From vtcs1::bitnet% Thu Apr 17 15:42:49 1986 Date: Thu, 17 Apr 86 15:42:45 est From: vtcs1::bitnet% (AIList-REQUEST@SRI-AI) To: fox Subject: AIList Digest V4 #73 Status: RO Received: From WISCVM(SMTPUSR1) by VTCS1 with RSCS id 3889 for FOX@VTCS1; Thu, 17-APR-1986 15:44 EST Received: from SRI-AI.ARPA by wiscvm.wisc.edu on 04/17/86 at 09:42:11 CST Mail-From: LAWS created at 8-Apr-86 22:17:45 Date: Tue 8 Apr 1986 22:14-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 V4 #73 To: AIList@SRI-AI ReSent-date: Wed 16 Apr 86 21:55:15-PST ReSent-From: Ken Laws ReSent-To: fox%vtcs1.bitnet@WISCVM.WISC.EDU AIList Digest Wednesday, 9 Apr 1986 Volume 4 : Issue 73 Today's Topics: Psychology - Survival Instinct & Emotions ---------------------------------------------------------------------- Date: 2 Apr 86 10:23:18 GMT From: ulysses!mhuxr!mhuxt!houxm!whuxl!whuxlm!akgua!gatech!seismo!ll-xn !mit-amt!mit-eddie!psi@ucbvax.berkeley.edu Subject: Re: Computer Dialogue Hi: Before the recent tragedy, there had been a number of instances where the space shuttle computers aborted the mission in the final seconds before launch. My explanation for this was that the on-board computers were displaying a form of 'programmed survival instinct.' In short: they were programmed to survive, and if the launch had continued, they might not have. Almost everyone I told explained this to back then was incredulous. "You don't actually _believe_ that the computer wanted to survive, do you?" was a typical comment. I feel this brings out an important point, though, which deals with simulation, feelings, and our understanding of The Real Thing. On a computer, simulating an event and the actual event may be indistinguishable. (This does not mean, as one of my friends believed, that in a computer simulation of a hurricane, the simulated victims of the storm would be rained upon by square-root symbols.;-)) For example, if a computer can run programs in the language Lisp and we then write a simulator for the language CLU in Lisp, then the computer can actually run programs in CLU. Now, what does this mean for feelings? Well, I won't go that far, but I would assert that a 'survival instinct' is a much simpler thing that can be simulated on a computer. The space shuttle computers could be thought of as programmed to survive, in just the same way that evolution has programmed animals to survive. No consciousness is necessary(yet), just a goal and a means to that goal. It should be noted that the means of continuing survival available to the space shuttle computers are very minimal right now, but even animals must draw upon a limited set of defenses in order to survive. The successes in AI so far have been in very restricted areas, to say the least. Certain well-understood human abilities have been simulated on computers. Where the ability is less understood, like that of a chess master, the simulation breaks down. Where something such as 'survival' may be understood, I challenge anyone to come up with a generalized theory of 'feelings.' A final point: whenever we understand something, it loses its magical properties for us. If, for example, we observe the complex behavior of some program, we may be amazed. When we look at the sources and see how it works, however, we will probably feel that there really is no magic there, and that we could have written the program ourselves. The same could be true of parts of the mind which we understand. The simpler facilities, like an instinct to survive may seem obvious, while others, such as the feeling of love may yet seem mystical. Maybe someday we will come to understand even that and be able to program it into computers. Ultimately Yours, Joseph J. Mankoski ***PSI*** {decvax!genrad, allegra, ihnp4}!mit-eddie!psi psi@mit-ai.ARPA In the fullness of time even parallel lines will meet. ------------------------------ Date: 3 Apr 86 20:43:57 GMT From: hplabs!hao!seismo!umcp-cs!venu@ucbvax.berkeley.edu (Venugopala R. Dasigi) Subject: Re: Computer Dialogue In article <1439@mit-eddie.MIT.EDU> psi@mit-eddie.UUCP writes: >thing that can be simulated on a computer. The space shuttle >computers could be thought of as programmed to survive, in just the >same way that evolution has programmed animals to survive. No >consciousness is necessary(yet), just a goal and a means to that goal. >It should be noted that the means of continuing survival available to >the space shuttle computers are very minimal right now, but even >animals must draw upon a limited set of defenses in order to survive. To me it appears that the ability to dynamically redefine the goal in a context-sensitive manner is also an important characteristic of the "survival instinct". While animals seem to have this ability, programming this ability into computers (in the same sense as in the case of animals) is perhaps very difficult. --- Venu Venugopala Rao Dasigi UUCP : {seismo,allegra,brl-bmd}!umcp-cs!venu CSNet : venu@umcp-cs ARPA : venu@mimsy.umd.edu US Mail: Dept. of CS, Univ. of Maryland, College Park MD 20742. ------------------------------ Date: 7 Apr 86 03:15:06 GMT From: ulysses!mhuxr!mhuxt!houxm!whuxl!whuxlm!akgua!gatech!seismo!rochester !rocksanne!sunybcs!ellie!colonel@ucbvax.berkeley.edu Subject: Re: survival instinct It depends on what you mean by "wanted." Even rocks are programmed to survive--they're hard. (The soft ones become dirt: survival of the fittest!) "This rock, for instance, has an I.Q. of zero. Ouch!" "What's the matter, Professor?" "It bit me!" Col. G. L. Sicherman UU: ...{rocksvax|decvax}!sunybcs!colonel CS: colonel@buffalo-cs BI: csdsicher@sunyabva ------------------------------ Date: 5 Apr 86 13:51:18 GMT From: ulysses!mhuxr!mhuxt!houxm!hounx!kort@ucbvax.berkeley.edu (B.KORT) Subject: Re: Computer Dialogue Joseph Mankoski writes a thought provoking article on whether survival logic in NASA computers has any connection to human survival instincts wired into to our brains from birth. I have been pondering this question myself. It seems to me that I have some autonomic responses to threat situations which appeear to be wired-in instincts. I note that I don't rely on them often. Most times, I rely on learned behavior to handle situations which might have called for fight/flight/freeze if I were living as a hunter-gatherer on the Savannahs some 20,000 years ago. Joseph asks for a theory of feelings. As it happens, I just wrote a brief article on the subject, which may or may not be suitable for publication after editorial comment and revision. Just for the hell of it, let me append the article and solicit comments from netters interested in this topic. ==================== Article on Feelings ======================== A Simplified Model of the Effects of Perceived Aggression in the Work Environment Barry Kort Copyright 1986 Introduction The work environment offers a mix of personalities. In this paper, I would like to examine the effects of one dimension along which personalities are perceived to differ, and trace the consequential effects. I would like to focus attention on the dimension aggressive...assertive...politic...nonassertive...nonaggressive. The effects that I wish to investigate are not the behavioral responses, but the more fundamental internal body sensations or somatic reactions which lie behind the subsequent behavioral response. The goal of this investigation is to discover the biological roots of somatic reactions to stressors in the work environment, and develop a useful model of the underlying dynamics. I make no claims that the model constructed here is complete or comprehensive. To do so is beyond my ken. Rather, I have attempted to construct a first crude model, which despite it's simplicity, can be advantageously applied to ameliorate a few of the ills that we encounter in the work environment. A Model of Nature of Aggressive Behavior It has been said that civilization is a thin veneer. Underneath our legacy of some 5000 years of civilization lies our evolutionary past. Deep within the human brain one can find the vestiges of our animal nature-the old mammalian brain, the old reptilian brain. Of principal interest here are two groups of structures responsible for much of our "wired-in" instincts. The cerebellum is responsible for much of our risk-taking, self-gratifying drives, including the aggressive sex drives. It is the cerebellum that says, "Go for it! This could be exciting! Damn the torpedoes, full speed ahead." The limbic system, on the other hand, is responsible for self-protective behaviors. The limbic system perceives the threats to one's safety or well-being, and initiates protective or counter measures. The limbic system says, "Hold it! This could be dangerous! We'd better go slow and avoid those torpedoes." Rising above it all resides the neocortex or cerebrum. This is the "new brain" of homo sapiens which is the seat of learning and intelligence. It is the part that gains knowledge of cause and effect patterns, and overrules the myopic attitude of the cerebellum and limbic system. Occasionally, the cerebral cortex is faced with a novel situation, where past experience and learning fail to provide adequate instruction in how to proceed. In that case, the usual patterns of regulation are ineffective, and the behavioral response may revert back to the more primitive instincts. Whether or not the cerebral cortex carries the day, the messages of the cerebellum and limbic system ricochet through the nervous system, leaving their signature here and there. In the next section, we explore how these messages manifest themselves in somatic sensations, commonly known as feelings. Somatic Reactions to Stress When an individual is presented with an unusual situation, the lack of an immediately obvious method of dealing with it may lead to an accumulation of stress which manifests itself somatically. For instance, first-time jitters may show up as a knotting of the stomach (butterflies), signaling fear (of failure). A perceived threat may cause increased heart rate, sweating, or a tightening of the skin on the back of the neck. (This latter phenomenon is commonly known as "raising of one's hackles," which in birds, causes the feathers to stand up in display mode, warning off the threatening invader.) Teeth clenching, which comes from repressing the urge to express anger, leads to a common affliction among adult males-temporal mandibular joint (TMJ). Leg shaking and pacing indicate a subliminal urge to flee, while cold feet corresponds to frozen terror (playing 'possum). All of these are variations on the fight/flight/freeze instincts mediated by the limbic system. They often occur without our conscious awareness. Another reaction is migraine headaches which arise when one is vexed by the situation at hand, and is searching without success for a rational solution. A person's awareness of and sensitivity to such somatic feelings may affect his mode of expression. The somasthetic cortex is the portion of the brain where the body stresses are registered, and this sensation may be the primary indication that a stressor is present in the environment. A challenge for every individual is to accurately identify which environmental stimulus is linked to which somatic response. Somatic responses such as those outlined above are intimately connected with our expressed feelings, which usually are translated into some behavioral response along the axis from aggressive to assertive to politic to nonassertive to nonaggresive. The challenge is to find and effectuate the middle ground between too much communication and too little. The goal of the communication is to identify the cause and effect link between the environmental stressor and the somatic reaction, and from the somatic reaction to the behavioral response. The challenge is all the more difficult because the most effective mode and intensity of the communication depends on the maturity of the other party. Acknowledgements The original sources for the ideas assembled in this paper are too diffuse to pinpoint with completeness or precision. However, I would like to acknowledge the influence of so many of my colleagues who took the time to contribute their ideas and experiences on the subject matter. I especially would like to thank Dr. John Karlin, Dr. R. Isaac Evan, and Dr. Laura Rogers who helped me shape and test the models presented here. ========================================================================= Comments are invited. --Barry Kort ...ihnp4!houxm!hounx!kort ------------------------------ End of AIList Digest ******************** From vtcs1::in% Thu Apr 10 18:55:45 1986 Date: Thu, 10 Apr 86 18:55:40 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #74 Status: RO AIList Digest Thursday, 10 Apr 1986 Volume 4 : Issue 74 Today's Topics: Policy - Discussion Style & Professional Ethics & Press Releases, Programming Languages - LetS Lisp Loop Notation ---------------------------------------------------------------------- Date: Thu, 3 Apr 86 11:52:18 GMT From: gcj%qmc-ori.uucp@cs.ucl.ac.uk Subject: Less on IQ tests for Computers, more on Editorial Policy? Scott Preece asks in Vol 4 # 66 :- ``Do we really want this list to be a battleground for unsubstantiated personal opinions on the potential for machine intelligence?'' Agreed that this a moderated digest, it is interesting to note that the net.ai forum is currently carrying a discussion of the cognitive (and emotional) abilities of an arbitrarily large number of toasters. Here is an example:- > In article <2345@jhunix.UUCP> ins_akaa@jhunix.UUCP (Ken Arromdee) writes: > >You are actually quite correct. There's one problem here. Toasters can > >store perhaps two or three bytes of information. Consider how many > >toasters would be required to be as complex as a human brain. > > > >And as for the future toasters, toasters' primary function is to affect > >items of a definite physical size (toast). > >-- > >Kenneth Arromdee > > Gee, I always thought that toasters' primary function was to affect > items of a definite physical size (bread). > -- > > When you meet a master swordsman, > show him your sword. > When you meet a man who is not a poet, > do not show him your poem. > - Rinzai, ninth century zen master > > --Nathan Hess > uucp: {allegra, ihnp4}!psuvax1!gondor!hess > csnet: