computer Society. HARDWARE All Other Papers Edmond L. Gallizzi Papers not clearly within one of the HICSS-18 Track Chairperson aforementioned tracks should be mailed Eckerd College to: St. Petersberg, FL 33733 Ralph H. Sprague, Jr. (813) 867-1166 HICSS-18 Conference Co-chairperson College of Business Administration SOFTWARE University of Hawaii Bruce D. Shriver 2404 Maile Way, E-303 HICSS-18 Track Chairperson Honolulu, HI 96822 Computer Science Dept. (808)948-7430 U. of Southwestern Louisiana P. O. Box 44330 Lafayette, LA 70504 Conference Co-Chairpersons (318) 231-6284 RALPH H. SPRAGUE, JR. BRUCE D. SHRIVER DECISION SUPPORT SYSTEM & KNOWLEDGE-BASED SYSTEMS Contributing Sponsor Coordinator Joyce Elam RALPH R. GRAMS HICSS-18 Track Chairperson College of Medicine Dept. of General Business Department of Pathology BEB 600 University of Florida U. of Texas at Austin Box J-275 Austin, TX 78712 Gainesville, FL 32610 (512) 471-3322 (904) 392-4571 MEDICAL INFORMATION PROCESSING FOR FURTHER INFORMATION Terry M. Walker Concerning Conference Logistics HICSS-18 Track Chairperson NEM B. LAU Computer Science Dept. HICSS-18 Conference Coordinator U. of Southwestern Louisiana Center for Executive Development P. O. Box 44330 College of Business Administration Lafayette, LA 70504 University of Hawaii (318) 231-6284 2404 Maile Way, C-202 Honolulu, HI 96822 (808) 948-7396 Telex: RCA 8216 UHCED Cable: UNIHAW The HICSS conference is a non-profit activity organized to provide a forum for the interchange of ideas, techniques, and applications among practitioners of the system sciences. It maintains objectivity to the systems sciences without obligation to any commercial enterprise. All attendees and speakers are expected to have their respective companies, organizations or universities bear the costs of their expenses and registration fees. ------------------------------ End of AIList Digest ******************** 11-Feb-84 21:33:00-PST,19687;000000000001 Mail-From: LAWS created at 11-Feb-84 21:27:53 Date: Sat 11 Feb 1984 20:58-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #17 To: AIList@SRI-AI AIList Digest Sunday, 12 Feb 1984 Volume 2 : Issue 17 Today's Topics: Jargon - Glossary of NASA Terminology, Humor - Programming Languages ---------------------------------------------------------------------- Date: 23 Jan 84 7:41:17-PST (Mon) From: hplabs!hao!seismo!flinn @ Ucb-Vax Subject: Glossary of NASA Terminology [Reprinted from the Space Digest by permission of the author. This strikes me as an interesting example of a "natural sublanguage." It does not reflect the growth and change of NASA jargon, however: subsequent discussion on the Space Digest indicates that many of the terms date back eight years and many newer terms are missing. The author and others are continuing to add to the list. -- KIL] I've been collecting examples of the jargon in common use by people at NASA Headquarters. Here is the collection so far: I have not made any of these up. I'd be glad to hear of worthy additions to the collection. The 'standard NASA noun modifiers' are nouns used as adjectives in phrases like 'science community' or 'planetary area.' Definitions have been omitted for entries whose meaning ought to be clear. -- Ted Flinn Action Item Actors in the Program Ancillary Ankle: 'Get your ankles bitten' = running into unexpected trouble. Ant: 'Which ant is steering this log?' = which office is in charge of a project. Appendice (pronounced ap-pen-di-see): some people, never having seen a document with only one appendix, think that this is the singular of 'appendices.' Area: Always as 'X Area,' where X is one of the standard NASA noun modifiers. Asterick: pronounced this way more often than not. Back Burner Bag It: 'It's in the bag' = it's finished. Ball of Wax Baseline: verb or noun. Basis: Always as 'X Basis,' where X is one of the standard NASA noun modifiers. Bean Counters: financial management people. Bed: 'Completely out of bed' = said of people whose opinions are probably incorrect. Belly Buttons: employees. Bench Scientists Bend Metal: verb, to construct hardware. Bending Your Pick: unrewarding activity. Bent Out of Shape: disturbed or upset, of a person. Big Picture Big-Picture Purposes Bite the Bullet Big-Ticket Item: one of the expensive parts. Black-belt Bureaucrat: an experienced and knowledgable government employee. Bless: verb, to approve at a high level of management. Blow One's Skirts Up: usually negative: 'that didn't blow their skirts up' = that didn't upset them. Blow Smoke: verb, to obfuscate. Blown Out of the Water Bottom Line Bounce Off: to discuss an idea with someone else. Brassboard (see Breadboard). Breadboard (see Brassboard). Bullet: one of the paragraphs or lines on a viewgraph, which are *never* numbered, but always labelled with a bullet. Bulletize: to make an outline suitable for a viewgraph. Bureaucratic Hurdles Burn: verb, to score points off a competitor. Burning Factor: one of the critical elements. Calibrate: verb, to judge the capabilities of people or organizations. Camel's Nose in the Tent Can of Worms Canned: finished, as 'it's in the can.' Can't Get There From Here. Capture a Mission: verb, to construct a launch vehicle for a space flight. Carve Up the Turkey Caveat: usually a noun. Centers: 'on N-week centers' = at N-week intervals. Choir, Preaching to the Clock is Ticking = time is getting short. Code: Every section at NASA centers or Headquarters has a label consisting of one or more letters or numbers, and in conversations or less formal memos, sections are always referred to by the code rather than the name: Code LI, Code 931, Code EE, etc. Commonality Community: 'X Community,' where X is one of the standard NASA noun modifiers. Concept: 'X Concept,' where X is one of the standard NASA noun modifiers. Concur: verb, to agree. Configure: verb. Constant Dollars: cost without taking inflation into account (see Real-Year Dollars). Contract Out Core X: The more important parts of X, where X is one of the nouns used as modifiers. Correlative Cost-Benefit Tradeoff Cross-Cut: verb, to look at something a different way. Crump: transitive verb, to cause to collapse. Crutch: flimsy argument. Cut Orders: to fill out a travel order form; left over from the days when this was done with mimeograph stencils. Cutting Edge Data Base Data Dump: a report made to others, usually one's own group. Data Point: an item of information. Debrief: transitive verb, to report to one's own staff after an outside meeting. Deep Yoghurt: bad trouble. Definitize: verb, to make precise or definite. De-integrate: verb, to take apart (not dis-). De-lid: verb, to take the top off an instrument. Delta: an increment to cost or content. Descope: verb, to redesign a project as a result of budget cuts (not the opposite of scope, q.v.). Development Concept Dialog: transitive verb. Disadvantage: transitive verb. Disgruntee: non-NASA person unhappy with program decisions. Dog's Breakfast Dollar-Limited Driver: an item making up a significant part of cost or schedule: 'X is the cost driver.' Drop-Dead Date: the real deadline; see 'hard deadline.' Ducks in a Row Egg on One's Face End Item: product. End-Run the System End to End Extent to Which Extramural Facilitize: verb, to make a facility out of something. Factor in: verb. Feedback: reaction of another section or organization to a proposition. Fill This Square Finalize Finesse The System First Cut: preliminary estimate. Fiscal Constraints Flag: verb, to make note of something for future reference. Flagship Program Flex the Parameters Flux and Change What Will Fly: 'see it if will fly.' Folded In: taken into account. Forest: miss the f. for the trees. Forgiving, unforgiving: of a physical system. Front Office Full-Up: at peak level. Future: promise or potential, as, 'a lot of potential future.' Futuristic Gangbusters Glitch Grease the Skids Green Door: 'behind the green door' = in the Administrator's offices. Go to Bat For Goal: contrasted to 'objective,' q.v. Grabber Gross Outline: approximation. Ground Floor Group Shoot = brainstorming session. Guidelines: always desirable to have. Guy: an inanimate object such as a data point. Hack: 'get a hack on X' = make some kind of estimate. Hard Copy: paper, as contrasted to viewgraphs. Hard Deadline: supposed deadline; never met. Hard Over: intransigent. Head Counters: personnel office staff. Hit X Hard: concentrate on X. Hoop: a step in realizing a program: 'yet to go through this hoop.' Humanoid Hypergolic: of a person: intransigent or upset in general. Impact: verb. Implement: verb. In-House Initialize Innovative Intensive: always as X-intensive. Intercompare: always used instead of 'compare.' Issue: always used instead of 'problem.' Key: adj., of issues: 'key issue; not particularly key'. Knickers: 'get into their knickers' = to interfere with them. Laicize: verb, to describe in terms comprehensible to lay people. Lashup = rackup. Lay Track: to make an impression on management ('we laid a lot of track with the Administrator'). Learning Curve Liaise: verb. Limited: always as X-limited. Line Item Link Calculation Liberate Resources: to divert funds from something else. Looked At: 'the X area is being looked at' = being studied. Loop: to be in the loop = to be informed. Love It! exclamation of approval. Low-Cost Machine = spacecraft. Man-Attended Experiment Marching Orders Matrix Micromanagement = a tendency to get involved in management of affairs two or more levels down from one's own area of responsibility. Milestone Mission Definition Mode: 'in an X mode.' Model-Dependent Muscle: 'get all the muscle into X' Music: 'let's all read from the same sheet of music.' Necessitate Nominal: according to expectation. Nominative: adj., meaning unknown. Nonconcur: verb, to disagree. Numb Nut: unskilled or incapable person. Objective: as contrasted with 'goal' (q.v.) Overarching Objective Oblectation Off-Load: verb. On Board: 'Y is on board' = the participation of Y is assured. On-Boards: employees or participants. On Leave: on vacation. On the Part Of On Travel: out of town. Open Loop Out-of-House Over Guidelines Ox: 'depends on whose ox is gored.' Package Paradigm Parking Orbit: temporary assignment or employment. Pathfinder Studies Pedigree: history of accumulation of non-NASA support for a mission. Peg to Hang X On Pie: 'another slice through this same pie is...' Piece of the Action Ping On: verb, to remind someone of something they were supposed to do. Pitch: a presentation to management. Placekeeper Planning Exercise Pony in This Pile of Manure Somewhere = some part of this mess may be salvageable. Posture Pre-Posthumous Prioritize Priority Listing Problem Being Worked: 'we're working that problem.' Problem Areas Product = end item. Programmatic Pucker Factor: degree of apprehension. Pull One's Tongue Through One's Nose: give someone a hard time. Pulse: verb, as, 'pulse the system.' Quick Look Rackup = lashup. Rainmaker: an employee able to get approval for budget increases or new missions. Rapee: a person on the receiving end of an unfavorable decision. Rattle the Cage: 'that will rattle their cage.' Real-Year Dollars: cost taking inflation into account, as contrasted with 'constant dollars.' Reclama Refugee: a person transferred from another program. Report Out: verb, used for 'report.' Resources = money. Resource-Intensive = expensive. ROM: 'rough order of magnitude,' of estimates. Rubric Runout Sales Pitch Scenario Scope: verb, to attempt to understand something. Scoped Out: pp., understood. Secular = non-scientific or non-technological. Self-Serving Sense: noun, used instead of 'consensus.' Shopping List Show Stopper Sign Off On something = approve. Space Cadets: NASA employees. Space Winnies or Wieners: ditto, but even more derogatory. X-Specific Speak to X: to comment on X, where X is a subject, not a person. Specificity Speed, Up To Spinning One's Wheels Spooks: DOD of similar people from other agencies. Staff: verb. Standpoint: 'from an X standpoint' Statussed: adj., as, 'that has been statussed.' Strap On: verb, to try out: 'strap on this idea...' Strawman String to One's Bow Street, On The: distributed outside one's own office. Stroking Structure: verb. Subsume Success-Oriented: no provision for possible trouble. Surface: verb, to bring up a problem. Surveille: verb. Suspense Date: the mildest form of imaginary deadline. Tail: to have one's tail in a crack = to be upset or in trouble. Tall Pole in the Tent: data anomaly. Tar With the Same Brush On Target Task Force Team All Set Up Tickler = reminder. Tiger Team Time-Critical: something likely to cause schedule trouble. Time Frame Torque the System Total X, where X is one of the standard NASA noun modifiers. Total X Picture Truth Model Unique Update: noun or verb. Up-Front: adj. Upscale Upper Management Vector: verb. Vector a Program: to direct it toward some objective. Ventilate the Issues: to discuss problems. Versatilify: verb, to make something more versatile. Viable: adj., something that might work or might be acceptable. Viewgraph: always mandatory in any presentation. Viz-a-Viz WAG = wild-assed guess. Wall to Wall: adj., pervasive. Watch: 'didn't happen on my watch...' Water Off a Duck's Back Waterfall Chart: one way of present costs vs. time. I'm Not Waving, I'm Drowning Wedge; Planning Wedge: available future-year money. Been to the Well Where Coming From Whole Nine Yards X-Wide X-wise Workaround: way to overcome a problem. Wrapped Around the Axle: disturbed or upset. ------------------------------ Date: Wed 8 Feb 84 07:14:34-CST From: Werner Uhrig Subject: The Best Languages in Town!!! (forwarded from USENET) [Reprinted from the UTexas-20 bboard.] From: bradley!brad Feb 6 16:56:00 1984 Laidback with (a) Fifth By John Unger Zussman From Info World, Oct 4, 1982 Basic, Fortran, Cobol... These programming Languages are well known and (more or less) well loved throughout the computer in- dustry. There are numerous other languages, however, that are less well known yet still have ardent devotees. In fact, these little-known languages generally have the most fanatic admirers. For those who wish to know more about these obscure languages - and why they are obscure - I present the following catalog. SIMPLE ... SIMPLE is an acronym for Sheer Idiot's Mono Pur- pose Programming Lingusitic Environment. This language, developed at the Hanover College for Technological Misfits, was designed to make it impossible to write code with errors in it. The statements are, therefore confined to BEGIN, END, and STOP. No matter how you arrange the statements, you can't make a syntax error. Programs written in SIMPLE do nothing useful. Thus they achieve the results of programs written in other languages without the tedious, frustrating process of testing and debug- ging. SLOBOL ... SLOBOL is best known for the speed, or lack of it, of its compiler. Although many compilers allow you to take a coffee break while they compile, SLOBOL compilers allow you to take a trip to Bolivia to pick up the coffee. Forty-three pro- grammers are known to have died of boredom sitting at their ter- minals while waiting for a SLOBOL program to compile. Weary SLO- BOL programmers often turn to a related (but infinitely faster) language, COCAINE. VALGOL ... (With special thanks to Dan and Betsy "Moon Unit" Pfau) - From its modest beginnings in southern California's San Fernando Valley, VALGOL is enjoying a dramatic surge of populari- ty across the industry. VALGOL commands include REALLY, LIKE, WELL and Y$KNOW. Vari- ables are assigned with the =LIKE and =TOTALLY operators. Other operators include the "CALIFORNIA BOOLEANS", FERSURE, and NOWAY. Repetitions of code are handled in FOR-SURE loops. Here is a sam- ple VALGOL program: 14 LIKE, Y$KNOW (I MEAN) START %% IF PI A =LIKE BITCHEN AND 01 B =LIKE TUBULAR AND 9 C =LIKE GRODY**MAX 4K (FERSURE)**2 18 THEN 4I FOR I=LIKE 1 TO OH MAYBE 100 86 DO WAH + (DITTY**2) 9 BARF(I) =TOTALLY GROSS(OUT) -17 SURE 1F LIKE BAG THIS PROGRAM ? REALLY $$ LIKE TOTALLY (Y*KNOW) VALGOL is characterized by its unfriendly error messages. For example, when the user makes a syntax error, the interpreter displays the message, GAG ME WITH A SPOON! LAIDBACK ... Historically, VALGOL is a derivative of LAID- BACK, which was developed at the (now defunct) Marin County Center for T'ai Chi, Mellowness, and Computer Programming, as an alternative uo the more intense atmosphere in nearby silicon val- ley. The center was ideal for programmers who liked to soak in hot tubs while they worked. Unfortunately, few programmers could survive there for long, since the center outlawed pizza and RC Cola in favor of bean curd and Perrier. Many mourn the demise of LAIDBACK because of its reputation as a gentle and nonthreatening language. For Example, LAIDBACK responded to syntax errors with the message, SORRY MAN, I CAN'T DEAL WITH THAT. SARTRE ... Named after the late existential philosopher. SARTRE is an extremely unstructured language. Statements in SAR- TRE have no purpose; they just are there. Thus, SARTRE programs are left to define their own functions. SARTRE programmers tend to be boring and depressed and are no fun at parties. FIFTH ... FIFTH is a precision mathematical language in which the data types refer to quantity. The data types range from CC, OUNCE, SHOT, and JIGGER to FIFTH (hence the name of the language), LITER, MAGNUM, and BLOTTO. Commands refer to in- gredients such as CHABLIS, CHARDONNAY, CABERNET, GIN, VERMOUTH, VODKA, SCOTCH and WHATEVERSAROUND. The many versions of the FIFTH language reflect the sophisti- cation and financial status of its users. Commands in the ELITE dialect include VSOP and LAFITE, while commands in the GUTTER di- alect include HOOTCH and RIPPLE. The latter is a favorite of frustrated FORTH programmers who end up using the language. C- ... This language was named for the grade received by its creator when he submitted it as a class project in a graduate programming class. C- is best described as a "Low-Level" pro- gramming language. In fact, the language generally requires more C- statements than machine-code statements to execute a given task. In this respect, it is very similar to COBOL. LITHP ... This otherwise unremarkable labuage is dis- tinguished by the absence of an "s" in its character set. pro- grammers and users must substitute "TH". LITHP is said to useful in prothething lithtth. DOGO ... Developed at the Massachussettes Institute of Obedi- ence Training. DOGO heralds a new era of computer-literate pets. DOGO commands include SIT, STAY, HEEL and ROLL OVER. An innova- tive feature of DOGO is "PUPPY GRAPHICS", in which a small cocker spaniel occasionally leaves a deposit as he travels across the screen. Submitted By Ian and Tony Goldsmith ------------------------------ End of AIList Digest ******************** 11-Feb-84 21:37:55-PST,12973;000000000001 Mail-From: LAWS created at 11-Feb-84 21:35:53 Date: Sat 11 Feb 1984 21:32-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #18 To: AIList@SRI-AI AIList Digest Sunday, 12 Feb 1984 Volume 2 : Issue 18 Today's Topics: AI and Meteorology - Summary of Responses ---------------------------------------------------------------------- Date: 11 Jan 84 16:07:00-PST (Wed) From: ihnp4!fortune!rpw3 @ Ucb-Vax Subject: Re: AI and Weather Forecasting - (nf) Article-I.D.: fortune.2249 As far as the desirability to use AI on the weather, it seems a bit out of place, when there is rumoured to be a fairly straightforward (if INCREDIBLY cpu-hungry) thermodynamic relaxation calculation that gives very good results for 24 hr prediction. It uses as input the various temperature, wind, and pressure readings from all of the U.S. weather stations, including the ones cleverly hidden away aboard most domestic DC-10's and L-1011's. Starting with those values as boundary conditions, an iterative relaxation is done to fill in the cells of the continental atmospheric model. The joke is of course (no joke!), it takes 26 hrs to run on a Illiac IV (somebody from Ames or NOAS or somewhere correct me, please). The accuracy goes up as the cell size in the model goes down, but the runtime goes up as the cube! So you can look out the window, wait 2 hours, and say, "Yup, the model was right." My cynical prediction is that either (1) by the time we develop an AI system that does as well, the deterministic systems will have obsoleted it, or more likely (2) by the time we get an AI model with the same accuracy, it will take 72 hours to run a 24 hour forecast! Rob Warnock UUCP: {sri-unix,amd70,hpda,harpo,ihnp4,allegra}!fortune!rpw3 DDD: (415)595-8444 USPS: Fortune Systems Corp, 101 Twin Dolphins Drive, Redwood City, CA 94065 ------------------------------ Date: 19 Jan 84 21:52:42-EST (Thu) From: ucbtopaz!finnca1 @ Ucb-Vax Subject: Re: "You cant go home again" Article-I.D.: ucbtopaz.370 It seems to me (a phrase that is always a copout for the ill-informed; nonetheless, I proceed) that the real payoff in expert systems for weather forecasting would be to capture the knowledge of those pre-computer experts who, with limited data and even fewer dollars, managed to develop their pattern-recognition facilities to the point that they could FEEL what was happening and forecast accordingly. I was privileged to take some meteorology courses from such an oldster many years ago, and it was, alas, my short-sightedness about the computer revolution in meteorology that prevented me from capturing some of his expertise, to buzz a word or two. Surely not ALL of these veterans have retired yet...what a service to science someone would perform if only this experise could be captured before it dies off. ...ucbvax!lbl-csam!ra!daven or whatever is on the header THIS time. ------------------------------ Date: 15 Jan 84 5:06:29-PST (Sun) From: hplabs!zehntel!tektronix!ucbcad!ucbesvax.turner @ Ucb-Vax Subject: Re: Re: You cant go home again - (nf) Article-I.D.: ucbcad.1315 Re: finnca1@topaz's comments on weather forecasting Replacing expertise with raw computer power has its shortcomings--the "joke" of predicting the weather 24 hours from now in 26 hours of cpu time is a case in point. Less accurate but more timely forecasts used to be made by people with slide-rules--and where are these people now? It wouldn't surprise me if the 20th century had its share of "lost arts". Archaelogists still dig up things that we don't know quite how to make, and the technological historians of the next century might well be faced with the same sorts of puzzles when reading about how people got by without computers. Michael Turner (ucbvax!ucbesvax.turner) ------------------------------ Date: Wed 8 Feb 84 15:29:01-PST From: Ken Laws Subject: Summary of Responses The following is a summary of the responses to my AIList request for information on AI and meteorology, spatial and temporal reasoning, and related matters. I have tried to summarize the net messages accurately, but I may have made some unwarranted inferences about affiliations, gender, or other matters that were not explicit in the messages. The citations below should certainly not be considered comprehensive, either for the scientific literature as a whole or for the AI literature. There has been relevant work in pattern recognition and image understanding (e.g., the work at SRI on tracking clouds in satellite images), mapping, database systems, etc. I have not had time to scan even my own collection of literature (PRIP, CVPR, PR, PAMI, IJCAI, AAAI, etc.) for relevant articles, and I have not sought out bibliographies or done online searches in the traditional meteorological literature. Still, I hope these comments will be of use. ------------------ Bob Giansiracusa (Dept of Computer Science, Penn State Univ, 814-865-9507) reports that he and Alistair Frazer (Penn State Meteo Dept.) are advising two meteorology/CS students who want to do senior/masters theses in AI. They have submitted a proposal and expect to hear from NSF in a few months. Capt. Roslyn (Roz) J. Taylor, Applied AI Project Officer, USAF, @RADC, has read two of the Gaffney/Racer papers entitled "A Learning Interpretive Decision Algorithm for Severe Storm Forecasting." She found the algorithm to be a "fuzzy math"-based fine-tuning algorithm in much the same spirit as a Kalman filter. The algorithm might be useful as the numerical predictor in an expert system. Jay Glicksman of the Texas Instruments Computer Science Lab suggests that we check out