hess@penn-state.CSNET > Bitnet: HESS@PSUVAXG.BITNET I would also like to extract this from the List_of_Lists :- > Contributions may be anything from tutorials to rampant speculation. In > particular, the following are sought: > Abstracts Reviews > Lab Descriptions Research Overviews > Work Planned or in Progress Half-Baked Ideas > Conference Announcements Conference Reports > Bibliographies History of AI > Puzzles and Unsolved Problems Anecdotes, Jokes, and Poems > Queries and Requests Address Changes (Bindings) The poetry of Rinzai is illuminating, cf Vol 4 # 50,53, and very apt. Gordon Joly ARPA: gcj%qmc-ori@ucl-cs.arpa UUCP: ...!ukc!qmc-cs!qmc-ori!gcj [I am unable to follow the logic of this message, but find it easier (and faster!) to let it pass than to engage in editorial debate with Gordon. Contributors should note that it is they, not I, who control the quality of AIList. My thanks to you all; keep up the good work. -- KIL] ------------------------------ Date: Fri, 4 Apr 86 12:53:55 GMT From: gcj%qmc-ori.uucp@cs.ucl.ac.uk Subject: World Times, April 1, 2284. A special analysis of the entries in the AI Digests of the mid 1980's has shown that all the entries written by "The Joka" were the products of an automated intelligent system. This result is regarded by some as an interesting twist on the Turing test. Other News. Today the World's first trial by computer was held. The jury consisted of 12 independent intelligent systems and they sat at the World Court in the U.N. The jury returned it's first verdict after a few seconds, and the judge commented on the impartiality of the jurors, unclouded by any emotion or form of prejudice. On trial was the off-world outlaw, Roy Baty... Reporter : PiQuan. ------------------------------ Date: Thu, 3 Apr 86 9:38:45 CST From: Glenn Veach Subject: Professional ethics. Over the past several months I have been receiving the AIList, and I must take this time to express some concerns of mine. I have seen several "policy notices" and debates raging where the authors have lowered themselves to the level of the "ad homina"(sp?) attack. One should have more substantive comments if one wishes to express criticism, and not resort to personal attacks. I am in no way opposed to healthy debate, even if it should become heated. However, there seems to be some dislike, on the part of many, of pointed criticism. I wish to admonish those who take part in this medium of intellectual exchange to express a little more common courtesy and professional ethic, if indeed either of these still remain. Let's drop the name-calling. I personally welcome criticism of AI, even if it (the criticism) may be in left field. After all, many think we are in left field, while we may hold that they are in left field. So, exactly where is left field? Perhaps it is dependent on ones own position? Also, we should remember that this is a monitored digest. I personally trust the discretion of Ken, who I think does a good job, to weed out any inappropriate notices. Thus, I would love to see this list continue to announce various product and research development, whether it be presented by a party directly involved in the development or someone farther removed. As long as it is not out and out advertisement, I, as well as others (I think), am interested in such postings. Enough for now... Glenn O. Veach Artificial Intelligence Laboratory Department of Computer Science University of Kansas Lawrence, KS 66045-2192 (913) 864-4482 veach%ukans.csnet@csnet-relay.csnet ------------------------------ Date: Wed 9 Apr 86 10:50:34-PST From: Pat Hayes Subject: Re: AIList Digest V4 #70 Part of this AIlist reads perilously like an advertisement, even though it is protected by Les Earnest's mention. Do we have to have whole 'product descriptions' ( ie advertising brochures ) put out over the net? Isn't that ( just slightly ) illegal? Pat Hayes ------------------------------ Date: Sun 16 Mar 86 22:01:03-PST From: Ken Laws Subject: Policy - Press Releases A press release typically contains factual information; the cost of transmitting it is small. Is it not always in the government's interest for me to pass on the information to those who may need it rather than to censor it (to avoid annoying those who don't)? Early net organizers were no doubt [rightly] worried about corporate PR departments broadcasting unwanted press releases to everyone on the net. The situation has changed. A press release judged appropriate for a narrow-topic discussion list by its moderator is unlikely to offend many (other than self-appointed censors) or to seriously waste the time of the list members. It will not mislead readers so long as it is clearly marked as a commercial message. The inherent bias of such messages is mitigated by the opportunity for immediate rebuttal and for submission of equally biased messages supporting other views. Any resulting controversy sparks interest and keeps the list active. Outright flaming or numbing repetition can be prevented by the moderator. If the moderator fails to intervene, comments from disgruntled readers will fill his (or her) mailbox and eventually become a metadiscussion within the list itself. Readers who get tired of all this can drop out. My view is that policy on commercial content (hardware hype, job ads, prices, whatever) within a discussion list should be set by the moderator and the list members -- not by conventions required for unmoderated message streams. The Arpanet administrators and host administrators will always hold the trump, of course; they can refuse to support any list that violates >>their<< standards. -- Ken Laws ------------------------------ Date: Fri, 28 Mar 1986 15:56 EST From: Dick@MC.LCS.MIT.EDU Subject: LetS -- a new Lisp loop notation [Forwarded from the MIT bboard by Laws@SRI-AI.] This message advertises a Common Lisp macro package called LetS (rhymes with process) which it is hoped will become a standard iteration facility in Common Lisp. LetS makes it possible to write a wide class of algorithms which are typically written as loops in a functional style which is similar to expressions written with the Common Lisp sequence functions. LetS supports a number of features which make LetS expressions more expressive than sequence expressions. However, the key feature of LetS is that every LetS expression is automatically transformed into an efficient iterative loop. As a result, unlike sequence expressions, LetS expressions are just as efficient as the traditional loop expressions they replace. An experimental version of LetS currently exists on the MIT-AI machine in the file "DICK;LETS BIN". Although LetS is written in Common Lisp, it has not yet been tested on anything other than a Symbolics Lisp Machine. For various detailed reasons it is unlikely to run on any other machine. Everyone who wants to is invited to borrow this file and try LetS out. I am very interested to hear any and all comments on LetS. Extensive documentation of LetS is in the file "DICK;LETSD >" also on the MIT-AI machine. Even people who do not have a Lisp Machine or are not able to access the code are invited to read this documentation and make comments on it. I am interested in getting as wide a feedback as possible. If you cannot access the documentation file directly, send me your US mail address and I will mail you a copy. The documentation is much too long to reliably send via computer mail. After an initial testing and feedback period, a final version of LetS which runs under all Common Lisps will be created along with formal documentation. This should happen within a couple of months. A very brief summary of lets is included at the end of this message. Dick Waters The advantages (with respect to conciseness, readability, verifiability and maintainability) of programs written in a functional style are well known. A simple example of the clarity of the functional style is provided by the Common Lisp program below. This function computes the sum of the positive elements of a vector. (defun sum-pos-vect (v) (reduce #'+ (remove-if-not #'plusp v))) A key feature of sum-pos-vect is that it makes use of an intermediate aggregate data structure (a sequence) to represent the selected set of vector elements. The use of sequences as intermediate quantities in computations makes it possible to use functional composition to express a wide variety of computations which are usually represented as loops. Unfortunately, as typically implemented, sequence expressions are extremely inefficient. The problem is that straightforward evaluation of a sequence expression requires the actual creation of the intermediate sequence objects. Since alternate algorithms using loops can often compute the same result without creating any intermediate sequences, the overhead engendered by using sequence expressions is quite reasonably regarded as unacceptable in many situations. A solution to the problem of the inefficiency of sequence expressions is to transform them into iterative loops which do not actually create any intermediate sequences before executing them. For example, sum-pos-vect might be transformed as shown below. (defun sum-pos-vect-transformed (v) (prog (index last sum element) (setq index 0) (setq last (length v)) (setq sum 0) L (if (not (< index last)) (return sum)) (setq element (aref v index)) (if (plusp element) (setq sum (+ element sum))) (setq index (1+ index)) (go L))) Several researchers have investigated the automatic transformation of sequence expressions into loops. For example, APL compilers transform many kinds of sequence expressions into loops. Unfortunately, there is a fundamental problem with the transformation of sequence expressions into loops. Although many sequence expressions can be transformed, many cannot. For example, Common Lisp provides a sequence function (reverse) which reverses the elements in a sequence. Suppose that a sequence expression enumerates a sequence, reverses it, and then reduces it to some value. This sequence expression cannot be computed without using intermediate storage for the enumerated sequence because the first element of the reversed sequence is taken from the last element of the enumerated sequence. There is no way to transform the sequence expression into an efficient loop without eliminating the reverse operation. A solution to the problems caused by the presence of non-transformable sequence operations is to restrict the kinds of sequence operations which are allowed so that every sequence expression is guaranteed to be transformable. For example, one could start by outlawing the operation reverse. LETS LetS supports a wide class of sequence expressions that are all guaranteed to be transformable into efficient loops. In order to avoid confusion with the standard Common Lisp data type sequence, the data type supported by LetS is called a series. Using LetS the program sum-pos-vect would be rendered as shown below. The function Evector converts the vector v into a series which contains the same elements in the same order. The function Tplusp is analogous to (remove-if-not #'plusp ...) except that it operates on a series. The function Rsum corresponds to (reduce #'+ ... :initial-value 0) except that it takes in a series as its argument. (defun sum-pos-vect-lets (v) (Rsum (Tplusp (Evector v)))) LetS automatically transforms the body of this program as shown below. The readability of the transformed code is reduced by the fact that it contains a large number of gensymed variables. However, the code is quite efficient. The only significant problem is that too many variables are used. (For example, the variable #:vector5 is unnecessary.) However, this problem need not lead to inefficiency during execution as long as a compiler which is capable of simple optimizations is available. (defun sum-pos-vect-lets-transformed (v) (let (#:index12 #:last4 #:sum21 #:element11 #:vector5) (tagbody (setq #:vector5 v) (setq #:index12 0) (setq #:last4 (length #:vector5)) (setq #:sum21 0) #:p0 (if (not (< #:index12 #:last4)) (go #:e9)) (setq #:index12 (1+ #:index12)) (setq #:element11 (aref #:vector5 #:index12)) (if (not (plusp #:element11)) (go #:p0)) (setq #:sum21 (+ #:element11 #:sum21)) (go #:p0) #:e9) #:sum21)) RESTRICTIONS ENFORCED BY LETS The key aspect of LetS is that it enforces a palatable (and not overly strict) set of easily understandable restrictions which guarantee that every series expression can be transformed into a highly efficient loop. This allows programmers to write series expressions which are much easier to work with than the loops they might otherwise write, without suffering a decrease in efficiency. There are two central restrictions which are enforced by LetS. First, every series must be statically identifiable so that transformation can occur at compile time rather than at run time. Second every series function is required to be "in-order". A series function is said to be in-order if it reads each input series in order, one element at a time, starting from the first one, and if it creates the output series (if any) in order, one element at a time, starting from the first one. In addition, the function must do this without using internal storage for more than one element at a time for each of the input and output series. For example, the series functions Evector, Tplusp, and Rsum are all in-order. In contrast, the function reverse is not in-order. (Reverse either has to read the input in reverse order, or save up the elements until the last one is read in.) OTHER FEATURES OF LETS Although efficiency is the main goal of LetS, LetS supports a number of features which are not directly related to efficiency per se. Most notable of these is implicit mapping of functions over series. Whenever an ordinary Lisp function is syntactically applied to a series, it is automatically mapped over the elements of the series. The following example illustrates implicit mapping. In the function below, the computation "(lambda (x) (expt (abs x) 3))" is implicitly mapped over the series of numbers generated by Evector. Implicit mapping of this sort is a commonly used feature of APL and is extremely convenient. (defun sum-cube-abs-vect (v) (Rsum (expt (abs (Evector v)) 3))) (sum-cube-abs-vect #(1 -2 3)) => (+ 1 8 27) => 36 New series functions can be defined by using the form defunS. The following example shows how the function Rsum could be defined. More complex forms can be defined by using the ordinary Common Lisp macro definition facilities to define macros which create appropriate series expressions. (defunS Rsum (numbers) (declare (series numbers)) (reduceS #'+ 0 numbers)) LetS provides two forms (LetS and LetS*) which are analogous to let and let*. As shown in the example below, These forms can be used to bind both ordinary variables (e.g., num-obs, mean, and deviation) and series variables (e.g., ob). Whether or not a variable is a series is determined by looking at the type of value produced by the expression which computes the value bound to it. (defun mean-and-deviation (observations) (letS* ((ob (Elist observations)) (num-obs (Rlength ob)) (mean (/ (Rsum ob) num-obs)) (deviation (- (/ (Rsum (expt ob 2)) num-obs) (expt mean 2)))) (list mean deviation))) The complete documentation of LetS compares LetS with the Common Lisp sequence functions and with the Zeta Lisp Loop macro. LetS supports essentially all of the functionality of the Loop macro in a style which looks like sequence functions and which is exactly as efficient as the loop macro. THE ANCESTRY OF LETS The LetS package described here is descended from an earlier package of the same name (See MIT/AIM-680a and "Expressional Loops", Proc. Eleventh ACM SIGACT-SIGPLAN Symposium on the Principles of Programming Languages, January 1984). The current system differs from the earlier system in a number of ways. In particular, the new system supports a much wider set of features. ------------------------------ End of AIList Digest ******************** From vtcs1::in% Thu Apr 10 18:55:55 1986 Date: Thu, 10 Apr 86 18:55:48 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #75 Status: RO AIList Digest Thursday, 10 Apr 1986 Volume 4 : Issue 75 Today's Topics: Games - Game-Playing Programs, Philosophy - Computer Consciousness & Wittgenstein and NL & Reply to Lucas on Formal Systems ---------------------------------------------------------------------- Date: Wed, 09 Apr 86 11:54:48 -0500 From: lkramer@dewey.udel.EDU Subject: Game-Playing Programs Re: Allen Sherzer's request for information of AI game-playing programs. I wrote a program last year for an expert systems course that plays the card game Spades. (ESP -- Expert Spades Player) It is implemented as a frame-based expert system written in minifrl (my revision of the frame primitives in Winston and Horn's Lisp) on top of Franz. The pro- gram is fairly simple-minded in that it doesn't learn from its mistakes or deal well with novel situations, but it still is able to play a fairly good game of Spades. In addition, since it is written as an expert system, its rule-base is easily modifiable. Mostow has written a (much more sophisticated) program that plays Hearts and is able to operationalize from fairly general advice. --1983, Mostow, D.J., Machine transformation of advice into a heuristic search procedure. In R.S. Michalski, J. Carbonell, and T. M. Mitchell, eds., Machine learning: An artificial Intelligence Approach. Tioga Press. ------------------------------ Date: 9 Apr 86 08:55:00 EST From: "CUGINI, JOHN" Reply-to: "CUGINI, JOHN" Subject: computer consciousness Thought I'd jump in here with a few points. 1. There's a metaphilosophers (don't ask me why the "meta") mailing list where folks thrash on about this stuff constantly, so if you care, listen in. Tune in to: MetaPhilosophers%MIT-OZ@MIT-MC. 2. There's a common problem with confusing epistemological questions (what would constitute evidence for computer consciousness) and ontological ones (so, is it *really* conscious). Those who subscribe to various verificationist fallacies are especially vulnerable, and indeed may argue that there is ultimately no distinction. The point is debatable, obviously, but we shouldn't just *assume* that the latter question (is it *really* conscious) is meaningless unless tied to an operational definition. After all, conscious experience is the classic case of a *private* phenomenon (ie, no one else can directly "look" at your experiences). If this means that consciousness fails a verificationist criterion of meaningfulness, so much the worse for verificationism. 3. Taking up the epistemological problem for the moment, it isn't as obvious as many assume that even the most sophisticated computer performance would constitute *decisive* evidence for consciousness. Briefly, we believe other people are conscious for TWO reasons: 1) they are capable of certain clever activities, like holding English conversations in real-time, and 2) they have brains, just like us, and each of us knows darn well that he/she is conscious. Clearly the brain causes/supports consciousness and external performance in ways we don't understand. A conversational computer does *not* have a brain; and so one of the two reasons we have for attributing consciousness to others does not hold. Analogy: suppose you know that cars can move, that they all have X-type-engines, and that there's something called combustion which depends on X-type-engines and which is instrumental in getting the cars to move. Let's say you have a combustion-detector which you tried out on one car and, sure enough, it had it, but then you dropped your detector and broke it. You're still pretty confident that the other cars have combustion. Now you see a very different type of vehicle which can move, but which does NOT have an X-type-engine - in fact you're not too sure whether it's really an engine at all. Now, is it just obvious that this other vehicle has combustion?? Don't we need to know a) a good definition of combustion, b) some details as to how X-type-engines and combustion are related? c) some details as to how motion depends on combustion, d) in what respects the new "engine" resembles/differs from X-type-engines, etc etc.? The point is that motion (performance) isn't *decisive* evidence for combustion (consciousness) in the absence of an X-type-engine (brain). John Cugini ------------------------------ Date: 2 Apr 86 08:58:24 GMT From: amdcad!cae780!leadsv!rtgvax!ramin@ucbvax.berkeley.edu (Pantagruel) Subject: Natural Language processing An issue that has propped up now and again through my studies has been the relation between current Natural Language/Linguistic research and the works of Ludwig Wittgenstein (especially through the whole Vienna School mess and later in his writings in "Philosophical Investigations"). It appears to me (in observing trends in such theories) and especially after the big hoopla over Frames that AI/Cognitive Research has spent the past 30 years experimenting through "Tractatus" and has just now warmed up to "P.I." The works of the Vienna School's context-free language analyses earlier in this century seems quite parallel to early context-free language parsing efforts. The later studies in P.I. with regards to the role of Natural Context and the whole Picture-Theory rot seems to have been a direct result of the failure of the context-free approach. Quite a few objections voiced nowadays by researchers on the futility of context-free analysis seems to be very similar to the early chapters in P.I. I still haven't gone through Wittgenstein with a fine enough comb as I would like... especially this latter batch of his notes that I saw a few weeks ago finally published and available publicly... But I still think there is quite a bit of merit to this fellow's study of language and cognition. Any opinions on this...? Any references to works to the contrary? I must be fair in warning that I hold Wittgensteins' works to contain the answers to some of the biggest issues facing us now... Personally, I'm holding out for someone to come up with some relevant questions... I think Bertrand Russell was correct in assessing L.W.'s significance... Please mail back to me for a livelier dialogue... The Net seems rather hostile nowadays... (but post to net if you think it merits a public forum)... "Pantagruel at his most vulgar..." = = = Alias: ramin firoozye | USps: Systems Control Inc. uucp: ...!shasta \ | 1801 Page Mill Road ...!lll-lcc \ | Palo Alto, CA 94303 ...!ihnp4 \...!ramin@rtgvax | ^G: (415) 494-1165 x-1777 = = = ------------------------------ Date: Fri, 4 Apr 86 13:34:54 est From: Stanley Letovsky Subject: Reply to Lucas At the conference on "AI an the Human Mind" held at Yale early in March 1986, a paper was presented by the British mathematician John Lucas. He claimed that AI could never succeed, that a machine was in principle incapable of doing all that a mind can do. His argument went like this. Any computing machine is essentially equivalent to a system of formal logic. The famous Godel incompleteness theorem shows that for any formal system powerful enough to be interesting, there are truths which cannot be proved in that system. Since a person can see and recognize these truths, the person can transcend the limitations of the formal system. Since this is true of any formal system at all, a person can always transcend a formal system, therefore a formal system can never be a model of a person. Lucas has apparently been pushing this argument for several decades. Marvin Minsky gave the rebuttal to this; he said that formal systems had nothing to do with AI or the mind, since formal systems required perfect consistency, whereas what AI required was machines that make mistakes, that guess, that learn and evolve. I was less sure of that refutation; although I agreed with Minsky, I was worried that because the algorithms for doing all that guessing and learning and mistake making would run on a computer, there was still a level of description at which the AI model must look like a consistent formal system. This is equivalent to the statement that your theory of the mind is a consistent theory. I was worried that Lucas could revive his argument at that level, and I wanted a convincing refutation. I have found one, which I will now present. First, we need to clarify the relationship between a running computer program and a system of formal logic. A running computer program is a dynamic object, it has a history composed of a succession of states of the machine. A formal system, by contrast, is timeless: it has some defining axioms and rules of inference, and a space of theorems and nontheorems implicitly defined by those axioms and rules. For a formal system to model a dynamic process, it must describe in its timeless manner the temporal behavior or history of the process. The axioms of the formal system, therefore, will contain a time parameter. They might look something like this: if the process is in a state of type A at time t1, it will be in a state of type B in the next instant. A more complicated problem is how the interaction between the computer program and the outside world is to be modelled within the formal system. You cannot simulate input and output by adding axioms to the formal system, because changing the axioms changes the identity of the system. Moreover, input and output are events in the domain of the running program; within the formal system they are just axioms or theorems which assert that such and such an input or output event occurred at such and such a time. The ideal solution to this problem is to include within the formal system a theory of the physics of the world as well as a theory of the mind. This means that you can't construct a theory of the mind until you have a theory of the rest of the universe, which seems like a harsh restriction. Of course, the theory of the rest of the universe need not be correct or very detailed; an extremely impoverished theory would simply be a set of assertions about sensory data received at various instants. Alternatively, you could ignore I/O completely and just concern yourself with a model of isolated thought; if we debunk Lucas' argument for this case we can leave it to him to decide whether to retreat to the high ground of embodied thinking machines. Therefore I will ignore the I/O issue. The next point concerns the type of program that an AI model of the mind is likely to be. Again, ignoring sensory and motor processing and special purpose subsystems like visual imagery or solid modelling, we will consider a simple model of the mind as a process whose task is belief fixation. That is, the job of the mind is to maintain a set of beliefs about the world, using some kind of abductive inference procedure: generate a bunch of hypotheses, evaluate their credibility and consistency using a variety of heuristic rules of evidence, and, on occasion, commit to believe a particular hypothesis. It is important to understand that the set of beliefs maintained by this program need not be consistent with each other. If we use the notation believes(Proposition,Instant) to denote the fact that the system believes a particular proposition at some instant, it is perfectly acceptable to have both believes(p,i) and believes(not(p),i) be theorems of the formal system which describes the program's behavior. The formal system must be a consistent description of the behavior of the program, or we do not have a coherent theory. The behavior of the program must match Lucas' (or some other person's) behavior or we do not have a correct theory. However the beliefs maintained by the program need not be a consistent theory of anything, unless Lucas happens to have some consistent beliefs about something. For those more comfortable with technical jargon, the formal system has a meta-level and an object level. The object level describes Lucas beliefs and is not necessarily consistent; the meta-level is our theory of Lucas' belief fixation process and had better be consistent. The object level is embedded in the meta-level using the modal operator "believes". What would it mean to formulate a Godel sentence for this system? To begin with, we seem to have a choice about where to formulate the Godel sentence: at the object level or the meta level. Formulating a Godel sentence for the object level, that is, the level of Lucas' beliefs, is clearly a waste of time, however. This level is not required to be consistent, and so Godel's trick of forcing us to choose between consistency and completeness fails: we have already rejected consistency. The