Kawaguchi, E. et al. (1979) An Understanding System of Natural Language and Pictorial Pattern in the World of Weather Reports IJCAI-6 Tokyo, pp. 469-474 It does not provide many details and he has not seen a follow up, but the paper may give some leads. This paper is evidently related to the Taniguchi et al. paper in the 6th Pat. Rec. proceedings that I mentioned in my query. Dr. John Tsotsos and his students at the Univ. of Toronto Laboratory for Computational Medicine have been working for several years on the ALVEN system to interpret heart images in X-ray films. Dr. Tsotsos feels that the spatial and temporal reasoning capabilities of the system would be of use in meteorology. The temporal reasoning includes intervals, points, hierarchies, and temporal sampling considerations. He has sent me the following reports: R. Gershon, Y. Ali, and M. Jenkin, An Explanation System for Frame-based Knowledge Organized Along Multiple Dimensions, LCM-TR83-2, Dec. 1983. J.K. Tsotsos, Knowledge Organization: Its Role in Representation, Decision-making and Explanation Schemes for Expert Systems, LCM-TR83-3, Dec. 1983. J.K. Tsotsos, Representational Axes and Temporal Cooperative Processes, Preliminary Draft. I regret that I have found time for only a cursory examination of these papers, and so cannot say whether they will be useful in themselves for meteorology or only as a source of further references in spatial and temporal reasoning. Someone else in my group is now taking a look at them. Others papers from Dr. Tsotsos group may be found in: IJACI77-79-81, PRIP81, ICPR82, PAMI Nov.80, and IEEE Computer Oct. 83. Stuart C. Shapiro at the Univ. of Buffalo (SUNY) CS Dept. added the following reference on temporal reasoning: Almeida, M. J., and Shapiro, S. C., Reasoning about the temporal structure of narrative texts. Proceedings of the Fifth Annual Meeting of the Cognitive Science Society, Rochester, NY, 1983. Fanya S. Montalvo at MIT echoed my interest in * knowledge representations for spatial/temporal reasoning; * inference methods for estimating meteorological variables from (spatially and temporally) sparse data; * methods of interfacing symbolic knowledge and heuristic reasoning with numerical simulation models; * a bibliography or guide to relevant literature. She reports that good research along these lines is very scarce, but suggests the following: As far as interfacing symbolic knowlege with heuristic reasoning with numerical simulation, Weyhrauch's FOL system is the best formalism I've seen/worked-with to do that. Unfortunately there are few references to it. One is Filman, Lamping, & Montalvo in IJCAI'83. Unfortunately it was too short. There's a reference to Weyhrauch's Prolegomena paper in there. Also there is Wood's, Greenfeld's, and Zdybel's work at BBN with KLONE and a ship location database; they're no longer there. There's also Mark Friedell's Thesis from Case Western Reserve; see his SIGGRAPH'83 article, also references to Greenfeld & Yonke there. Oh, yes, there's also Reid Simmons, here at MIT, on a system connecting diagrams in geologic histories with symbolic descriptions, AAAI'83. The work is really in bits and pieces and hasn't really been put together as a whole working formalism yet. The issues are hard. Jim Hendler at Brown reports that Drew McDermott has recently written several papers about temporal and spatial reasoning. The best one on temporal reasoning was published in Cognitive Science about a year ago. Also, one of Drew's students at Yale recently did a thesis on spatial reasoning. David M. Axler, MSCF Applications Manager at Univ. of Pennsylvania, suggests: A great deal of info about weather already exists in a densely-encoded form, namely proverbs and traditional maxims. Is there a way that this system can be converted to an expert system, if for no other reason than potential comparison between the analysis it provides with that gained from more formal meteorological approaches? If this is of interest, I can provide leads to collections of weather lore, proverbs, and the like. If you're actually based at SRI, you're near several of the major folklore libraries and should have relatively easy access (California is the only state in the union with two grad programs in the field, one at Berkeley (under the anthro dept.), and one at UCLA) to the material, as both schools have decent collections. I replied: The use of folklore maxims is a good idea, and one fairly easy to build into an expert system for prediction of weather at a single site. (The user would have to enter observations such as "red sky at night" since pattern recognition couldn't be used. Given that, I suspect that a Prospector-style inference net could be built that would simultaneously evaluate hypotheses of "rain", "fog", etc., for multiple time windows.) Construction of the system and evaluation of the individual rules would make an excellent thesis project. Unfortunately, I doubt that the National Weather Service or other such organization would be interested in having SRI build such a "toy" system. They would be more interested in methods for tracking storm fronts and either automating or improving on the map products they currently produce. As a compromise, one project we have been considering is to automate a book of weather forecasting rules for professional forecasters. Such rule books do exist, but the pressures of daily forecasting are such that the books are rarely consulted. Perhaps some pattern recognition combined with some man-machine dialog could trigger the expert system rules that would remind the user of relevant passages. Dave liked the project, and suggested that there may be additional unofficial rule sources such as those used by the Farmer's Almanac publishers. Philip Kahn at UCLA is interested in pattern recognition, and recommends the book REMOTE SENSING: Optics and Optical Systems by Philip N. Slater Addison-Wesley Publ. Co., Reading, MA, 1980 for information on atmospherics, optics, films, testing/reliability, etc. Alex Pang at UCLA is doing some non-AI image processing to aid weather prediction. He is interested in hearing about AI and meteorology. Bill Havens at the University of British Columbia expressed interest, particularly in methods that could be implemented on a personal computer. Mike Uschold at Edinburgh and Noel Kropf at Columbia University (Seismology Lab?) have also expressed interest. ------------------ My thanks to all who replied. -- Ken Laws ------------------------------ End of AIList Digest ******************** 15-Feb-84 00:12:57-PST,12576;000000000001 Mail-From: LAWS created at 15-Feb-84 00:12:06 Date: Tue 14 Feb 1984 17:27-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #19 To: AIList@SRI-AI AIList Digest Wednesday, 15 Feb 1984 Volume 2 : Issue 19 Today's Topics: Requests - OPS5 & IBM LISP, LISP - Timings, Bindings - G. Spencer-Brown, Knowledge Acquisition - Regrets, Alert - 4-Color Problem, Brain Theory - Definition, Seminars - Analogy & Causal Reasoning & Tutorial Discourse ---------------------------------------------------------------------- Date: Mon 13 Feb 84 10:06:53-PST From: Ted Markowitz Subject: OPS5 query I'd like to find out some information on acquiring a copy of the OPS5 system. Is there a purchase price, is it free-of-charge, etc. Please send replies to G.TJM@SU-SCORE Thanks. --ted ------------------------------ Date: 1 Feb 1984 15:14:48 EST From: Robert M. Simmons Subject: lisp on ibm Can anyone give me pointers to LISP systems that run on IBM 370's under MVS? Direct and indirect pointers are welcome. Bob Simmons simmons@edn-unix ------------------------------ Date: 11 Feb 84 17:54:24 EST From: John Subject: Timings of LISPs and Machines I dug up these timings, they are a little bit out of date but seem a little more informative. They were done by Dick Gabriel at SU-AI in 1982 and passed along by Chuck Hedrick at Rutgers. Some of the times have been updated to reflect current machines by myself. These have been marked with the date of 1984. All machines were measured using the function - an almost Takeuchi function as defined by John McCarthy (defun tak (x y z) (cond ((not (< y x)) z) (t (tak (tak (1- x) y z) (tak (1- y) z x) (tak (1- z) x y))))) ------------------------------------------ (tak 18. 12. 6.) On 11/750 in Franz ordinary arith 19.9 seconds compiled On 11/780 in Franz with (nfc)(TAKF) 15.8 seconds compiled (GJC time) On Rutgers-20 in Interlisp/1984 13.8 seconds compiled On 11/780 in Franz (nfc) 8.4 seconds compiled (KIM time) On 11/780 in Franz (nfc) 8.35 seconds compiled (GJC time) On 11/780 in Franz with (ffc)(TAKF) 7.5 seconds compiled (GJC time) On 11/750 in PSL, generic arith 7.1 seconds compiled On MC (KL) in MacLisp (TAKF) 5.9 seconds compiled (GJC time) On Dolphin in InterLisp/1984 4.81 seconds compiled On Vax 11/780 in InterLisp (load = 0) 4.24 seconds compiled On Foonly F2 in MacLisp 4.1 seconds compiled On Apollo (MC68000) PASCAL 3.8 seconds (extra waits?) On 11/750 in Franz, Fixnum arith 3.6 seconds compiled On MIT CADR in ZetaLisp 3.16 seconds compiled (GJC time) On MIT CADR in ZetaLisp 3.1 seconds compiled (ROD time) On MIT CADR in ZetaLisp (TAKF) 3.1 seconds compiled (GJC time) On Apollo (MC68000) PSL SYSLISP 2.93 seconds compiled On 11/780 in NIL (TAKF) 2.8 seconds compiled (GJC time) On 11/780 in NIL 2.7 seconds compiled (GJC time) On 11/750 in C 2.4 seconds On Rutgers-20 in Interlisp/Block/84 2.225 seconds compiled On 11/780 in Franz (ffc) 2.13 seconds compiled (KIM time) On 11/780 (Diablo) in Franz (ffc) 2.1 seconds compiled (VRP time) On 11/780 in Franz (ffc) 2.1 seconds compiled (GJC time) On 68000 in C 1.9 seconds On Utah-20 in PSL Generic arith 1.672 seconds compiled On Dandelion in Interlisp/1984 1.65 seconds compiled On 11/750 in PSL INUM arith 1.4 seconds compiled On 11/780 (Diablo) in C 1.35 seconds On 11/780 in Franz (lfc) 1.13 seconds compiled (KIM time) On UTAH-20 in Lisp 1.6 1.1 seconds compiled On UTAH-20 in PSL Inum arith 1.077 seconds compiled On Rutgers-20 in Elisp 1.063 seconds compiled On Rutgers-20 in R/UCI lisp .969 seconds compiled On SAIL (KL) in MacLisp .832 seconds compiled On SAIL in bummed MacLisp .795 seconds compiled On MC (KL) in MacLisp (TAKF,dcl) .789 seconds compiled On 68000 in machine language .7 seconds On MC (KL) in MacLisp (dcl) .677 seconds compiled On SAIL in bummed MacLisp (dcl) .616 seconds compiled On SAIL (KL) in MacLisp (dcl) .564 seconds compiled On Dorado in InterLisp Jan 1982 (tr) .53 seconds compiled On UTAH-20 in SYSLISP arith .526 seconds compiled On SAIL in machine language .255 seconds (wholine) On SAIL in machine language .184 seconds (ebox-doesn't include mem) On SCORE (2060) in machine language .162 seconds (ebox) On S-1 Mark I in machine language .114 seconds (ebox & ibox) I would be interested if people who had these machines/languages available could update some of the timings. There also isn't any timings for Symbolics or LMI. John. ------------------------------ Date: Sun, 12 Feb 1984 01:14 EST From: MINSKY%MIT-OZ@MIT-MC.ARPA Subject: AIList Digest V2 #14 In regard to G Spencer Brown, if you are referring to author of the Laws of Form, if that's what it was called: I believe he was a friend of Bertrand Russell and that he logged out quite a number of years ago. ------------------------------ Date: Sun, 12 Feb 84 14:18:04 EST From: Brint Subject: Re: "You cant go home again" I couldn't agree more (with your feelings of regret at not capturing the expertise of the "oldster" in meterological lore). My dad was one of the best automotive diagnosticians in Baltimore until his death six years ago. His uncanny ability to pinpoint a problem's cause from external symptoms was locally legendary. Had I known then what I'm beginning to learn now about the promise of expert systems, I'd have spent many happy hours "picking his brain" with the (unfilled) promise of making us both rich! ------------------------------ Date: Mon 13 Feb 84 22:15:08-EST From: Jonathan Intner Subject: The 4-Color Problem To Whom It May Concern: The computer proof of the 4 - color problem can be found in Appel, K. and W. Haken ,"Every planar map is 4-colorable-1 : Discharging", "Every planar map is 4-colorable-2: Reducibility", Illinois