more serious problem concerns a Godel sentence formulated for the meta-level, which must be consistent. The general form of a Godel sentence is G: not(provable(G)) where "provable" is a predicate which you embed in the system in a clever way, and which captures the notion of provability within the system. The meaning of such a sentence is "This sentence is not a theorem", and therein lies the Godelian dilemma: if the sentence is true, the system is incomplete because not all statable truths are theorems. If the sentence is false, then the system is inconsistent, because G is both true and false. This dilemma holds for all "sufficiently powerful" systems, and we assume that our model of Lucas falls into this category, and that one can therefore write down a Godel sentence for the model. What is critical to realize, however, is that the Godel sentence for our model of Lucas is not a belief of Lucas' according to the model. The form of the Godel sentence G: not(provable(G)) is syntactically distinct from the form of an assertion about Lucas' beliefs, believes(p,t) Nothing stops us from having believes(G,t) be provable in the system, despite the fact that G is not itself provable in the system. (Actually, the last sentence is incorrect, since it is illegal to put G inside the scope of the "believes" operator. G is a meta-level sentence, and only object level sentences are permitted inside "believes". The object level and the meta level are not allowed to share any symbols. If you want to talk about Lucas's beliefs about the model of himself, you will have to embed Lucas' model of the model of himself at the object level, but we can ignore this technicality.) This point is crucial: the Godel sentence for our theory of Lucas as a belief-fixing machine is not a theorem ascribing any beliefs to Lucas. Therefore the fact that Lucas can arrive at a belief that the Godel sentence is true is perfectly compatible with the fact that the system cannot prove G as a theorem. Lucas' argument depends on the claim that if he believes G, he transcends the formal system: this is his mistake. Lucas can believe whatever he wants about what sentences can or can't be proved within the model of himself. The only way his beliefs have any bearing on the correctness of the model is if the model predicts that Lucas will believe something he doesn't, or disbelieve something he believes. In other words, the usual criteria of science apply to judging the correctness of the model, and no Godelian sophistry can invalidate the model a priori. Lucas' argument has a certain surface plausibility to it. Its strength seems to depend on the unwarranted assumption that the theorems of the formal system correspond directly to the beliefs of the mind being modelled by that system. This is a naive and completely fallacious assumption: it ignores the fact that minds are temporal processes, and that they are capable of holding inconsistent beliefs. When these issues are taken into account, Lucas' argument falls flat. ------------------------------ End of AIList Digest ******************** From vtcs1::in% Thu Apr 10 18:56:17 1986 Date: Thu, 10 Apr 86 18:56:09 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #76 Status: R AIList Digest Thursday, 10 Apr 1986 Volume 4 : Issue 76 Today's Topics: Seminars - NL Interfaces to Expert Systems (Villanova) & Minsky (SIU-Edwardsville) & Frames and Objects in Modeling and Simulation (SU) & Machine Inductive Inference (UPenn) & Conditionals and Inheritance (CMU) & Knowledge Retrieval as Specialized Inference (CMU) & Ontology and Efficiency in a Belief Reasoner (UPenn) & Probabilistic Inference: Theory and Practice (SMU), Conference - Southern California AI Conference Program ---------------------------------------------------------------------- Date: Fri, 4 Apr 86 13:09 EST From: Tim Finin Subject: Seminar - NL Interfaces to Expert Systems (Villanova) I got an announcement in the mail this week about the first meeting of the DELAWARE VALLEY AI ASSOCIATION. It will be held at Villanova University (Tolentine Hall, room 215) on April 21st at 7:30pm. The meeting will discuss the organizational structure of the association, introduce the current officers, and feature a talk by Bonnie Webber on "Natural Language Interfaces to Expert Systems". DIRECTIONS: from rt. 320 North turn right onto route 30. At the first light, turn right into the parking lot. Walk across route 30 and proceed along the walkway towards the chapel. Turn left at the Chapel to Tolentine Hall, which is about 50 yards to the right. For more information, call 215-265-1980. ------------------------------ Date: 8 Apr 1986 13:30-EST From: ISAACSON@USC-ISI.ARPA Subject: Seminar - Minsky (SIU-Edwardsville) Marvin Minsky will be in the St. Louis area on Tuesday and Wednesday, April 22, 23. He'll give a talk at Southern Illinois University at Edwardsville on: THE SOCIETY OF MIND Science Labs Bldg., Room 1105 Tuesday, 7:30 pm April 22, 1986 Admission is free and people in the St. Louis area are welcome. ------------------------------ Date: Tue 8 Apr 86 16:27:21-PST From: Christine Pasley Subject: Seminar - Frames and Objects in Modeling and Simulation (SU) CS529 - AI In Design & Manufacturing Instructor: Dr. J. M. Tenenbaum Title: Frames and Objects: Application to Modeling And Simulation Speaker: Richard Fikes and Marilyn Stelzner From: Intellicorp Date: Wednesday, April 9, 1986 Time: 4:00 - 5:30 Place: Terman 556 We will describe the characteristic features of frame-based knowledge representation facilities and indicated how they can provide a foundation for a variety of knowledge-system functions. We will focus on how frames can contribute to a knowledge sytem's reasoning activities and how they can be used to organize and direct those activities. Application to engineering modelling and simulation will be discussed. Visitors welcome. ------------------------------ Date: Tue, 8 Apr 86 12:00 EST From: Tim Finin Subject: Seminar - Machine Inductive Inference (UPenn) Forwarded From: Dale Miller on Tue 8 Apr 1986 at 8:35 UPenn Math-CS Logic Seminar SOME RECENT RESEARCH ON MACHINE INDUCTIVE INFERENCE Scott Weinstein Tuesday, 8 April 1986, 4:30 - 6:00, 4N30 DRL The talk will survey some recent (and not so recent) results on the inference of r.e. sets and first-order structures. ------------------------------ Date: 8 Apr 1986 1416-EST From: Lydia Defilippo Subject: Seminar - Conditionals and Inheritance (CMU) Speaker: Rich Thomason Date: Thursday, April 17 Time: 3:00 pm Place: 4605 Topic: CONDITIONALS AND INHERITANCE This talk will provide motivation and an overview of an NSF-sponsored research project that has recently begun here, involving David Touretzky, Chuck Cross, Jeff Horty, and Kevin Kelly. The portion of the project on which I will concentrate aims at bringing logical work on conditionals to bear on nonmonotonic reasoning, and in particular on inheritance theory. Some of the background for the theory consists in the need for a qualitative approach to "belief kinematics" (or knowledge revision, or database update), as opposed to a quantitative approach such as the Bayesian one. The logic of conditionals provides some principles for such an approach, where the conditionals are interpreted as indicative expressions of willingness to make belief transitions. Although we have many firm intuitions about inheritance in particular cases, it is difficult to establish a correct general definition of nonmonotonic inheritance for arbitrary semantic nets. I will show how a definition of inheritance generates a definition of validity for simple conditional expressions, and will suggest that this can be used as a criterion to judge inheritance definitions. I will present some results relating particular inheritance definitions to conditional logics. These results depend on a kind of ad hoc update procedure for semantic nets. I will suggest that a better procedure might be obtained by considering nets with both monotonic and nonmonotonic links. If time permits, I will develop some analogies between semantic nets and Gentzen systems or natural deduction. ------------------------------ Date: 8 April 1986 1615-EST From: Betsy Herk@A.CS.CMU.EDU Subject: Seminar - Knowledge Retrieval as Specialized Inference (CMU) Speaker: Alan M. Frisch, University of Rochester Date: Tuesday, April 22 Time: 3:30 - 5:00 Place: 5409 Wean Hall Title: Knowledge retrieval as specialized inference Artificial intelligence reasoning systems commonly contain a large corpus of declarative knowledge, called a knowledge base (KB), and provide facilities with which the system's components can retrieve this knowledge. Consistent with the necessity for fast retrieval is the guiding intuition that a retriever is, at least in simple cases, a pattern matcher, though in more complex cases it may perform selected inferences such as property inheritance. Seemingly at odds with this intuition, the thesis of this talk is that the entire process of retrieval can be viewed as a form of inference and hence the KB as a representation, not merely a data structure. A retriever makes a limited attempt to prove that a queried sentence is a logical consequence of the KB. When constrained by the no-chaining restriction, inference becomes indistinguishable from pattern-matching. Imagining the KB divided into quanta, a retriever that respects this restriction cannot combine two quanta in order to derive a third. The techniques of model theory are adapted to build non-procedural specifications of retrievability relations, which determine what sentences are retrievable from what KB's. Model-theoretic specifications are presented for four retrievers, each extending the capabilities of the previous one. Each is accompanied by a rigorous investigation into its properties, and a presentation of an efficient, terminating algorithm that can be proved to meet the specification. ------------------------------ Date: Wed, 9 Apr 86 15:01 EST From: Tim Finin Subject: Seminar - Ontology and Efficiency in a Belief Reasoner (UPenn) Forwarded From: Bonnie Webber Forwarded From: Glenda Kent ONTOLOGY AND EFFICIENCY IN A BELIEF REASONER Anthony S. Maida Department of Computer Science Penn State University This talk describes the implementation of, and theoretical influences underlying, a belief reasoner called the "Belief Space Engine." A belief reasoner is a program that reasons about the "beliefs" of other agents. The Belief Space Engine uses specialized data structures, called belief spaces, to compute a certain class of inferences about the beliefs of other agents efficiently. Theoretically, the architecture is motivated by a syntactic simulation ontology, which is an alternative to the possible-worlds ontology. In order to encode this ontology, a meta description facility has been implemented. This talk is organized as follows. First, we explain the semantic difficulties with belief reasoning that stem from interactions between belief, equality, and quantification. Next, we argue for the sufficiency of the syntactic simulation ontology to address the difficulties we described. Then we show how the ontology is partially embodied in the Belief Space Engine. Finally, we show that the Belief Space Engine is robust in this domain by programming several examples. Thursday, April 10, 1986 Room 216 - Moore School 3:00 - 4:30 p.m. Refreshments Available ------------------------------ Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Seminar - Probabilistic Inference: Theory and Practice (SMU) Title: Probabilistic Inference: Theory and Practice Speaker: Won D. Lee University of Illinois at Urbana- Champaign Location: 315SIC Time: 2:00 PM This talk presents a system and a methodology for probabilistic learning from examples. First, I present a new methodology, Probabilistic Rule Generator (PRG), of variable-valued logic synthesis which can be applied effectively to noisy data. Then a new system, Probabilistic Inference, which can generate concepts with limited time and/or resources is defined. It is discussed how PRG can be a practical tool for Probabilistic Inference. A departure from the classical viewpoint in logic minimization, and in knowledge acquisition is reported. ------------------------------ Date: Wed, 9 Apr 86 19:52:30 PST From: cottrell@nprdc.arpa (Gary Cottrell) Subject: Conference - Southern California AI Conference Program Southern California Conference on Artificial Intelligence Saturday, April 26, 1986 Peterson Hall UCSD Sponsored by San Diego SIGART and SCAIS 9:00am Registration Desk Opens 10:00am-12:00pm Invited Overviews 10:00am-10:25am AI Environment and Research at UCLA Michael G. Dyer and Josef Skrzypek, UCLA AI Lab 10:30am-10:55am Ai Research at USC Peter Norvig, USC 11:00am-11:25am Parallel Distributed Processing: Explorations in the Microstructure of Cognition David E. Rumelhart, Institute for Cognitive Science, UCSD 11:30am-11:55am Human Computer Interaction: Research at the Intelligent Systems Group Jim Hollan, Intelligent Systems Group, UCSD 12:00-1:00 Buffet Lunch 1:00pm-3:00pm SCAIS Session I: Expert Systems 1:00pm-1:15pm RAMBOT: A connectionist expert system that learns by example Michael C. Mozer, Institute for Cognitive Science, UCSD 1:20pm-1:35pm A small expert system that learns George S. Levy, Counseling and Consulting Associates, San Diego 1:40pm-1:55pm A knowledge based selection system Xi-an Zhu, Dept. of Electrical Engineering, USC 2:00pm-2:15pm STYLE Counselor: An expert system to select ties Jeffrey Blake, Peter Tenereillo, and Jeff Wicks Department of Mathematical Sciences, SDSU 2:20pm-2:35pm A health and nutrition expert system Marwan Yacoub, Department of Mathematical Sciences, SDSU 2:40pm-2:55pm An inexact reasoning scheme based on intervals of probabilities Koenraad Lecot, Computer Science Dept., UCLA 1:00pm-3:00pm SCAIS Session 2: Vision and Natural Language 1:00pm-1:15pm A Scheme-based PC vision workstation Michael Stiber and Josef Skrzypek, CS Dept., UCLA and CRUMP Inst. 1:20pm-1:35pm Early Vision: 3-D silicone solution to lightness constancy Paul C. H. Lin and Josef Skrzypek, CS Dept., UCLA and CRUMP Inst. 1:40pm-1:55pm A connectionist computing architecture for textural segmentation Edmond Mesrobian and Josef Skrzypek, CS Dept., UCLA and CRUMP Inst. 2:00pm-2:15pm ANIMA: Analogical Image Analysis Arthur Newman, Computer Science Dept., UCLA 2:20pm-2:35pm Representing pragmatic knowledge in lexical memory Michael Gasser, Artificial Intelligence Laboratory, UCLA 2:40pm-2:55pm The role of mental spaces in establishing universal principles for the semantic interpretation of cliches Michelle Gross, Linguistics Dept., UCSD 1:00pm-3:00pm SIGART Session 1 1:00pm-1:25pm Using commonsense knowledge for prepositional phrase attachment K. Dahlgren, IBM 1:30pm-1:55pm Social Intelligence Les Gasser, Computer Science Dept., USC 2:00pm-2:25pm A unified algebraic theory of logic and probability Philip Calabrese, LOGICON 2:30pm-2:55pm Learning while searching in constraint- satisfaction problems Rina Dechter, & Hughes AI Center Cognitive Systems Lab, UCLA 3:00-3:30 Coffee Break 3:30pm-5:30pm SCAIS Session 3: Connectionist Models & Learning 3:30pm-3:45pm Toward optimal parameter selection in the back-propagation algorithm Yves Chauvin, Institute for Cognitive Science, UCSD 3:50pm-4:05pm Inverting a connectionist network mapping by back-propagation of error Ron Williams, Institute for Cognitive Science, UCSD 4:10pm-4:25pm Learning internal representations from gray scale images Gary Cottrell and Paul Munro, Institute for Cognitive Science, UCSD 4:30pm-4:45pm Decomposition in perceptron systems Rik Verstraete, Computer Science Dept., UCLA 4:50pm-5:05pm Adaptive Self-Organizing Logic Networks Tony Martinez, *** 5:10pm-5:25pm Human understanding in diverse environments Louis Rossi, Harvey Mudd College 3:30pm-5:30pm SCAIS Session 4: Miscellaneous (HMI, Planning, Problem Solving, Knowledge Representation) 3:30pm-3:45pm Producing coherent interactions in a tutoring system Balaji Narasimhan, Computer Science Dept., USC 3:50pm-4:05pm AQUA: An intelligent UNIX advisor Alex Quilici, Artificial Intelligence Laboratory, UCLA 4:10pm-4:25pm Errors in parsing problem descriptions Eric Hestenes, Problem Solving Group, UCSD 4:30pm-4:45pm Constraint based problem solving Mitchell Saywitz, Computer Science Dept., USC 4:50pm-5:05pm An approach to planning and scheduling for robot assembly lines Xiaodong Xia, Computer Science Dept., USC 5:10pm-5:25pm Changes of mind: Revision of "interpretation" in episodic memory Antoine Cornuejols, Computer Science Dept. UCLA 3:30pm-5:30pm SIGART Session 2 3:30pm-355pm Facilitating parametric analyses with AI methodologies N. T. Gladd, JAYCOR 4:00pm-4:25pm Computer Chess: Arguments and examples for a knowledge-based approach Danny Kopec, Dept. of Mathematical Sciences, SDSU 4:30pm-4:55pm Artificial Intelligence applications in information retrieval Mark Chignell, Dept. of Industrial & Systems Engineering, USC ------------------------------ End of AIList Digest ******************** From vtcs1::in% Fri Apr 11 19:02:24 1986 Date: Fri, 11 Apr 86 19:02:18 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #77 Status: R AIList Digest Friday, 11 Apr 1986 Volume 4 : Issue 77 Today's Topics: Bibliographies - AI Subject Codes & Report Sources & Technical Reports #1 ---------------------------------------------------------------------- Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: AI Subject Codes The following is a list of subject codes that are being put in the %K field of all bibliographies going out to AILIST. Hopefully, this will be of assistance to people in finding material on their favorite subfield of artificial intelligence. This searching is best done with the bib or refer utilities but could be done less conveniently with more general-purpose utilities. For example, if one is interested in applications of expert systems to electrical engineering one would search for AI01 and AA04. ______ AI areas AI01 Expert Systems, Rule Based Systems AI02 Natural Language AI03 Search (Minimax, Consistant Labelling, alpha-beta, etc.) AI04 Learning AI05 Speech Understanding AI06 Vision, Pattern Recogniton AI07 Robotics AI08 Cognitive Science AI09 Planning AI10 Logic Programming (material on prolog only will be under T02) AI11 Theorem Proving AI12 Neural Networks, Genetic Algorithms, etc. AI13 Decision Support AI14 Symbolic Math Application Areas AA01 Medicine AA02 Chemistry AA03 Geology, Mineral Extraction, Petroleum Extraction and Geology AA04 Electrical Engineering AA05 Other Engineering, Unclassifiable Engineering AA06 Financial, Business, Marketing, Accounting, Etc. AA07 Education AA08 Software Engineering, Automatic Programming, Computer Configuration and Operation AA09 Data Bases AA10 Biology AA11 Social Sciences AA12 Statistics AA13 Mathematics AA14 Information Retrieval AA15 User Interfaces to other Software AA16 Other Physcial Science AA17 Game Playing AA18 Military Applications AA19 Operating Equipment, e. g. pilots associate, autonomous land vehicle AA20 Process Control AA21 Diagnostic and Maintenance Systems (Other than Medical) AA22 Configuration Systems AA23 Agriculture AA24 Legal AA25 Art, Humanities, Music, Architecture, entertainment etc. Geographical Areas GA01 Japan GA02 United States GA03 Europe GA04 Canada Tools for AI T01 Lisp T02 Prolog T03 Expert System Tools Hardware for AI H01 Microcomputers H02 Lisp Machines H03 Parallel Processing H04 Supercomputers, e. g. Crays Other Areas O01 User Interfaces for AI systems O02 Software Engineering Issues in the Construction of AI programs O03 Real Time O04 Fuzzy Logic, Uncertainty Issues, etc. O05 Social Aspects of AI Article Types AT01 Advertisements AT02 Product Announcements AT03 Examples of AI Hype AT04 Market Predictions AT05 Interviews with Executives of Companies AT06 Other Interviews AT07 Book Reviews AT08 Tutorial Articles AT09 Bibliography AT10 Announcements of Company University Interactions AT11 New Bindings AT12 Letters to the Editors AT13 Corrections AT14 Pronouncements of Famous People AT15 BOOK AT16 Company Business, e. g. new financing, revenue announcements, joint marketing agreements etc. AT17 Software Reviews AT18 Articles on AI topic eduation AT19 Notes about Grantsmanship and Research Milieu type issues AT20 History of AI topics AT21 Bibliography ------------------------------ Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Report Sources Naomi Schulman, Publications COMPUTER SYSTEMS LABORATORY STANFORD UNIVERSITY Stanford, CA 94305 UCLA COMPUTER SCIENCE DEPARTMENT University of California 3713 Boelter Hall Los Angeles, CA 90024 Cindy Hathaway, technical reports secretary, computer science department, louisiana state university, baton rouge, louisiana 70803, or cindy@lsu on csnet California Institute of Technology Computer Science, 256-80 Pasadena California 91125 Electrical Engineering and Computer Science Departments Stevens Institute of Technology Castle Point Station Hoboken, New Jersey 07030 Computer Science Department University of Rochester Rochester, New York 14627 Ms. Sally Goodall Technical Reports Librarian Computer Science Department SUNY Albany LI 67A Albany, New York 12222 Technical Reports Department of Computer Science Campus Box 1045 Washington University St. Louis, Missouri 63130 Department of Computer Science 136 Lind Hall University of Minnesota, Twin cities 207 Church Street SE Minneapolis, Minnesota 55455 IBM T.J. Watson Research Center Distribution Services, F-11, Stormytown P.O. Box 218 Yorktown Heights, NY 10598 ------------------------------ Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Technical Reports #1 %A Bruce Abramson %T A Cure for Pathological Behavior in Games that use Minimax %R CUCS-153-85 %I Columbia University %C New York City %K AI03 AA17 %A Peter K. Allen %A Ruzena Bajcsy %T Integrating Sensory Data for Object Recognition Tasks %R CUCS-184-85 %I Columbia University %C New York City %K AI06 %A Peter Kirby Allen %T Object Recognition Using Vision and Touch %R CUCS-220-85 %I Columbia University %C New York City %K AI06 AI07 %A Terrance E. Boult %T Reproducing Kernels for Visual Surface Interpolation %R CUCS-186-85 %I Columbia University %C New York City %K AI06 %A Terrance E. Boult %T Visual Surface Interpolation: A Comparison of Two Methods %R CUCS-189-85 %I Columbia University %C New York City %K AI06 %A Galina Datskovsky %T Menu Interfaces to Expert Systems: Overview and Evaluation %R CUCS-168-84 %I Columbia University %C New York City %K O01 AI01 %A Galina Datskovsky %T Natural Language Interfaces to Expert Systems %R CUCS-169-85 %I Columbia University %C New York City %K AI01 O01 AI02 %A Thomas Ellman %T Generalizing Logic Circuit Designs by Analyzing Proofs of Correctness %R CUCS-190-85 %I Columbia University %C New York City %K AA04 %A Bruce K. Hilyer %A David Elliot Shaw %T Execution of OPS5 Production Systems on a Massively Parallel Machine %R CUCS-147-84 %I Columbia University %C New York City %K AI01 H03 %A Bruce K. Hillyer %T A Knowledge-Based Expert Systems Primer and Catalog %R CUCS-195-85 %I Columbia University %C New York City %K AI01 AT08 %A Hussaein A. H. Ibrahim %A John R. Kender %A David Elliot Shaw %T On the Application of Massively Parallel SIMD Tree Machines to Certain Intermediate-Level Vision Tasks %I Columbia University %C New York City %R CUCS-221-85 %K AI06 H03 %A Husein A. H. Ibrahim %A John R. Kender %A David Elliot Shaw %T SIMD Tree Algorithms for Image Correlation %R CUCS-222-86 %I Columbia University %C New York City %K AI06 H03 %A Toru Ishida %A Salvatore Stolfo %T Towards the Parallel Execution of Rules in Production Systems Programs %R CUCS-154-84 %I Columbia University %C New York City %K H03 AI01 %A John R. Kender %A David Lee %A Terrance Boult %T Information Based Complexity Applied to Optimal Recovery of the 2 1/2-D Sketch %R CUCS-170-85 %I Columbia University %C New York City %K AI06 %A Richard E. Korf %T Macro-Operators: A Weak Method for Learning %R CUCS-156-85 %I Columbia University %C New York City %K AI04 %A Richard E. Korf %T Depth-First Iterative-Depending: An Optimal Admissible Tree Search %R CUCS-197-85 %I Columbia University %C New York City %K AI03 %A Michael Lebowitz %T The Use of Memory in Text Processing %R CUCS-200-85 %I Columbia University %C New York City %K AI02 Researcher Patent %A Michael Lebowitz %T Integrated Learning: Controlling Explanation %R CUCS-201-85 %I Columbia University %C New York City %K AI04 Unimem %A MIchael Lebowitz %T Story Telling and Generalizations %R CUCS-202-85 %I Columbia University %C New York City %K AI02 %A Michael Lebowitz %T Researcher: An Experimental Intelligent Information Systems %R CUCS-171-85 %I Columbia University %C New York City %K AI02 AA14 %A David Lee %T Contributions to Information-Based Complexity, Image Understanding, and Logic Circuit Desing %R CUCS-182-85 %I Columbia University %C New York City %K AI06 AA04 %A David Lee %T Optimal Algorithms for Image Understanding: Current Status and Future Plans %R CUCS-183-85 %I Columbia University %C New York City %K AI06 %A Mark D. Lerner %A Michael von Biema %A Gerald Q. Maguire, Jr. %R CUCS-146-85 %I Columbia University %C New York City %K PSL PPSL H03 T01 %A Mark D. Lerner %A Gerald Q. Maguire, Jr. %A Salvatore J. Stolfo %T An Overview of the DADO Parallel Computer %R CUCS-157-85 %I Columbia University %C New York City %K H03 AI01 T03 AI10 %A Andy Lowery %A Stephen Taylor %A Salvatore J. Stolfo %T LPS Algorithms %R CUCS-203-84 %I Columbia University %C New York City %K AI10 H03 %A Kevin Matthews %T Taking the Initiative for System Goals in Cooperative Dialogue %R CUCS-150-85 %I Columbia University %C New York City %K advisor AI02 %A Kevin Matthews %T Initiatory and Reactive System Roles in Human Computer Discourse %R CUCS-151-85 %I Columbia University %C New York City %K advisor AI02 %A Kathleen R. McKeown %A Myron Wish %A Kevin Matthews %T Tailoring Explanations for the User %R CUCS-172-85 %I Columbia University %C New York City %K AI01 AI02 O01 %A Katthleen R. McKeown %T The Need for Text Generation %R CUCS-173-85 %I Columbia University %C New York City %K AI01 AI02 O01 %A Kathleen R. McKeown %T Discourse Strategies for Generating Natural Language Text %R CUCS-204-85 %I Columbia University %C New York City %K AI02 AA09 %A Mark L. Moerdler %A John R. Kender %T Surface Orientation and Segmentation from Perspective Views of Parallel-Line Textures %R CUCS-159-85 %I Columbia University %C New York City %K AI06 %A Luanne Burns %A Alexander Pasik %T A Generic Framework for Expert Data Analysis Systems %R CUCS-163-85 %I Columbia University %C New York City %K AI01 AA09 %A Alexander Pasik %A Jans Christensen %A Douglas Gordin %A Agata Stancato-Pasik %A Salvatore Stolfo %T Explanation and Acquisition in Expert System Using Support Knowledge %R CUCS-164-85 %I Columbia University %C New York City %K AI01 AA01 DTEX %A K. S. Roberts %T Equivalent Descriptions of Generalized Cylinders %R CUCS-210-85 %I Columbia University %C New York City %A Salvatore J. STolfo %A Daniel P. Miranker %T The DADO Production System Machine %R CUCS-213-84 %I Columbia University %C New York City %K H03 AI01 %A Salvatore J. STolfo %A Daniel M. Miranker %A Russel C. Mills %T More Rules May Mean Faster Parallel Execution %I Columbia University %C New York City %R CUCS-175-85 %K RETE H03 AI01 %A Salvatore J. STolfo %A Daniel M. Miranker %A Russel C. Mills %T A Simple Preprocessing Scheme to Extract and Balance Implicit Parallelism in the Concurrent Match of Production Rules %R CUCS-174-85 %I Columbia University %C New York City %K H03 AI01 RETE T02 AI10 %A Peter Waldes %A Janet Lustgarten %A Salvatore J. Stolfo %T Are Maintenance Expert Systems Practical Now? %R CUCS-166-85 %I Columbia University %C New York City %K AI01 AA04 Automated Cable Expert Telephone AA21 %A J. F. Traub %T Information Based Complexity %R CUCS-162-85 %I Columbia University %C New York City %K AI03 %X information-based complexity is based on the assumption that it is partial, contaminated and it costs in comparison to ordinary complexity theory in which information is complete, exact and free. [There were several reports on this subject. I am only including one in this bibliography as it is not clear whether it is related to AI or not. Contact Columbia for more info if desired. LEFF] %A Kenneth Hal Wasserman %T Unifying Representation and Generalization: Understanding Hierarchically Structured Objects %R TCUCS-177-85 %I Columbia University %C New York City %K AA06 %X describes a system to understand upper-level corporate management hierarchies %A Ursula Wolz %T Analyzing User Plans to Produce Informative Responses by a Programmers' Consultant %R CUCS-218-85 %I Columbia University %C New York City %K AA08 AI02 AI09 AA15 %A Othar Hansson %A Andrew E. Mayer %A Mordechai M. Yung %T Generating Admissible Heuristics by Criticizing Solutions to Relaxed Models %R CUCS-219-85 %I Columbia University %C New York City %K AI03 %A Carolyn L. Talcott %T The Essence of Rum A Theory of the Intensional and Extensional Aspects of Lisp-type Computation %D AUG 1985 %R STAN-CS-85-1060 %I Stanford University Computer Science %K AI11 T01 %X $9.50 %A David E. Smith %A Michael R. Genesereth %T Controlling Recursive Inference %D JUN 1985 %R STAN-CS-85-1063 %I Stanford University Computer Science %K AI11 %X $3.75 %A Matthew L.Ginsberg %T Decison Procedures %D MAY 1985 %R STAN-0CS-85-1064 %I Stanford University Computer Science %K H03 %X The assumption of common rationality that is provably optimal (in a formal sense) and which enables us to characterize precisely the communication needs of the participants in multi-agent interactions. .br $2.75 %A William J. Clancey %T Review of Sowa's Conceptual Structures %D MAR 1985 %R STAN-CS-85-1065 %I Stanford University Computer Science %K AT07 %X $2.75 %A William J. Clancey %T Heuristic Classification %D JUN 1985 %R STAN-CS-85-1066 %I Stanford University Computer Science %K AI01 %X $4.75 %A William J. Clancey %T Acquiring, Representing, and Evaluating a Competence Model of Diagnostic Strategy %D AUG 1985 %R STAN-CS-85-1067 %I Stanford University Computer Science %K AI01 AA01 %X $4.95 %A Mark H. Richer %A William J. Clancey %T Guidon-Watch: A Graphic Interface for Viewing a Knowledge-Based System %D AUG 1985 %R STAN-CS-85-1068 %I Stanford University Computer Science %K AI01 O01 %X $2.75 %A John D. Hobby %T Digitized Brush Trajectories %D SEP 1985 %R STAN-CS-85-1070 %I Stanford University Computer Science %K AI06 %X $5.75 %A Russel Greiner %T Learning by Understanding Analogies %D SEP 1985 %R STAN-CS-85-1071 %I Stanford University Computer Science %K AI04 %X $15.00 %A Bruce G. Buchanan %T Expert Systems: Working Systems and The Research Literature %D OCT 1985 %R STAN-Cs-85-1075 %K AI01 %X $3.00 %A Deepinder P. Sidhu %T Protocol Verification Using Prolog %D JUL 1985 %R TR #85-21 %I Iowa State University %K AA08 Communications ISO/ISI T02 %A M. Attisha %A M. Yazdani %T A Microcomputer-based Tutor for Teaching Arithmetic Skills %D 1983 %I Department of Computer Science, University of Exeter, UK %K H01 AA07 GA03 %A M. Attisha %A M. Yazdani %T An Expert System for Diagnosing childrens' Multiplication Errors %D 1983 %I Department of Computer Science, University of Exeter, UK %K H01 AI01 AA07 GA03 %A A. Attisha %T A Microcomputer-based Tutoring System for Self-Improving and Teaching Techniques in Arithmetic Skills %D 1983 %I Department of Computer Science, University of Exeter, UK %K H01 AA07 PET Non Borrow Subtraction Algorithm Buggy Debuggy GA03 %A M. Yazdani %T Artificial Intelligence and Education %D 1984 %I Department of Computer Science, University of Exeter, UK %R Research Report NO. R122 %K H01 AI01 AA07 T02 GA03 %A J. Barchan %A B. Woodmansee %A M. Yazdani %T A PROLOG-based tool for French Grammar Analysis %I Department of Computer Science, University of Exeter, UK %R Research Report No. R126 %K AI02 GA03 FROG AA07 T02 %A M. Yazdani %T Intelligent Tutoring Systems: An Overview %I Department of Computer Science, University of Exeter, UK %R Working Paper No. W. 136 %K AI01 AA07 GA03 %A M. Yazdani %T Artificial Intelligence, Powerful Ideas and Education %I Department of Computer Science, University of Exeter, UK %R Working Paper No. W 138 %K Computer Assisted Learning AA07 GA03 %A Bruce Abramson %T An Explanation of and Cure for Minimax Pathology %R CSD-850034 %I University of California, Los Angeles %K AA17 AI03 %X The minimax procedure has long been the standard method of evaluating nodes in game trees. The general assumption underlying its use in game-playing programs is that increasing search depth improves play. Recent work has shown that this assumption is not always valid; for a large class of games and evaluation functions, searching deeper decreases the probability of making a correct move. This phenomenon is called game tree pathology. Two structural properties of game trees have been suggested as causes of pathology: independence among the values of sibling nodes, and uniform depth of wins and losses. This paper examines the relationship between uniform win depth and pathology from two angles. First, it proves mathematically that as search deepens, an evaluation function that does not .ul ask whether wins can be forced from mid-game positions becomes decreasingly likely to choose forced wins. Second, it experimentally illustrates the connection between recognition of mid-game wins and pathological behavior. Two evaluation functions, which differ only in their ability to recognize wins in mid-game, are run on a series of games. Despite recognizing fewer mid-game wins than the theoretically predicted minimum needed to avoid pathology, the function that checked for them cleared up the pathological behavior of the one that did not. The analytic and empirical aspects of this paper combine to form one major result: As search deepens, so does the probability that failing to check for forced wins will change the game's outcome. This strengthens the hypothesis that uniform win depth is the cause of pathology. $1.50 %A Michael Dyer %A Eric Quilici %T Human Problem Understanding and Advice Giving: A Computer Model %R CSD-850039 %I University of California, Los Angeles %K AA15 AI02 Aqua Unix Advisor AI09 %X How are people able to understand someone else's problem and provide them with advice? How are people able to develop novel solutions to problems they have never seen before? The thesis presented here is a first step toward answering these questions, presenting a computer model of the process of problem understanding and advice giving. The problems we consider are typical planning problems that novice computer users encounter. We view advice giving as a memory search problem, guided by heuristics for problem understanding, advice generation, and plan creation. In this thesis we describe a representational system for user planning problems, show how advice can be generated using a taxonomy of planning problems and associated heuristics for advice formulation, present heuristics that can be used to repair failed plans and to create new plans by combining existing pl ans in novel ways, and suggest a memory organization for planning knowledge that allows for efficient retrieval of relevant planning experiences. The theory discussed in this thesis is implemented in a computer program called AQUA (Alex Quilici's UNIX|+ Advisor). AQUA takes natural language descriptions of problems users are having with the UNIX operating system and provides natural language advice that explains their failures and suggests solutions. AQUA is also able to create solutions for problems that it has not been presented with before. $7.75 %A Judea Pearl %A Azaria Paz %T Graphoids A Graph-Based Logic for Reasoning About Relevance Relations %I University of California, Los Angeles %R CSD-850038 %K AI11 %X We consider 3-place relations I (x,z,y) where, x,y, and z are three non-intersecting sets of elements, (e.g., propositions), and I (x,z,y) stands for the statement: "Knowing z renders x irrelevant to y". We give sufficient conditions on I for the existence of a (minimal) graph G such that I (x,z,y) can be validated by testing whether z separates x from y in G. These conditions define a GRAPHOID. The theory of graphoids uncovers the axiomatic basis of probabilistic dependencies and ties it to vertex-separation conditions in graphs. The defining axioms can also be viewed as inference rules for deducing which propositions are relevant to each other, given a certain state of knowledge. ------------------------------ End of AIList Digest ******************** From vtcs1::in% Fri Apr 11 19:02:05 1986 Date: Fri, 11 Apr 86 19:02:01 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #78 Status: R AIList Digest Friday, 11 Apr 1986 Volume 4 : Issue 78 Today's Topics: Bibliography - Technical Reports #2 ---------------------------------------------------------------------- Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Technical Reports #2 %A C. V. Srinivasan %T Knowledge Processing Versus Programming: CK-LOG vs PROLOG %R DCS-TR-160 %I Rutgers University Laboratory for Computer Science %K AI10 CK-LOG T02 %A B. A. Nadel %T The Consistent Labeling Problem, Part 1: Background and Problem Formulation %R DCS-TR-164 %I Rutgers University Laboratory for Computer Science %A B. A. Nadel %T the Consistent Labeling Problem, Part 2: Subproblems, Enumerations and Constraint Satisfiability %R DCS-TR-165 %I Rutgers University Laboratory for Computer Science %A B. Nadel %T The Consistent Labeling Problem, Part 3: The Generalized Backtracking Algorithm %R DCS-TR-166 %I Rutgers University Laboratory for Computer Science %A B. A. Nadel %T The Consistent Labeling Problem, Part 4: The Generalized Forward Checking and Word-Wise Forward Checking Algorithms %R DCS-Tr-167 %I Rutgers University Laboratory for Computer Science %K AI03 %A T. M. Mitchell %A B. M. Keller %A S. T. Kedar-Cabelli %T Explanation-Based Generalization: A Unifying View %R ML-TR-2 %I Rutgers University Laboratory for Computer Science %K analogy %A S. T. Kedar-Cabelli %T Analogy - From a Unified Perspective %R ML-Tr-3 %I Rutgers University Laboratory for Computer Science %A R. M. Kellar %A S. T. Kedar-Cabelli %T Machine Learning Research at Rutgers University %R ML-Tr-4 %I Rutgers University Laboratory for Computer Science %K AI04 %X collection of research summaries in learning from Rutgers University %A R. Kurki-Suonio %T Towards Programming with Knowledge Expressions %I Carnegie Mellon Computer Science %K H03 %D AUG 1985 %A J. Laird %A P. Rosenbloom %A A. Newell %T Chunking in Soar: The Anatomy of a General Learning Mechanism %I Carnegie Mellon Computer Science %D SEP 1985 %K AI04 %A B. D. Lucas %T Generalized Image Matching by the Method of Differences %D JUL 1984 %I Carnegie Mellon Computer Science %K AI06 %A J. B. Saxe %T Decomposable Searching Problems and Circuit Optimization by Retiming: Two Studies in General Transformations of Computational Structures %D AUG 1985 %I Carnegie Mellon Computer Science %K AA04 AI03 %A E. S. Cohen %A E. T. Smith %A L. A. Iverson %T Constraint-Based Tiled Windows %D OCT 1985 %I Carnegie Mellon Computer Science %K AA15 %A A. Hisgen %T Optimization of User-Defined Abstract Data Types: A Program Transformation Approach %D SEP 1985 %I Carnegie Mellon Computer Science %K AA08 %A A. J. Kfoury %A Pawl Urzyczyn %T Necessary and Sufficient Conditoins for the Universality of Programming Formalisms %D MAY 1985 %R 85-007 %I Boston University Computer Science Department %K AA08 %X $4.00 %A Bipin Indurkhya %T Constrained Semantic Transference: A Formal Theory of Metaphors %D JUN 1985 %R 85-008 %I Boston University Computer Science Department %K AI02 %X $3.00 %A Bipin Indurkhya %T Approximate Semantic Transference: A Computational Theory: A Computational Theory of Metaphors and Analogies %D OCT 1985 %R BUCS 85-012 %I Boston University Computer Science Department %K AI02 AI11 %X $3.00 %A Weiguo Wang %T Computational Linguistics Technical Notes %D NOV 1985 %R BUCS 85-013 %I Boston University Computer Science Department %K AI02 %X $3.00 %A Gerhart %T A Test Data Generation Method Using Prolog %R TR-85-02 %I Wang Institute of Graduate Studies %K AA08 %A Velasco %T Computer Vison and Image Understanding %R TR-85-09 %I Wang Institute of Graduate Studies %K AI06 %A Gerhart %T Software Engineering Perspectives on Prolog %R TR-85-13 %I Wang Institute of Graduate Studies %K T02 O02 %A Gerhart %T A Detailed Look at Some Prolog Code: A Course Scheudler %R TR-85-14 %I Wang Institute of Graduate Studies %K O02 T02 %A Gerhart %T Several Prolog Packages %R Tr-85-15 %I Wang Institute of Graduate Studies %K T02 %A Van Nguyen %A David Gries %A Susan Owicki %R CSL T.R. 85-270 %T A MODEL AND TEMPORAL PROOF SYSTEM FOR NETWORKS OF PROCESSES %D February 1985 %I Stanford University Computer Systems Laboratories %K AI11 AA08 %X 12 pages.....