Journal of Mathematics, 21, 429-567 (1977). I haven't looked at this myself, but I understand from Mike Townsend (a Prof here at Columbia) that the proof is a real mess and involves thousands of special cases. Jonathan Intner INTNER@COLUMBIA-20.ARPA ------------------------------ Date: 11 Feb 1984 13:50-PST From: Andy Cromarty Subject: Re: Brain, a parallel processor? What are the evidences that the brain is a parallel processor? My own introspection seem to indicate that mine is doing time-sharing. -- Rene Bach You are confusing "brain" with "mind". ------------------------------ Date: 10 Feb 1984 15:23 EST (Fri) From: "Daniel S. Weld" Subject: Revolving Seminar [Forwarded by SASW@MIT-MC.] Wednesday, February 15, 4:00pm 8th floor playroom Structure-Mapping: A Theoretical Framework for Analogy Dedre Gentner The structure-mapping theory of analogy describes a set of principles by which the interpretation of an analogy is derived from the meanings of its terms. These principles are characterized as implicit rules for mapping knowledge about a base domain into a target domain. Two important features of the theory are (1) the rules depend only on syntactic properties of the knowledge representation, and not on the specific content of the domains; and (2) the theoretical framework allows analogies to be distinguished cleanly from literal similarity statements, applications of general laws, and other kinds of comparisons. Two mapping principles are described: (1) Relations between objects, rather than attributes of objects, are mapped from base to target; and (2) The particular relations mapped are determined by @u(systematicity), as defined by the existence of higher-order relations. Psychological experiments supporting the theory are described, and implications for theories of learning are discussed. COMING SOON: Tomas Lozano-Perez, Jerry Barber, Dan Carnese, Bob Berwick, ... ------------------------------ Date: Mon 13 Feb 84 09:15:36-PST From: Juanita Mullen Subject: SIGLUNCH ANNOUNCEMENT - FEBRUARY 24, 1984 [Reprinted from the Stanford SIGLUNCH distribution.] Friday, February 24, 1984 LOCATION: Chemistry Gazebo, between Physical & Organic Chemistry 12:05 SPEAKER: Ben Kuipers, Department of Mathematics Tufts University TOPIC: Studying Experts to Learn About Qualitative Causal Reasoning By analyzing a verbatim protocol of an expert's explanation we can derive constraints on the conceptual framework used by human experts for causal reasoning in medicine. We use these constraints, along with textbook descriptions of physiological mechanisms and the computational requirements of successful performance, to propose a model of qualitative causal reasoning. One important design decision in the model is the selection of the "envisionment" version of causal reasoning rather than a version based on "causal links." The envisionment process performs a qualitative simulation, starting with a description of the structure of a mechanism and predicting its behavior. The qualitative causal reasoning algorithm is a step toward second-generation medical diagnosis programs that understand how the mechanisms of the body work. The protocol analysis method is a knowledge acquisition technique for determining the conceptual framework of new types of knowledge in an expert system, prior to acquiring large amounts of domain-specific knowledge. The qualitative causal reasoning algorithm has been implemented and tested on medical and non-medical examples. It will be the core of RENAL, a new expert system for diagnosis in nephrology, that we are now developing. ------------------------------ Date: 12 Feb 84 0943 EST (Sunday) From: Alan.Lesgold@CMU-CS-A (N981AL60) Subject: colloquium announcement [Forwarded from the CMU-C bboard by Laws@SRI-AI.] THE INTELLIGENT TUTORING SYSTEM GROUP LEARNING RESEARCH AND DEVELOPMENT CENTER UNIVERSITY OF PITTSBURGH AN ARCHITECTURE FOR TUTORIAL DISCOURSE BEVERLY P. WOOLF COMPUTER AND INFORMATION SCIENCE DEPARTMENT UNIVERSITY OF MASSACHUSETTS WEDNESDAY, FEBRUARY 15, 2:00 - 3:00, LRDC AUDITORIUM (SECOND FLOOR) Human discourse is quite complex compared to the present ability of machines to handle communication. Sophisticated research into discourse is needed before we can construct intelligent interactive systems. This talk presents recent research in the areas of discourse generation, with emphasis on teaching and tutoring dialogues. This talk describes MENO, a system where hand tailored rules have been used to generate flexible responses in the face of student failures. The system demonstrates the effectiveness of separating tutoring knowledge and tutoring decisions from domain and student knowledge. The design of the system suggests a machine theory of tutoring and uncovers some of the conventions and intuitions of tutoring discourse. This research is applicable to any intelligent interface which must reason about the users knowledge. ------------------------------ End of AIList Digest ******************** 17-Feb-84 09:38:25-PST,18372;000000000001 Mail-From: LAWS created at 17-Feb-84 09:36:31 Date: Fri 17 Feb 1984 09:22-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #20 To: AIList@SRI-AI AIList Digest Friday, 17 Feb 1984 Volume 2 : Issue 20 Today's Topics: Lisp - Timing Data Caveat, Bindings - G. Spencer Brown, Logic - Nature of Undecidability, Brain Theory - Parallelism, Expert Systems - Need for Perception, AI Culture - Work in Progress, Seminars - Learning & Automatic Deduction & Commonsense Reasoning ---------------------------------------------------------------------- Date: 16 Feb 1984 1417-PST From: VANBUER at USC-ECL.ARPA Subject: Timing Data Caveat A warning on the TAK performance testing: this code only exercises function calling and small integer arithmetic, and none of things most heavily used in "real" lisp programming: CONSing, garbage collection, paging (ai stuff is big after all). Darrel J. Van Buer ------------------------------ Date: Wed, 15 Feb 84 11:15:21 EST From: John McLean Subject: G. Spencer-Brown and undecidable propositions G. Spencer-Brown is very much alive. He spent several months at NRL a couple of years ago and presented lectures on his purported proof of the four color theorem. Having heard him lecture on several topics previously, I did not feel motivated to attend his lectures on the four color theorem so I can't comment on them first hand. Those who knew him better than I believe that he is currently at Oxford or Cambridge. By the way, he was not a friend of Russell's as far as I know. Russell merely said something somewhat positive about LAWS OF FORM. With respect to undecidability, I can't figure out what Charlie Crummer means by "undecidable proposition". The definition I have always seen is that a proposition is undecidable with respect to a set of axioms if it is independent, i.e,. neither the proposition nor its negation is provable. (An undecidable theory is a different kettle of fish altogether.) Examples are Euclid's 5th postulate with respect to the other 4, Goedel's sentence with respect to first order number theory, the continuum hypothesis with respect to set theory, etc. I can't figure out the claim that one can't decide whether an undecidable proposition is decidable or not. Euclid's 5th postulate, Goedel's sentence, and the continuum hypothesis have been proven to be undecidable. For simple theories, such as sentential logic (i.e., no quantifiers), there are even algorithms for detecting undecidability. John McLean ------------------------------ Date: Wed, 15 Feb 84 11:18:43 PST From: Charlie Crummer Subject: G. Spencer-Brown and undecidable propositions Thanks for the lead to G. S-B. I think I understand what he is driving at with THE LAWS OF FORM so I would like to see his alledged 4-color proof. Re: undecidability... Is it true that all propositions can be proved decidable or not with respect to a particular axiomatic system from WITHIN that system? My understanding is that this is not generally possible. Example (Not a proof of my understanding): Is the value of the statement "This statement is false." decidable from within Boolean logic? It seems to me that from within Boolean logic, i.e. 2-valued logic, all that would be seen is that no matter how long I crank I never seem to be able to settle down to a unique value. If this proposition is fed to a 2-valued logic program (written in PROLOG, LISP, or whatever language one desires) the program just won't halt. From OUTSIDE the machine, a human programmer can easily detect the problem but from WITHIN the Boolean system it's not possible. This seems to be an example of the halting problem. --Charlie ------------------------------ Date: 16 Feb 1984 12:22 EST (Thu) From: "Steven C. Bagley" Subject: Quite more than you want to know about George Spencer Brown Yes, Spencer Brown was associated with Russell, but since Lord Russell died recently (1970), I think it safe to assume that not ALL of his associates are dead, yet, at least. There was a brief piece about Spencer Brown in "New Scientist" several years ago (vol. 73, no. 1033, January 6, 1977, page 6). Here are two interesting quotes: "What sets him apart from the many others who have claimed a proof of the [four-color] theorem are his technique, and his personal style. Spencer Brown's technique rests on a book he wrote in 1964 called `Laws of Form.' George Allen and Unwin published it in 1969, on the recommendation of Bertrand Russell. In the book he develops a new algebra of logic -- from which the normal Boolean algebra (a means of representing propositions and arguments with symbols) can be derived. The book has had a mixed reputation, from `a work of genius' to `pretentious triviality.' It is certainly unorthodox, and mixes metaphysics and mathematics. Russell himself was taken with the work, and mentions it in his autobiography.... The style of the man is extravagant -- he stays at the Savoy -- and all-embracing. He was in the Royal Navy in the Second World War; has degrees in philosophy and psychology (but not mathematics); was a lecturer in logic at Christ Church College, Oxford; wrote a treatise on probability; a volume of poetry, and a novel; was a chief logic designer with Mullard Equipment Ltd where his patented design of a transistorised elevator logic circuit led to `Laws of Form'; has two world records for gliding; and presently lectures part-time in the mathematics department at the University of Cambridge while also managing his publishing business." I know of two reviews of "Laws of Form": one by Stafford Beer, the British cyberneticist, which appeared in "Nature," vol. 223, Sept 27, 1969, and the other by Lancelot Law Whyte, which was published in the British Journal of the Philosophy of Science, vol 23, 1972, pages 291-292. Spencer Brown's probability work was published in a book called "Probability and Scientific Inference", in the late 1950's, if my memory serves me correctly. There is also an early article in "Nature" called "Statistical Significance in Psychical Research", vol. 172, July 25, 1953, pp. 154-156. A comment by Soal, Stratton, and Trouless on this article appeared in "Nature" vol 172, Sept 26, 1953, page 594, and a reply by Spencer Brown immediately follows. The first sentence of the initial article reads as follows: "It is proposed to show that the logical form of the data derived from experiments in psychical research which depend upon statistical tests is such as to provide little evidence for telepathy, clairvoyance, precognition, psychokinesis, etc., but to give some grounds for questioning the practical validity of the test of significance used." Careful Spencer Brown watchers will be interested to note that this article lists his affliation as the Department of Zoology and Comparative Anatomy, Oxford; he really gets around. His works have had a rather widespread, if unorthodox, impact. Spencer Brown and "Laws of Form" are mentioned in Adam Smith's Powers of Mind, a survey of techniques for mind expansion, contraction, adjustment, etc., e.g., EST, various flavors of hallucinogens, are briefly noted in Aurthur Koestler's The Roots of Coincidence, which is, quite naturally enough, about probability, coincidence, and synchronicity, and