$2.40 .br A model and a sound and complete proof system for networks of processes in which component processes communicate exclusively through messages is given. The model, an extension of the trace model, can describe both synchronous and asynchronous networks. The proof system uses temporal-logic assertions on sequences of observations - a generalization of traces. The use of observations (traces) makes the proof system simple, compositional and modular, since internal details can be hidden. The expressive power of temporal logic makes it possible to prove temporal properties (safety, liveness, precedence, etc.) in the system. The proof system is language-independent and works for both synchronous and asynchronous networks. %A W. E. Cory %T Verification of Hardware Design Correctness; Symbolic Execution Techniques and Criteria for Consistency %R TR 83-241 %I Stanford University Computer Systems Laboratory %X 118 pages, $6.15 %A S. Demetrescu %T High Speed Image Rasterization Using a Higly Parallel Smart bulk Memory %R TR 83-244 %I Stanford University Computer Systems Laboratory %K AI06 H03 %X 38 pages $3.40 %A A. L. Lansky %A S. S. Owicki %T GEM: A Tool for Concurrency Specification and Verification %R TR 83-251 %I Stanford University Computer Systems Laboratory %K AI11 AA08 %X 16 pages $2.55 %H TR84-018 %A Krzysztof J. Kochut %T UW LISP Manual %R LSU Computer Science Technical Report 84-018 %K T01 %H TR84-025 %A E. T. Lee %T Application of Fuzzy Languages to Medical Pattern Recognition %R LSU Computer Science Technical Report TR84-025 %K AI06 %H TR84-026 %A E. T. Lee %T Similarity Directed Chromosome Image Processing %R LSU Computer Science Technical Report TR84-026 %K AI06 %H TR84-029 %A S. S. Iyengar %A T. Sadler %A S. Kundu %T A Technique for Representing a Tree Structure with Predicates by a Forest Data Structure %R LSU Computer Science Technical Report TR84-029 %K AI10 %H TR84-030 %A Rajendra T. Dodhiawala %A George R. Cross %T A Distributed Problem-Solving Approach to Point Pattern Matching %R LSU Computer Science Technical Report TR84-030 %K AI06 %H TR84-034 %A W. G. Rudd %A George R. Cross %T Design of an Expert System for Insect Pest Management %R LSU Computer Science Technical Report TR84-034 %K AA23 AI01 %A George R. Cross %A Ellen R. Foxman %A Daniel L. Sherrell %T Using an Expert System to Teach Marketing Strategy %R LSU Computer Science Technical Report TR85-001 %K AI01 AA06 AA07 %H TR85-003 %A Cary G. deBessonet %A George R. Cross %T An Artificial Intelligence Application in the Law: CCLIPS, A Computer Program that Processes Legal Information %R LSU Computer Science Technical Report TR85-003 %K AA24 %H TR85-028 %A Sukhamay Kundu %T A Theory of Multi-Relations for Uncertain Facts %R LSU Computer Science Technical Report TR85-028 %K O04 %H TR85-032 %A Rajendra T. Dodhiawala %A George R. Cross %T Analysis of Cosmic Ray Tracks Using Distributed Problem-Solving %R LSU Computer Science Technical Report TR85-032 %K AI06 %H TR85-033 %A George R. Cross %A Cary G. deBessonet %A Teri Broemmelsiek %A Glynn Durham %A Rittick Gupta %A Mohd Nasiruddin %T The Implementation of CCLIPS %R LSU Computer Science Technical Report TR85-034 %H TR85-034 %A Mohd. Nasiruddin %A M. Srikanth %A George R. Cross %T A Confidence Factor Extension to the YAPS Expert System Development Tool %R LSU Computer Science Technical Report TR85-034 %K O04 T03 AI01 %H TR85-035 %A Cary G. deBessonet %A George R. Cross %T Some AI Techniques Used for Decision Making in Conceptual Retrieval %R LSU Computer Science Technical Report TR85-035 %K AI13 %H TR85-037 %A Cary G. deBessonet %A George R. Cross %T Distinguishing Legal Language-Types for Conceptual Retrieval %R LSU Computer Science Technical Report TR85-037 %K AA24 AA14 AI02 %H TR85-038 %A Zvieli Arie %T A Fuzzy Relational Calculus %R LSU Computer Science Technical Report TR85-038 %K O04 %A Eric Mjolness %T Neutral Networks, Pattern Recognition and Fingerprint Hallucination %R 5198:TR:85 %X 8.00 PHD THESIS %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %K AI06 AI12 %A B. H. Thompson %A Frederick B. Thompson %T Customizing One's Own Interface Uisng English as a Primary Language %R 5165:TR:84 %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %X $4.00 %K AI02 %A Remy Sanouillet %T ASK French - A French Natural Language Syntax %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 5164:TR:84 %K AI02 %X 13.00 Master's Thesis %A Michael Newton %T Combined Logical and Functional Programming Language %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 5172:TR;85 %K AI10 %X $6.00 %A Howard Derby %T Using Logic Programming for Compiling APL %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 5134:TR:84 %K AI10 %X $2.00 %A Bozena H. Thompson %T Linguistic Analysis of Natural Language Communication with Computers %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 5128:TM:84 %K AI02 %X $3.00 %A Bozena Thompson %A Fred Thompson %T ASK As Window to the World %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %X $3.00 %R 5114:TM:84 %K AI02 %A Alain Martin %T General Proof Rule for Procedures in Predicate Transformer Semantics %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 5075:TR:83 %K AI11 AA08 %X $2.00 %A David Trawick %T Robust Sentence Analysis and Habitability %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 5074:TR:83 %K AI02 %X $10.00 %A Bozena H. Thompson %A Frederick B. Thompson %T Introducing ASK, A Simple Knowledge System, Conference on App'l Natural Language Processing %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 5054:TM:82 %K AI02 %X $3.00 %A Bozena Thompson %A Frederick B. Thompson %A Tai-Ping Ho %T Knowledgeable Contexts for User Interfaction, Proc Nat'l Computer Conference %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 5051:TM:82 %K AI02 %X $2.00 %A Barry Megdal %T VLSI Computational Structures Applied to Fingerprint Image Analysis %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 5015:TR:82 %K AI06 %X $15.00 %A Charles R. Lang %T Concurrent, Asynchronous Garbage Collection Among Cooperating Processors %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 4724:TR:82 %K H03 %X $2.00 %A Sheue-Ling Lien %T Toward A Theorem Proving Architecture %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 4653:TR:81 %K AI11 %X $10.00 MS Thesis %A Leonid Rudin %T Lambda Logic %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 4521:TR:81 %K AI10 %X $8.00 MS THESIS %A Tzu-mu Lin %T From Geometry to Logic %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 4298:TR:81 %K AI10 AA13 %X $7.00 MS THESIS %A Jim Kajiya %T Toward A Mathematical Theory of Perception %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 4116:TR:79 %K AI08 AI06 %X $25.00 PHD THESIS %A Fred Thompson %A B. Thompson %T Shifting to a Higher Gear in a Natural Language System %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 4128:TM:81 %K AI02 %X $2.00 %A Bozena H. Thompson %A Frederick Thompson %T REL System and REL English, REL Report no. 22 %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 3999:TM:76 %K AI02 %X $3.00 %A B. H. Thompson %A Fred B. Thompson %T Rapidly Extendible Natural Language %I California Institute of Technology, Computer Science %C Pasadena, California 91125 %R 3975:TM:80 %K AI02 %X $3.00 %A Garrett M. Odell %A J. T. Bonner %T How the Dictyostelium Discoideum Grex Crawls %R 85-1 %I Computer Science Department, Rensselear Polytechnic Institute %K AI07 %A David L. Spooner %A Michael A. Milicia %A Donald B. Faatz %T Modeling Mechanical CAD Data With Data Abstraction and Object-Oriented Techniques %R 85-19 %I Computer Science Department, Rensselear Polytechnic Institute %K AA05 %A N. Prywes %A B. Szymanski %T Programming Supercomputers in an Equational Language %R 85-24 %I Computer Science Department, Rensselear Polytechnic Institute %K AI10 H04 %A N. Prywes %A Y. Shi %A J. Tseng %A B. Szymanski %T Supersystem Programming with the Model Equational Language %R 85-26 %I Computer Science Department, Rensselear Polytechnic Institute %K AI10 H04 %A Martin Hardwick %A Lin Kan %A Goutam Sinha %A Subhendu Lahiri %A Zia Mohammed %A Nisar Yakoob %T Design and Implementation of a Data Manager for Design Objects %R 85-34 %I Computer Science Department, Rensselear Polytechnic Institute %K AA05 %A D. Nagel %T Some Considerations on Extracting Definitional Information About Relations %R CBM-TM-85 %D APR 1980 %I Rutgers University, Department of Computer Science %K AI02 AA09 AI04 %X Several of the current systems in Artificial Intelligence are represented in binary relational databases and rely on the semantics of relations as a source of knowledge for information retrieval. Examples of these systems include those developed by Lindsay [5,6], Raphael [10], Elliott [2], Brown [1], and Sridharan [11]. In these systems inferences can be made from a set of properties specified for each relation. Inferences can also be made from specified associations between relations. One interesting aspect is the degree to which making these inferences can be automated. Some methods are proposed in this paper for using machine learning to extract relational properties and recognize semantic ties between relations so that this definitional information will not have to be prespecified. In some cases these methods may not technically be categorized as learning because they primarily involve summarization. It is also difficult to pin down what is encompassed by semantics. However, this paper discusses concepts of learning, and the presentation is directed at capturing semantics. .sp 1 Extracting definitional information or more broadly, learning semantics of relations, provides a base for the study of interesting databases. This could be done in a symbiotic system where the interaction between the researcher and the system provides a means for improving the performance of the system in general and obtaining new insights in the scientific data. It could also be coupled with a system for automatic theory formation. Presently, applications using semantics of relations for making inferences have been most successful in areas where properties and relationships are well understood such as kinship relations. ------------------------------ End of AIList Digest ******************** From vtcs1::in% Fri Apr 11 19:02:37 1986 Date: Fri, 11 Apr 86 19:02:31 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #79 Status: R AIList Digest Friday, 11 Apr 1986 Volume 4 : Issue 79 Today's Topics: Bibliography - Technical Reports #3 ---------------------------------------------------------------------- Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Technical Reports #3 %A N. S. Sridharan %T Representational Facilities of AIMDS: A Sampling %R CBM-TM-86 %D 1/82 %I Rutgers University, Department of Computer Science %K AI01 %X The quest for fundamental and general mechanisms of intelligence, especially problem solving and heuristic search techniques, that guided early research in Artificial Intelligence has given way in the last decade to the search for equally fundamental and general methods for structuring and representing knowledge. This is the result of the realization that a duality exists between knowledge and search: Knowledge of the task domain can abbreviate search and search thru a problem space can yield new knowledge. AIMDS is one of the recently developed systems which permits experimentation with knowledge representation in the course of building an AI program. %A C. F. Schmidt %T The Role of Object Knowledge in Human Planning %R CBM-TM-87 %I Rutgers University Department of Computer Science %K AI08 AI09 %D 1/82 %X AI research on planning provides an important reference point from which the cognitive psychologist can build an understanding of human planning. It is argued that the human planning context differs from this reference point due to the incomplete knowledge that persons typically possess about the situation within which the plan will be executed. Various types of general functional knowledge about objects are then defined. This knowledge serves as a source of default assumptions for use in the planning process, and thus allows planning to continue despite the absence of complete knowledge of the planning situation. However, such assumption-based expectations must be tested. From this point of view, planning must also include a process for a kind of hypothesis testing and plan revision. The implications of this claim are briefly discussed. %A S. Amarel %T Initial Thoughts on Characterization of Expert Systems %R CBM-TM-88 %I Rutgers University Department of Computer Science %D 1/82 %K AI01 %X Expertise in a given domain is commonly characterized by skillful, high performance, problem solving activity in the domain. An expert solves problems in a domain more rapidly, more accurately, and with less conscious deliberation about his plan of attack than a novice does. An excellent discussion of general characteristics of expert behavior appears in a recent article in @u(Science) by Larkin et al. [1]. .sp 1 Expert behavior is equivalent to high performance problem solving behavior in a specific domain. It requires: knowledge of the domain, knowledge of problem solving schemas and methods, knowledge/experience about solution of specific problems in the domain with given methods, knowledge about special properties and regularities in the problem space, and highly effective ways of @u(using) all these bodies of knowledge in approaching the solution of new problems in the domain. Essentially, expert problem solving requires the conceptualization/formulation of a given problem within a framework wherein knowledge is embodied in definitions of states, moves, constraints, evaluation functions, etc. in such a way that solutions are attained with very little search. In other words, an expert problem solver works within a highly 'appropriate' problem representation: he describes situations and problem types within 'appropriate' conceptual frameworks, he specifies problem decompositions that minimize subproblem interactions, he often uses hierarchies of abstractions in his planning, he uses 'macromoves' where a novice would painstakingly have to piece together elementary moves, and he has rules for early recognition of unpromising as well as of promising developments. An expert problem solver behaves as if the great variety of knowledge sources needed for his solution-construction activities are available to him in a @u(compiled) procedural form. .sp 1 Usually, expertise in a domain requires @u(problem solving experience) in the domain. One can be scholar in a domain, and not an expert--if he does not know how to effectively @u(apply) domain knowledge to a variety of specific situations. Also, expertise implies a certain amount of robustness in performance-- which means that it is not sufficient to know how to handle a few 'textbook' cases; it is important to be able to handle a broad range of variations. %A S. Amarel %T Review of Characteristics of Current Expert Systems %R CBM-TM-89 %I Rutgers University Department of Computer Science %D 3/81 %K AI01 %X This report does not cover all current work in the area of Expert systems. It is intended to introduce a set of dimensions for characterizing Expert systems and to describe some of the important Expert systems that are now in existence (or are under active development) in terms of these dimensions. .sp 1 We have a dual purpose: (a) to illustrate via concrete examples the dimensions that are being introduced, and (b) to show what is the current state of the field from the perspective of this system of dimensions. .sp 1 We are using here ten main dimensions, and an optional eleventh called @ux(Special Features), which provides added flexibility for the presentation of relevant information about a system. Two of the main dimensions, @ux(Performance) and @ux(Utility), are concerned with the quality of the system's behavior and the impact of the system on the domain of application and on AI. Another two dimensions are concerned with the system's scope, its ability to handle situations that are outside its area of major expertise, and its ability to improve: they are called @u(Breadth, Intelligence, Robustness) and @u(Expertise improvement ability). The remaining six dimensions