are mentioned, again, in "The Dyadic Cyclone," by Dr. John C. Lilly, dolphin aficionado, and consciousness expander, extraordinaire. If this isn't an eclectic enough collection of trivia about Spencer Brown, keep reading. Here is quote from his book "Only Two Can Play This Game", written under the pseudonym of James Keys. "To put it bluntly, it looks as if the male is so afraid of the fundamentally different order of being of the female, so terrified of her huge magical feminine power of destruction and regeneration, that he doesn't look at her as she really is, he is afraid to accept the difference, and so has repressed into his unconscious the whole idea of her as ANOTHER ORDER OF BEING, from whom he might learn what he could not know of himself alone, and replaced her with the idea of a sort of second-class replica of himself who, because she plays the part of a man so much worse than a man, he can feel safe with because he can despise her." There are some notes at the end of this book (which isn't really a novel, but his reflections, written in the heat of the moment, about the breakup a love affair) which resemble parts of "Laws of Form": "Space is a construct. In reality there is no space. Time is also a construct. In reality there is no time. In eternity there is space but no time. In the deepest order of eternity there is no space....In a qualityless order, to make any distinction at all is at once to construct all things in embryo...." And last, I have no idea of his present-day whereabouts. Perhaps try writing to him c/o Cambridge University. ------------------------------ Date: Thu, 16 Feb 84 13:58:28 PST From: Charlie Crummer Subject: Quite more than you want to know about George Spencer Brown Thank you for the copious information on G. S-B. If I can't get in touch with him now, it will be because he does not want to be found. After the first reading of the first page of "The Laws of Form" I almost threw the book away. I am glad, however, that I didn't. I have read it several times and thought carefully about it and I think that there is much substance to it. --Charlie ------------------------------ Date: 15 Feb 84 2302 PST From: John McCarthy Subject: Serial or parallel It seems to me that introspection can tell us that the brain does many things serially. For example, a student with 5 problems on an examination cannot set 5 processes working on them. Indeed I can't see that introspection indicates that anything is done in parallel, although it does indicate that many things are done subconsciously. This is non-trivial, because one could imagine a mind that could set several processes going subconsciously and then look at them from time to time to see what progress they were making. On the other hand, anatomy suggests and physiological experiments confirm that the brain does many things in parallel. These things include low level vision processing and probably also low level auditory processing and also reflexes. For example, the blink reflex seems to proceed without thought, although it can be observed and in parallel with whatever else is going on. Indeed one might regard the blink reflex and some well learned habits as counter-examples to my assertion that one can't set parallel processes going and then observe them. All else seems to be conjecture. I'll conjecture that a division of neural activity into serial and parallel activities developed very early in evolution. For example, a bee's eye is a parallel device, but the bee carries out long chains of serial activities in foraging. My more adventurous conjecture is that primate level intelligence involves applying parallel pattern recognition processes evolve in connection with vision to records of the serial activities of the organism. The parallel processes of recognition are themselves subconscious, but the results have to take part in the serial activity. Finally, seriality seems to be required for coherence. An animal that seeks food by locomotion works properly only if it can go in one direction at a time, whereas a sea anemone can wave all its tentacles at once and needs only very primitive seriality that can spread in a wave of activity. Perhaps someone who knows more physiology can offer more information about the division of animal activity into serial and parallel kinds. ------------------------------ Date: Wed, 15 Feb 84 22:40:48 pst From: finnca1%ucbtopaz.CC@Berkeley Subject: Re: "You cant go home again" Date: Sun, 12 Feb 84 14:18:04 EST From: Brint I couldn't agree more (with your feelings of regret at not capturing the expertise of the "oldster" in meterological lore). My dad was one of the best automotive diagnosticians in Baltimore [...] Ah yes, the scarcest of experts these days: a truly competent auto mechanic! But don't you still need an expert to PERCEIVE the subtle auditory cues and translate them into symbolic form? Living in the world is a full time job, it seems. Dave N. (...ucbvax!ucbtopaz!finnca1) ------------------------------ Date: Monday, 13 Feb 1984 18:37:35-PST From: decwrl!rhea!glivet!zurko@Shasta Subject: Re: The "world" of CS [Forwarded from the Human-Nets digest by Laws@SRI-AI.] The best place for you to start would be with Sheri Turkle, a professor at MIT's STS department. She's been studying both the official and unofficial members of the computer science world as a culture/society for a few years now. In fact, she's supposed to be putting a book out on her findings, "The Intimate Machine". Anyone heard what's up with it? I thought it was supposed to be out last Sept, but I haven't been able to find it. Mez ------------------------------ Date: 14 Feb 84 21:50:52 EST From: Michael Sims Subject: Learning Seminar [Forwarded from the Rutgers bboard by Laws@SRI-AI.] MACHINE LEARNING BROWN BAG SEMINAR Title: When to Learn Speaker: Michael Sims Date: Wednesday, Feb. 15, 1984 - 12:00-1:30 Location: Hill Center, Room 254 (note new location) In this informal talk I will describe issues which I have broadly labeled 'when to learn'. Most AI learning investigations have concentrated on the mechanisms of learning. In part this is a reasonable consequence of AI's close relationship with the 'general process tradition' of psychology [1]. The influences of ecological and ethological (i.e., animal behavior) investigations have recently challenged this research methodology in psychology, and I believe this has important ramifications for investigations of machine learning. In particular, this influence would suggest that learning is something which takes place when an appropriate environment and an appropriate learning mechanism are present, and that it is inappropriate to describe learning by describing a learning mechanism without describing the environment in which it operates. The most cogent new issues which arise are the description of the environment, and the confronting of the issue of 'when to learn in a rich environment'. By a learning system in a 'rich environment' I mean a learning system which must extract the items to be learned from sensory input which is too rich to be exhaustively stored. Most present learning systems operate in such a restrictive environment that there is no question of what or when to learn. I will also present a general architecture for such a learning system in a rich environment, called a Pattern Directed Learning Model, which was motivated by biological learning systems. References [1] Johnston, T. D. Contrasting approaches to a theory of learning. Behavioral and Brain Sciences 4:125-173, 1981. ------------------------------ Date: Wed 15 Feb 84 13:16:07-PST From: Richard Treitel Subject: "Automatic deduction" and other stuff [Forwarded from the Stanford bboard by Laws@SRI-AI.] A reminder that the seminar on automatic reasoning / theorem proving /logic programming / mumble mumble mumble which I advertised earlier is going to begin shortly, under one title or another. It will tentatively be on Wednesdays at 1:30 in MJH301. If you wish to be on the mailing list for this, please mail to me or Yoni Malachi (YM@SAIL). But if you are already on Carolyn Talcott's mailing list for the MTC seminars, you will probably be included on the new list unless you ask not to be. For those interested specifically in the MRS system, we plan to continue MRS meetings, also on Weds., at 10:30, starting shortly. I expect to announce such meetings on the MRSusers distribution list. To get on this, mail to me or Milt Grinberg (GRINBERG@SUMEX). Note that MRSusers will contain other announcements related to MRS as well. - Richard ------------------------------ Date: Wed 15 Feb 84 Subject: McCarthy Lectures on Commonsense Knowledge [Forwarded from the Stanford CSLI newsletter by Laws@SRI.] MCCARTHY LECTURES ON THE FORMALIZATION OF COMMONSENSE KNOWLEDGE John McCarthy will present the remaining three lectures of his series (the first of the four was held January 20) at 3:00 p.m. in the Ventura Hall Seminar Room on the dates shown below. Friday, Feb. 17 "The Circumscription Mode of Nonmonotonic Reasoning" Applications of circumscription to formalizing commonsense facts. Application to the frame problem, the qualification problem, and to the STRIPS assumption. Friday, March 2 "Formalization of Knowledge and Belief" Modal and first-order formalisms. Formalisms in which possible worlds are explicit objects. Concepts and propositions as objects in theories. Friday, March 9 "Philosophical Conclusions Arising from AI Work" Approximate theories, second-order definitions of concepts, ascription of mental qualities to machines. ------------------------------ End of AIList Digest ******************** 22-Feb-84 17:14:20-PST,15264;000000000001 Mail-From: LAWS created at 22-Feb-84 17:09:04 Date: Wed 22 Feb 1984 16:28-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #21 To: AIList@SRI-AI AIList Digest Thursday, 23 Feb 1984 Volume 2 : Issue 21 Today's Topics: Waveform Analysis - EEG/EKG Request, Laws of Form - Comment, Review - Commercial NL Review in High Technology, Humor - The Adventures of Joe Lisp, Seminars - Computational Discovery & Robotic Planning & Physiological Reasoning & Logic Programming & Mathematical Expert System ---------------------------------------------------------------------- Date: Tue, 21 Feb 84 22:29:05 EST From: G B Reilly Subject: EEG/EKG Scoring Has anyone done any work on automatic scoring and interpretation of EEG or EKG outputs? Brendan Reilly [There has been a great deal of work in these areas. Good sources are the IEEE pattern recognition or pattern recognition and image processing conferences, IEEE Trans. on Pattern Analysis and Machine Intelligence, IEEE Trans. on Computers, and the Pattern Recognition journal. There have also been some conferences on medical pattern recognition. Can anyone suggest a bibliography, special issue, or book on these subjects? Have there been any AI (as opposed to PR) approaches to waveform diagnosis? -- KIL] ------------------------------ Date: 19-Feb-84 02:14 PST From: Kirk Kelley Subject: G. Spencer-Brown and the Laws of Form I know of someone who talked with G. on the telephone about six years ago somewhere in Northern California. My friend developed a quantum logic for expressing paradoxes, and some forms of schyzophrenia, among other things. Puts fuzzy set theory to shame. Anyway, he wanted to get together with G. to discuss his own work and what he perceived in the Laws of Form as very fundamental problems in generality due to over-simplicity. G. refused to meet without being paid fifty or so dollars per hour. Others say that the LoF's misleading notation masks the absence of any significant proofs. They observe that the notation uses whitespace as an implicit operator, something that becomes obvious in an attempt to parse it when represented as character strings in a computer. I became interested in the Laws of Form when it first came out as it promised to be quite an elegant solution to the most obscure proofs of Whitehead and Russell's Principia Mathematica. The LoF carried to perfection a very similar simplification I attempted while studying the same logical foundations of mathematics. One does not get too far into the proofs before getting the distinct feeling that there has GOT to be a better way. It would be interesting to see an attempt to express the essence of Go:del's sentence in the LoF notation. -- kirk ------------------------------ Date: Fri 17 Feb 84 10:57:18-PST From: Ken Laws Subject: Commercial NL Review in High Technology The February issue of High Technology has a short article on natural language interfaces (to databases, mainly). The article and business outlook section mention four NL systems currently on the market, led by AIC's Intellect ($70,000, IBM mainframes), Frey Associate's Themis ($24,000, DEC VAX-11), and Cognitive System's interface. (The fourth is not named, but some OEMs and licensees of the first two are given.) The article says that four more systems are expected out this year, and discusses Symantec's system ($400-$600, IBM PC with 256 Kbytes and hard disk) and Cal Tech's ASK (HP9836 micro, licensed to HP and DEC). ------------------------------ Date: Tue, 14 Feb 84 11:21:09 EST From: Kris Hammond Subject: *AI-LUNCH* [Forwarded from a Yale bboard by Shrager@CMU-PSY-A.] THE ADVENTURES OF JOE LISP, T MAN Brought to you by: *AI-LUNCH*, its hot, its cold, its more than a lunch... This week's episode: The Case of the Bogus Expert Part I It was late on a Tuesday and I was dead in my seat from nearly an hour of grueling mail reading and idle chit-chat with random passers by. The only light in my office was the soft glow from my CRT, the only sound was the pain wracked rattle of an over-heated disk. It was raining out, but the steady staccato rhythm that beat its way into the skulls of others was held back by the cold concrete slabs of my windowless walls. I like not having windows, but that's another story. I didn't hear her come in, but when the scent of her perfume hit me, my head swung faster than a Winchester. She was wearing My-Sin, a perfume with the smell of an expert, but that wasn't what impressed me. What hit me was her contours. She had a body with all the right variables. She wore a dress with a single closure that barely hid the dynamic scoping of what was underneath. Sure I saw her as an object, but I guess I'm just object oriented. It's the kind of operator I am. After she sat down and began to tell her story I realized that her sophisticated look was just cover. She was a green kid, still wet behind the ears. In fact she was wet all over. As I said, it was raining outside. It's an easy inference. It seems the kid's step-father had disappeared. He had been a medical specialist, diagnosis and prescription, but one day he started making wild claims about knowledge and planning and then he vanished. I had heard of this kind before. Some were specialists. Some in medicine, some in geology, but all were the same kind of guy. I looked the girl in the eye and asked the one question she didn't want to hear, "He's rule-based, isn't he?". She turned her head away and that was all the answer I needed. His kind were cold, unfeeling, unchanging, but she still loved him and wanted him back again. Once I got a full picture of the guy I was sure that I knew where to find him, California. It was the haven for his way of thinking and acting. I was sure that he had been swept up by the EXPERTS. They were a cult that had grown up in the past few years, promising fast and easy enlightenment. What they didn't tell you was that the price was your ability to understand itself. He was there, as sure as I was a T Man. I knew of at least one operative in California who could be trusted, and I knew that I had to talk to him before I could do any further planning. I reached for the phone and gave him a call. The conversation was short and sweet. He had resource conflicts and couldn't give me a hand right now. I assumed that it had to be more complex than that and almost said that resource conflicts aren't that easy to identify, but I had no time to waste on in fighting while the real enemy was still at large. Before he hung up, he suggested that I pick up a radar detector if I was planning on driving out and asked if I could grab a half-gallon of milk for him on the way. I agreed to the favor, thanked him for his advice and wished him luck on his tan... That's all for now kids. Tune in next week for the part two of: The Case of the Bogus Expert Starring JOE LISP, T MAN And remember kids, Wednesdays are *AI-LUNCH* days and 11:45 is the *AI-LUNCH* time. And kids, if you send in 3 box tops from *AI-LUNCH* you can get a JOE LISP magic decoder ring. This is the same ring that saved JOE LISP only two episodes ago and is capable of parsing from surface to deep structure in less than 15 transformations. Its part plastic, part metal and all bogus, so order now. ------------------------------ Date: 17 February 1984 11:55 EST From: Kenneth Byrd Story Subject: Computational Discovery of Mathamatical Laws [Forwarded from the MIT-MC bboard by Laws@SRI-AI.] TITLE: "The Computational Discovery of Mathematical Laws: Experiments in Bin Packing" SPEAKER: Dr. Jon Bentley, Bell Laboratories, Murray Hill DATE: Wednesday, February 22, 1984 TIME: 3:30pm Refreshments 4:15pm Lecture PLACE: Bldg. 2-338 Bin packing is a typical NP-complete problem that arises in many applications. This talk describes experiments on two simple bin packing heuristics (First Fit and First Fit Decreasing) which show that they perform extremely well on randomly generated data. On some natural classes of inputs, for instance, the First Fit Decreasing heuristic finds an optimal solution more often than not. The data leads to several startling conjectures; some have been proved, while others remain open problems. Although the details concern the particular problem of bin packing, the theme of this talk is more general: how should computer scientists use simulation programs to discover mathematical laws? (This work was performed jointly with D.S. Johnson, F.T. Leighton and C.A. McGeoch. Tom Leighton will give a talk on March 12 describing proofs of some of the conjectures spawned by this work.) HOST: Professor Tom Leighton THIS SEMINAR IS JOINTLY SPONSORED BY THE COMBINATORICS SEMINAR & THE THEORY OF COMPUTATION SEMINAR ------------------------------ Date: 17 Feb 1984 15:14 EST (Fri) From: "Daniel S. Weld" Subject: Revolving Seminar [Forwarded from the MIT-OZ bboard by SASW@MIT-MC.] [I am uncertain as to the interest of AIList readers in robotics, VLSI and CAD/CAM design, graphics, and other CS-related topics. My current policy is to pass along material relating to planning and high-level reasoning. Readers with strong opinions for or against such topics should write to AIList-Request@SRI-AI. -- KIL] AUTOMATIC SYNTHESIS OF FINE-MOTION STRATEGIES FOR ROBOTS Tomas Lozano Perez The use of force-based compliant motions enables robots to carry out tasks in the presence of significant sensing and control errors. It is quite difficult, however, to discover a strategy of such motions to achieve a task. Furthermore, the choice of motions is quite sensitive to details of geometry and to error characteristics. As a result, each new task presents a brand new and difficult problem. These factors motivate the need for automatic synthesis for compliant motions. In this talk I will describe a formal approach to the synthesis of compliant motion strategies from geometric description of assembly operations. (This is joint work [no pun intended -- KIL] with Matt Mason of CMU and Russ Taylor of IBM) ------------------------------ Date: Fri 17 Feb 84 09:02:29-PST From: Sharon Bergman Subject: Ph.D. Oral [Forwarded from the Stanford bboard by Laws@SRI-AI.] PH.D. ORAL USE OF ARTIFICIAL INTELLIGENCE AND SIMPLE MATHEMATICS TO ANALYZE A PHYSIOLOGICAL MODEL JOHN C. KUNZ, STANFORD/INTELLIGENETICS 23 FEBRUARY 1984 MARGARET JACKS HALL, RM. 146, 2:30-3:30 PM The objective of this research is to demonstrate a methodology for design and use of a physiological model in a computer program that suggests medical decisions. This methodology uses a physiological model based on first principles and facts of physiology and anatomy. The model includes inference rules for analysis of causal relations between physiological events. The model is used to analyze physiological behavior, identify the effects of abnormalities, identify appropriate therapies, and predict the results of therapy. This methodology integrates heuristic knowledge traditionally used in artificial intelligence programs with mathematical knowledge traditionally used in mathematical modeling programs. A vocabulary for representing a physiological model is proposed. ------------------------------ Date: Tue 21 Feb 84 10:47:50-PST From: Juanita Mullen Subject: ANNOUNCEMENT [Forwarded from the Stanford SIGLUNCH distribution by Laws@SRI-AI.] Thursday, February 23, 1984 Professor Kenneth Kahn Upssala University will give a talk: "Logic Programming and Partial Evaluation as Steps Toward Efficient Generic Programming" at: Bldg. 200, (History Building), Room 107, 12 NOON PROLOG and extensions to it embedded in LM PROLOG will be presented as a means of describing programs that can be used in many ways. Partial evaluation is a process that automatically produces efficient, specialized versions of programs. Two partial evaluators, one for LISP and one for PROLOG, will be presented as a means for winning back efficiency that was sacrificed for generality. Partial evaluation will also be presented as a means of generating compilers. ------------------------------ Date: 21 Feb 84 15:27:53 EST From: DSMITH@RUTGERS.ARPA Subject: Rutger's University Computer Science Colloquium [Forwarded from the Rutgers bboard by Laws@SRI-AI.] COLLOQUIUM Department of Computer Science SPEAKER: John Cannon Dept. of Math University of Sydney Syndey, AUSTRIA TITLE: "DESIGN AND IMPLEMENTATION OF A PROGRAMMING LANGUAGE/EXPERT SYSTEMS FOR MODERN ALGEBRA" Abstract Over the past 25 years a substantial body of algorithms has been devised for computing structural information about graphs. In order to make these techniques more generally available, I have undertaken the development of a system for group theory and related areas of algebra. The system consists of a high-level language (having a Pascal-like syntax) supported by an extensive library. In that the system attempts to plan, at a high level, the most economical solution to a problem, it has some of the attributes of an expert system. This talk will concentrate on (a) the problems of designing appropriate syntax for algebra and, (b) the implementation of a language professor which attempts to construct a model of the mathematical microworld with which it is dealing. DATE: Friday, February 24, 1984 TIME: 2:50 p.m. PLACE: Hill Center - Room 705 * Coffee served at 2:30 p.m. * ------------------------------ End of AIList Digest ******************** 29-Feb-84 14:04:15-PST,16017;000000000001 Mail-From: LAWS created at 29-Feb-84 13:58:16 Date: Wed 29 Feb 1984 13:46-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #22 To: AIList@SRI-AI AIList Digest Wednesday, 29 Feb 1984 Volume 2 : Issue 22 Today's Topics: Robotics - Personal Robotics Request, Books - Request for Laws of Form Review, Expert Systems - EURISKO Information Request, Automated Documentation Tools - Request, Mathematics - Fermat's Last Theorem & Map Coloring, Waveform Analysis - EEG/EKG Interpretation, Brain Theory - Parallelism, CS Culture - Computing Worlds ---------------------------------------------------------------------- Date: Thu 16 Feb 84 17:59:03-PST From: PIERRE@SRI-AI.ARPA Subject: Information about personal robots? Do you know anything about domestic robots? personal robots? I'm interested by the names and adresses of companies, societies, clubs, universities involved in that field. Does there exist any review about this? any articles? Do you work or have you heard of any projects in this field? Thank you to answer at Pierre@SRI-AI.ARPA Pierre ------------------------------ Date: 23 Feb 84 13:58:28 PST (Thu) From: Carl Kaun Subject: Laws of Form I hope that Charlie Crummer will share some of the substance he finds in "Laws of Form" with us (ref AIList Digest V2 #20). I myself am more in the group that does not understand what LoF has to say that is new, and indeed doubt that it does say anything unique. ------------------------------ Date: Fri, 24 Feb 84 15:32 MST From: RNeal@HIS-PHOENIX-MULTICS.ARPA Subject: EURISKO I have just begun reading the AI digests (our copy starts Nov 3 1983) and I am very interested in the one or two transactions dealing with EURISKO. Could someone explain what EURISKO does, and maybe give some background of its development? On a totally different note, has anyone done any AI work on lower order intelligence (ie. that using instinct) such as insects, reptiles, etc.? Seems they would be easier to model, and I just wondered if anyone had attempted to make a program which learns they way they do and the things they do . I don't know if this belongs in AI or some simulation meeting (is there one?). >RUSTY< ------------------------------ Date: 27 Feb 1984 07:26-PST From: SAC.LONG@USC-ISIE Subject: Automated Documentation Tools Is anyone aware of software packages available that assist in the creation of documentation of software, such as user manuals and maintenance manuals? I am not looking for simple editors which are used to create text files, but something a little more sophisticated which would reduce the amount of time one must invest in creating manuals manually (with the aid of a simple editor). If anyone has information about such, please send me a message at: SAC.LONG@USC-ISIE or Steve Long 1018-1 Ave H Plattsmouth NE 68048 or (402)294-4460 or reply through AIList. Thank you. -- Steve ------------------------------ Date: 16 Feb 84 5:36:12-PST (Thu) From: decvax!genrad!wjh12!foxvax1!minas @ Ucb-Vax Subject: Re: Fermat's Last Theorem & Undecidable Propositions Article-I.D.: foxvax1.317 Could someone please help out an ignorant soul by posting a brief (if that is, indeed, possible!) explanation of what Fermat's last theorem states as well as what the four-color theorem is all about. I'm not looking for an explanation of the proofs, but, simply, a statement of the propositions. Thanks! -phil minasian decvax!genrad!wjh12!foxvax1!minas ------------------------------ Date: 15 Feb 84 20:15:33-PST (Wed) From: ihnp4!mit-eddie!rh @ Ucb-Vax Subject: Re: Four color... Article-I.D.: mit-eddi.1290 I had thought that 4 color planar had been proved, but that the "conjectures" of 5 colors for a sphere and 7 for a torus were still waiting. (Those numbers are right, aren't they?) Randwulf (Randy Haskins); Path= genrad!mit-eddie!rh ------------------------------ Date: 17 Feb 84 21:33:46-PST (Fri) From: decvax!dartvax!dalcs!holmes @ Ucb-Vax Subject: Re: Four color... Article-I.D.: dalcs.610 The four colour problem is the same for a sphere as it is for the infinite plane. The problem for a torus was solved many years ago. The torus needs exactly 7 colours to paint it. Ray ------------------------------ Date: 26 Feb 1984 21:38:16-PST From: utcsrgv!utai!tsotsos@uw-beaver Subject: AI approach to ECG analysis One of my PhD students, Taro Shibahara, has been working on an expert system for arrhythmia analysis. The thesis should be finished by early summer. A preliminary paper discussing some design issues appeared in IJCAI-83. System name is CAA - Causal Arrhythmia Analyzer. Important contributions: Two distinct KB's, one of signal domain the other of the electrophysiological domain, communication via a "projection" mechanism, causal relations to assist in prediction, use of meta-knowledge within a frame-based representation for statistical knowledge. The overall structure is based on the ALVEN expert system for left ventricular performance assessment, developed here as well. John Tsotsos Dept. of Computer Science University of Toronto [Ray Perrault also suggested this lead. -- KIL] ------------------------------ Date: 24 Feb 84 10:07:36-PST (Fri) From: decvax!mcnc!ecsvax!jwb @ Ucb-Vax Subject: computer ECG Article-I.D.: ecsvax.2043 At least three companies are currently marketing computer ECG analysis systems. They are Marquette Electronics, IBM, Hewlett-Packard. We use the Marquette system which works quite well. Marquette and IBM use variants of the same program (the "Bonner" program below, original development funded by IBM.) Apparently because of fierce competition, much current information, particularly with regard to algorithms, is proprietary. Worst in this regard (a purely personal opinion) is HP who seems to think nobody but HP needs to know how they do things and physicians are too dumb to understand anyway. Another way hospitals get computer analysis of ECG's is through "Telenet" who offers telephone connection to a time sharing system (I think located in the Chicago area). Signals are digitized and sent via a modem through standard phone lines. ECG's are analyzed and printed information is sent back. Turn-around time is a few minutes. They offer an advantage to small hospitals by offering verification of the analysis by a Cardiologist (for an extra fee). I understand this service has had some financial problems (rumors). Following is a bibliography gathered for a lecture to medical students about computer analysis of ECG's. Because of this it is mainly from more or less clinical literature and is oriented toward methods of validation (This is tough, because reading of ECG's by cardiologists, like many clinical decisions, is partly a subjective process. The major impact of these systems so far has been to force the medical community to develop objective criteria for their analysis.) BIBLIOGRAPHY Computer Analysis of the Electrocardiogram August 29, 1983 BOOK Pordy L (1977) Computer electrocardiography: present status and criteria. Mt. Kisco, New York, Futura PAPERS Bonner RE, Crevasse L, Ferrer MI, Greenfield JC Jr (1972) A new computer program for analysis of scalar electrocardiograms. Computers and Biomedical Research 5:629-653 Garcia R, Breneman GM, Goldstein S (1981) Electrogram computer analysis. Practical value of the IBM Bonner-2 (V2MO) program. J. Electrocardiology 14:283-288 Rautaharju PM, Ariet M, Pryor TA, et al. (1978) Task Force III: Computers in diagnostic electrocardiography. Proceedings of the Tenth Bethesda Conference, Optimal Electrocardiography. Am. J. Cardiol. 41:158-170 Bailey JJ et al (1974) A method for evaluating computer programs for electrocardiographic interpretation I. Application to the experimental IBM program of 1971. Circulation 50:73-79 II. Application to version D of the PHS program and the Mayo Clinic program of 1968. Circulation 50:80-87 III. Reproducibility testing and the sources of program errors. Circulation 50:88-93 Endou K, Miyahara H, Sato (1980) Clinical usefulness of computer diagnosis in automated electrocardiography. Cardiology 66:174-189 Bertrand CA et al (1980) Computer interpretation of electrocardiogram using portable bedside unit. New York State Journal of Medicine. August 1980(?volume):1385-1389 Jack Buchanan Cardiology and Biomedical Engineering University of North Carolina at Chapel Hill (919) 966-5201 decvax!mcnc!ecsvax!jwb ------------------------------ Date: Friday, 24-Feb-84 18:35:44-GMT From: JOLY G C QMA (on ERCC DEC-10) Subject: re: Parallel processing in the brain. To compare the product of millions of years of evolution (ie the human brain) with the recent invention of parallel processors seems to me to be like trying to effect an analysis of the relative properties of chalk and cheese. Gordon Joly. ------------------------------ Date: Wed, 29 Feb 84 13:17:04 PST From: Dr. Jacques Vidal Subject: Brains: Serial or Parallel? Is the brain parallel? Or is the issue a red herring? Computing and thinking are physical processes and as all physical processes unfold in time are ultimately SEQUENTIAL even "continu- ous" ones although the latter are self-timed (free-running, asyn- chronous) rather than clocked. PARALLEL means that there are multiple tracks with similar func- tions like availability of multiple processors or multiple lanes on a superhighway. It is a structural characteristic. CONCURRENT means simultaneous. It is a temporal characteristic. REDUNDANT means that there is structure beyond that which is minimally needed for function, perhaps to insure integrity of function under perturbations. In this context, PARALLELISM, i.e. the deployment of multiple processors is the currency with which a system designer may pur- chase these two commodities: CONCURRENCY and REDUNDANCY (a neces- sary but not sufficient condition). Turing machines have zero concurrency. Almost everything else that computes exhibit some. Conventional processor architectures and memories are typically concurrent at the word level. Microprogram are sequences of concurrent gate events. There exist systems that are completely concurrent and free- running. Analog computers and combinational logic circuits have these properties. There, computation progresses by chunk between initial and final states. A new chunk starts when the system is set to a new initial state. Non-von architectures have moved away from single track computing and from the linear organization of memory cells. With cellular machines another property appears: ADJACENCY. Neighboring proces- sors use adjacency as a form of addressing. These concepts are applicable to natural automata: Brains cer- tainly employ myriads of processors and thus exhibit massive parallelism. From the numerous processes that are simultaneously active (autonomous as well as deliberate ones) it is clear that brains utilize unprecedented concurrency. These proces- sors are free-running. Control and data flows are achieved through three-dimensional networks. Adjacency is a key feature in most of the brain processes that have been identified. Long dis- tance communication is provided for by millions of parallel path- ways, carrying highly redundant messages. Now introspection indicates that conscious thinking is limited to one stream of thought at any given time. That is a limitation of the mechanisms supporting consciousness amd some will claim that it can be overcome. Yet even a single stream of thinking is cer- tainly supported by many concurrent processes, obvious when thoughts are spoken, accompanied by gestures etc... Comments? ------------------------------ Date: 18 Feb 1984 2051-PST From: Rob-Kling Subject: Computing Worlds [Forwarded from Human-Nets Digest by Laws@SRI-AI.] Sherry Turkle is coming out with a book that may deal in part with the cultures of computing worlds. It also examines questions about how children come to see computer applications as alive, animate, etc. It was to be called, "The Intimate Machine." The title was appropriated by Neil Frude who published a rather superficial book with an outline very similar to that Turkle proposed to some publishers. Frude's book is published by New American Library. Sherry Turkle's book promises to be much deeper and careful. It is to be published by Simon and Schuster under a different title. Turkle published an interesting article called, "Computer as Rorschach" in Society 17(2)(Jan/Feb 1980). This article examines the variety of meanings that people attribute to computers and their applications. I agree with Greg that computing activities are embedded within rich social worlds. These vary. There are hacker worlds which differ considerably from the worlds of business systems analysts who develop financial applications in COBOL on IBM 4341's. AI worlds differ from the personal computing worlds, and etc. To date, no one appears to have developed a good anthropological account of the organizing themes, ceremonies, beliefs, meeting grounds, etc. of these various computing worlds. I am beginning such a project at UC-Irvine. Sherry Turkle's book will be the best contribution (that I know of) in the near future. One of my colleagues at UC-Irvine, Kathleen Gregory, has just completed a PhD thesis in which she has studied the work cultures within a major computer firm. She plans to transform her thesis into a book. Her research is sensitive to the kinds of langauage categories Greg mentioned. (She will joining the Department of Information and Computer Science at UC-Irvine in the Spring.) Also, Les Gasser and Walt Scacchi wrote a paper on personal computing worlds when they were PhD students at UCI. It is available for $4 from: Public Policy Research Organization University of California, Irvine Irvine,Ca. 92717 (They are now in Computer Science at USC and may provide copies upon request.) Several years ago I published two articles which examine some of the larger structural arrangments in computing worlds: "The Social Dynamics of Technical Innovation in the Computing World" ^&Symbolic Interaction\&, 1(1)(Fall 1977):132-146. "Patterns of Segmentation and Intersection in the Computing World" ^&Symbolic Interaction\& 1(2)(Spring 1978): 24-43. One section of a more recent article, "Value Conflicts in the Deployment of Computing Applications" ^&Telecommunications Policy\& (March 1983):12-34. examines the way in which certain computer-based technologies such as automated offices, artificial intelligence, CAI, etc. are the foci of social movements. None of my papers examine the kinds of special languages which Greg mentions. Sherry Turkle's book may. Kathleen Gregory's thesis does, in the special setting of one major computing vendor's software culture. I'll send copies of my articles on request if I recieve mailing addresses. Rob Kling University of California, Irvine ------------------------------ End of AIList Digest ******************** 29-Feb-84 14:34:31-PST,13158;000000000001 Mail-From: LAWS created at 29-Feb-84 14:30:58 Date: Wed 29 Feb 1984 14:11-PST From: AIList Moderator Kenneth Laws Reply-to: AIList@SRI-AI US-Mail: SRI Int., 333 Ravenswood Ave., Menlo Park, CA 94025 Phone: (415) 859-6467 Subject: AIList Digest V2 #23 To: AIList@SRI-AI AIList Digest Thursday, 1 Mar 1984 Volume 2 : Issue 23 Today's Topics: Seminars - VLSI Knowledge Representation & Machine Learning & Computer as Musical Scratchpad & Programming Language for Group Theory & Algorithm Animation Conference - Very Large Databases Call for Papers ---------------------------------------------------------------------- Date: Wed 22 Feb 84 16:36:20-PST From: Joseph A. Goguen Subject: Hierarchical Software Processor [Forwarded by Laws@SRI-AI.] An overview of HISP by K. Futatsugi Special Lecture at SRI, 27 February 1984 HISP (hierarchical software processor) is an experimental language/system, which has been developed at ETL (Electrotechnical Laboratory, Japan) by the author's group, for hierarchical software development based on algebraic specification techniques. In HISP, software development is simply modeled as the incremental construction of a set of hierarchically structured clusters of operators (modules). Each module is the constructed as a result of applying one of the specific module building operations to the already existing modules. This basic feature makes it possible to write inherently hierarchical and modularized software. This talk will inroduce HISP informally by the use of simple examples. The present status of HISP implementation and future possibilities will also be sketched. ------------------------------ Date: Thu 23 Feb 84 00:26:45-MST From: Subra Subject: Very High Level Silicon Compilation [Forwarded by Laws@SRI-AI. This talk was presented at the SRI Computer Science Laboratory.] VERY HIGH LEVEL SILICON COMPILATION: THEORY AND PRACTICE P.A.Subrahmanyam Department of Computer Science University of Utah The possibility of implementing reasonably complex special purpose systems directly in silicon using VLSI technologies has served to underline the need for design methodologies that support the development of systems that have both hardware and software components. It is important in the long run for automated design aids that support such methodologies to be based on a uniform set of principles -- ideally, on a unifying theoretical basis. In this context, I have been investigating a general framework to support the analytic and synthetic tasks of integrated system design. Two of the salient features of this basis are: - The formalism allows various levels of abstraction involved in the software/hardware design process to be modelled. For example, functional (behavioral), architectural (system and chip level), symbolic layout, and electrical (switch-level)-- are explicitly modelled as being typical of the levels of abstraction that human "expert designers" work with. - The formalism allows for explicit reasoning about behavioral, spatial, temporal and performance criteria. The talk will motivate the general problem, outline the conceptual and theoretical basis, and discuss some of our preliminary empirical explorations in building integrated software-hardware systems using these principles. ------------------------------ Date: 22 Feb 84 12:19:09 EST From: Giovanni Subject: Machine Learning Seminar [Forwarded from the Rutgers bboard by Laws@SRI-AI.] *** MACHINE LEARNING SEMINAR AND PIZZA LUNCHEON *** Empirical Exploration of Problem Reformulation and Strategy Acquisition Authors: N.S. Sridharan and J.L. Bresina Location: Room 254, Hill Center, Busch Campus, Rutgers Date: Wednesday, February 29, 1984 Time: Noon - 1:30 pm Speaker: John L. Bresina The problem solving ability of an AI program is critically dependent on the nature of the symbolic formulation of the problem given to the program. Improvement in performance of the problem solving program can be made by improving the strategy of controlling and directing search but more importantly by shifting the problem formulation to a more appropriate form. The choice of the initial formulation is critical, since certain formulations are more amenable to incremental reformulations than others. With this in mind, an Extensible Problem Reduction method is developed that allows incremental strategy construction. The class of problems of interest to us requires dealing with interacting subgoals. A variety of reduction operator types are introduced corresponding to different ways of handling the interaction among subgoals. These reduction operators define a generalized And/Or space including constraints on nodes with a correspondingly generalized control structure for dealing with constraints and for combining solutions to subgoals. We consider a modestly complex class of board puzzle problems and demonstrate, by example, how reformulation of the problem can be carried out by the construction and modification of reduction operators. ------------------------------ Date: 26 Feb 84 15:16:08 EST From: BERMAN@RU-BLUE.ARPA Subject: Seminar: The Computer as Musical Scratchpad [Forwarded from the Rutgers bboard by Laws@SRI-AI.] SEMINAR: THE COMPUTER AS MUSICAL SCRATCHPAD Speaker: David Rothenburg, Inductive Inference, Inc. Date: Monday, March 5, 1984 Place: CUNY Graduate Center, 33 West 42nd St., NYC Room: 732 Time: 6:30 -- 7:30 p.m. The composer can use a description language wherein only those properties and relations (of and between protions of the musical pattern) which he judges significant need be specified. Parameters of these unspecified properties and relations are assigned at random. It is intended that this description of the music be refined in response to iterated auditions. ------------------------------ Date: Sun 26 Feb 84 17:06:23-CST From: Bob Boyer Subject: A Programming Language for Group Theory (Dept. of Math) [Forwarded from the UTexas-20 bboard by Laws@SRI-AI.] DEPARTMENT OF MATHEMATICS COLLOQUIUM A Programming Language for Group Theory John Cannon University of Sydney and Rutgers University Monday, February 27, 4pm The past 25 years has seen the emergence of a small but vigorous branch of group theory which is concerned with the discovery and implementation of algorithms for computing structural information about both finite and infinite groups. These techniques have now reached the stage where they are finding increasing use both in group theory research and in its applications. In order to make these techniques more generally available, I have undertaken the development of what in effect is an expert system for group theory. Major components of the system include a high-level user language (having a Pascal-like syntax) and an extensive library of group theory algorithms. The system breaks new ground in that it permits efficient computation with a range of different types of algebraic structures, sets, sequences, and mappings. Although the system has only recently been released, already it has been applied to problems in topology, algebraic number theory, geometry, graphs theory, mathematical crystalography, solid state physics, numerical analysis and computational complexity as well as to problems in group theory itself. ------------------------------ Date: 27 Feb 1984 2025-PST (Monday) From: Forest Baskett Subject: EE380 - Wednesday, Feb. 29 - Sedgewick on Algorithm Animation [Forwarded from the Stanford bboard by Laws@SRI-AI.] EE380 - Computer Systems Seminar Wednesday, February 29, 4:15 pm Terman Auditorium Algorithm Animation Robert Sedgewick Brown University The central thesis of this talk is that it is possible to expose fundamental characteristics of computer programs through the use of dynamic (real-time) graphic displays, and that such algorithm animation has the potential to be useful in several contexts. Recent research in support of this thesis will be described, including the development of a conceptual framework for the process of animation, the implementation of a software environment on high-performance graphics-based workstations supporting this activity, and the use of the system as a principal medium of communication in teaching and research. In particular, we have animated scores of numerical, sorting, searching, string processing, geometric, and graph algorithms. Several examples will be described in detail. [Editorial remark: This is great stuff. - Forest] ------------------------------ Date: 23 Feb 84 16:32:24 PST (Thu) From: Gerry Wilson Subject: Conference Call for Papers CALL FOR PAPERS ================ 10'th International Conference on Very Large Data Bases The tenth VLDB conference is dedicated to the identification and encouragement of research, development, and application of advanced technologies for management of large data bases. This conference series provides an international forum for the promotion of an understanding of current research; it facilitates the exchange of experiences gained in the design, construction and use of data bases; it encourages the discussion of ideas and future research directions. In this anniversary year, a special focus is the reflection upon lessons learned over the past ten years and the implications for future research and development. Such lessons provide the foundation for new work in the management of large data bases, as well as the merging of data bases, artificial intelligence, graphics, and software engineering technologies. TOPICS: Data Analysis and Design Intelligent Interfaces Multiple Data Types User Models Semantic Models Natural Language Dictionaries Knowledge Bases Graphics Performance and Control Data Representation Workstation Data Bases Optimization Personal Data Mangement Measurement Development Environments Recovery Expert System Applications Message Passing Designs Security Protection Real Time Systems Semantic Integrity Process Control Concurrency Manufacturing Engineering Design Huge Data Bases Data Banks Implementation Historical Logs Languages Operating Systems Multi-Technology Systems Applications Distributed Data Bases Office Automation Distribution Management Financial Management Heterogeneous and Homogeneous Crime Control Local Area Networks CAD/CAM Hardware Data Base Machines Associative Memory Intelligent Peripherals LOCATION: Singapore DATES: August 29-31, 1984 TRAVEL SUPPORT: Funds will be available for partial support of most participants. HOW TO SUBMIT: Original full length (up to 5000 words) and short (up to 1000 words) papers are sought on topics such as those above. Four copies of the submission should be sent to the US Program Chairman: Dr. Umeshwar Dayal Computer Corporation of America 4 Cambridge Center Cambridge, Mass. 02142 [Dayal@CCA-UNIX] IMPORTANT DATES: Papers Due: March 15, 1984 Notification: May 15, 1984 Camera Ready Copy: June 20, 1984 For additional information contact the US Conference Chairman: Gerald A. Wilson Advanced Information & Decision Systems 201 San Antonio Circle Suite 286 Mountain View, California 94040 [Wilson@AIDS] ------------------------------ End of AIList Digest ********************