are concerned with the type of tasks performed by the system, its structure and its means of interacting with users: they are called @u(Task type, Main Method, Mode of Knowledge Representation, User Interface for main task, Explanation facilities) and @u(Reasoning under Uncertainty). .sp 1 The systems considered are DENDRAL, CASNET/GLAUCOMA, MACSYMA, MYCIN, INTERNIST, PROSPECTOR and CRYSALIS. .sp 1 This report covers material which was prepared for inclusion in the Chapter 'What are Expert Systems' (co-authored with Ron Brachman, Carl Engelman, Robert Engelmore, Edward Feigenbaum and David Wilkins) of a book on Expert Systems which is currently under preparation; the book is based on the Rand Workshop on Expert Systems which took place in San Diego, California on August 25-28, 1980. %A John Kastner %A Sholom M. Weiss %T A Precedence Scheme for Selection and Explanation of Therapies %R CB-TM-90 %I Rutgers University Department of Computer Science %D 3/81 %K AA01 AI01 %X A general scheme to aid in the selection of therapies is described. A topological sorting procedure within a general production rule representation is introduced. The procedure is used to choose among competing therapies on the basis of precedence rules. This approach has a degree of naturalness that lends itself to automatic explanation of the choices made. A system has been implemented using this approach to develop an expert system for planning therapies for patients diagnosed as having ocular herpes simples. An abstracted example of the system's output on an actual case is given. %A P. Politakis %A S. M. weiss %T A System for Empherical Experimentation with Expert Knwoledge %R CBM-TM-91 %I Rutgers University, Department of Computer Science %D 1/82 %K AI01 AA01 rheumatology %X An approach to the acquisition of expert knowledge is presented based on the comparison of dual sources of knowledge: expert-modeled rules and cases with known conclusions. A system called SEEK has been implemented to give to the expert interactive advice about rule refinement. SEEK uses a simple frame model for expressing expert-modeled rules. The advice takes the form of suggestions of possible experiments in generalizing or specializing rules in the model. This approach has proven particularly valuable in assisting the expert in domains where two diagnoses are difficult to distinguish. Examples are given from an expert consultation system being developed for rheumatology. %A G. Drastal %A C. Kulikowski %T Knowledge Based Acquisition of Rules for Medical Diagnosis %R CBM-TM-92 %I Rutgers University, Department of Computer Science %D 10/81 %K AA01 AI01 %X Medical consultation systems in the EXPERT framework contain rules written under the guidance of expert physicians. We present a methodology and preliminary implementation of a system which learns compiled rule chains from positive case examples of a diagnostic class and negative examples of alternative diagnostic classes. Rule acquisition is guided by the constraints of physiological process models represented in the system. Evaluation of the system is proceeding in the area of glaucoma diagnosis, and an example of an experiment in this domain is included. %A N. S. Sridharan %T AIMDS: Applications and Performance Enhancements %R CBM-TM-93 %I Rutgers University, Department of Computer Science %D 1/82 %K AI01 AA24 %X AIMDS is a programming environment (language, editors, display drivers, file system) in which several programs are being constructed for modeling commonsense reasoning and legal argumentation. The main obstacle to realistic applications in these and other areas is system performance when the knowledge bases used are scaled up one or two orders of magnitude. The other obstacle is user performance resulting from the complexity of constructing and debugging large scale knowledge bases. This proposal argues that performance enhancement of AIMDS as a system is needed and that the usual solutions of software tuning have been exhausted and that new hardware ideas fitted to the characteristics of the task need to be experimented with. We adopt as important constraints: the requirement that existing programs should receive graded enhancement of performance, maintaining continuity of application programs; that user programs should not reflect changing machine configurations or architectures. Redesign and recoding of AIMDS should provide the necessary opacity to the user. With these constraints in mind, we suggest interim solutions and long-term solutions. The interim solutions include: converting large-address space personal Lisp machines with bit-mapped graphics; fast coding of low-level functionalities via microprogramming. The long-term solutions include the building and testing of multiprocessors. The long-term solutions open up a number of rather difficult software and hardware research problems whose solutions depend upon having good facilities to experiment in the search for answers. %A B. Lantz %T The AIMDS Interactive Command Parser %R CBM-TM-94 %I Rutgers University, Department of Computer Science %D 9/82 %K AA24 AI01 T03 %X Characters entered by the user are parsed immediately in order to provide interactive services to the user while he is entering commands. Services provided to the user include immediate verification of syntax, supplying the user with information about the correct syntax and semantics of a command, completion of long descriptive atom names, pretty printing the entered command, and defining special functions for selected characters. The parser accepts user defined grammars, thus providing a useful command parser for a great variety of applications. %A B. Lantz %T The AIMDS On-Line Documentation Facility %R CBM-TM-95 %I Rutgers University, Department of Computer Science %D 9/82 %K AA24 AI01 T03 %X The documentation system for the AIMDS language is designed to be suitable for both beginning and expert users, and to be capable of serving the needs of a changing system such as AIMDS. The documentation must be quickly and easily updatable, and the updated information should be available to, and easily used by, a wide variety of users. %X This paper is a short description of the documentation system for the AIMDS language. It includes a discussion of the considerations taken during the design of the documentation system, a description of the implemented system, and instructions for using the system for other documentation tasks. %A J. Roach %A N. S. Sridharan %T Implementing AIMDS on a Multiprocessor Machine. Some Considerations %R CBM-TM-96 %D 4/83 %I Rutgers University, Department of Computer Science %K AA24 T03 H03 AI01 %X As a possible long term solution for performance enhancement of AIMDS, a Lisp based multiprocessor system was proposed. Converting an existing AI knowledge based system from the current uniprocessor environment into a multiprocessor based regime is a largely unexplored research question. This report discusses some of the issues raised by such a proposal and attempts to evaluate some of the current models of parallel processing in regards to implementing an AIMDS based system. An extensive bibliography with commentary is included. %A George A. Drastal %A Casimir A. Kulikowski %T Knowledge-Based Acquisition of Rule for Medical Diagnosis %R CBM-TM-97 %I Rutgers University, Department of Computer Science %D 11/82 %K AI01 AA01 T03 %X Medical consultation systems in the EXPERT framework contain rules written under the guidance of expert physicians. We present a methodology and preliminary implementation of a system that learns compiled rule chains from positive case examples of a diagnostic class and negative examples of alternative diagnostic classes. Rule acquisition is guided by the constraints of physiological process models represented in the system. Evaluation of the system is proceeding in the area of glaucoma diagnosis, and an example of an experiment in this domain is included. %A S. Weiss %A K. Kern %A C. Kulikowski %A M. Uschold %T A Guide to the Use of the EXPERT Consultation System %R CBM-TR-94 %I Rutgers University, Department of Computer Science %D 1/82 %K T03 AI01 %X EXPERT is a system for designing and applying consultation models. An EXPERT model consists of hypotheses (conclusions), findings (observations), and rules for logically relating findings to hypotheses. Three phases of model development are outlined for users of the system. These include: the design of a decision-making model, compilation of the model, and consultation using the model. The facilities of the system are described, and examples of models and consultation sessions are presented. %A R. Banerji %A T. Mitchell %T Description Languages and Learning Algorithms: A Paradigm for Comparison %R CBM-TR-107 %D 1/82 %I Rutgers University, Department of Computer Science %K AI04 Inductive inference, learning, generalization, description languages. %X We propose and apply a framework for comparing various methods for learning descriptions of classes of objects given a set of training exemplars. Such systems may be usefully characterized in terms of their descriptive languages, and the learning algorithms they employ. The basis for our characterization and comparison is a general-to-specific partial ordering over the description language, which allows characterizing learning algorithms independent of the description language with which they are associated. Two existing learning systems are characterized within this framework, and correspondences between them made clear. ------------------------------ End of AIList Digest ******************** From vtcs1::in% Sat Apr 12 06:42:57 1986 Date: Sat, 12 Apr 86 06:42:52 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #80 Status: R AIList Digest Saturday, 12 Apr 1986 Volume 4 : Issue 80 Today's Topics: Queries - Shape & LOOPS on a XEROX, Application - Automatic Documentation, Policy - Press Releases, Journal - AI Expert, Review - Spang Robinson Report, Volume 2 No. 4, Philosophy - Lucas on AI & Computer Consciousness ---------------------------------------------------------------------- Date: Thu 10 Apr 86 09:52:08-PST From: Ken Laws Subject: Shape Jerry Hobbs has asked me "What is a hook and what is a ring that we know the ring can hang on the hook?" More specifically, what do we have to know about hooks and rings in general (for default reasoning) and about a particular hook-like object and ring-like object (dimensions, radius of curvature, surface normals, clearances, tolerances, etc.) in order to say whether a particular ring may be placed on a particular hook and whether it is likely to stay in place once put there. Can we reason about the functionality of shapes (in this and in other "mechanics" problems) without resorting to full CAD/CAM descriptions, physics, and simulation? How do people (e.g., children) reason about shape, particularly in the intuitively obvious cases where tolerances are not critical? Can anyone suggest a good lead? -- Ken Laws ------------------------------ Date: Fri, 11 Apr 86 15:03 CST From: Brick Verser Subject: LOOPS running on a XEROX Does anybody have any information pertaining to applications running under Loops on Xerox hardware? ------------------------------ Date: Thu, 10 Apr 86 20:30:50 pst From: saber!matt@SUN.COM (Matt Perez) Subject: Re: towards better documentation > > I am interested in creating an expert system to serve as on-line > documentation. The intent is to abrogate the above law and > corollaries. Does anyone know of such a system or any effort(s) to > produce one? > Contact Mark Miller of Computer*Thought, in Dallas, Texas. They may have what you are looking for. Mark is a pretty friendly guy and may also point you to the right literature, etc. The other place I can think of where there's something like this in development is the work being done by Prof. Wilensky at UCBerkely: The Unix Consultant. Matt Perez ------------------------------ Date: Thu 10 Apr 86 09:03:06-PST From: GARVEY@SRI-AI.ARPA Subject: Re: Policy - Press Releases Eschew Policy!! Let Ken handle it; if you don't like what he let's through, don't read it (I have ^O on my terminal for just such situations). Nobody wastes your time but you.... (And maybe me.) Cheers, Tom [Unfortunately ^O doesn't work for those reading the "unexploded" digest format. Most mail programs haven't adapted to the digest formats yet. -- KIL] ------------------------------ Date: Thu 10 Apr 86 11:17:53-PST From: Ken Laws Subject: AI Expert The April issue of CACM has an ad (p. A-31) for AI Expert, a new journal for AI programmers. "No hype, no star wars nonsense, no pipe dreams. AI Expert will focus on practical AI programming methods and applications." AI Expert, 2443 Fillmore Street, Suite 500, San Francisco, CA 94115; $27 for 13 issues, $33 Canada, $39 worldwide, $57 airmail. ------------------------------ Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Spang Robinson Report, Volume 2 No. 4 Summary Summary of The Spang Robinson Report, Volume 2 No. 4 April 1986 Packaging Financial Expertise Activities at specific companies and available products: Applied Expert Systems (APEX) Plan Power: Expert system for personal financial planning. (less than $50K including Xerox 1186 on which to run on) Arthur D. Little: Personal Financial Planning System (in test), Equity Trader's Assistant, Cash Trader's Assistant, insurance personnel selection system, investment manager's work station, bond indenture advice system. (in development) (will run on Symbolics 3670 and use databases residing on IBM mainframe) Cognitive Systems: Courtier, a stock portfolio management system. One version is design for individual use at public terminals with another to assist bank portfolio managers. Runs on Apollo's and DEC VAX. Human Edge Software: is supporting development of Financial Statement Analysis, an expert business evaluation program, and a busines plan expert for IBM PC's. Palladian Software: Financial Advisor which is designed to help with corporate financial decison-making and project evaluation capabilities. It is based upon net present value. Prophecy Development Corporation: Profit tool, a brokerage and financial services shell. Runs on MS-DOS computers and costs $1995.00. Sterling Wentworth Corporation: Planman, $4500.00 Database $2000.00 For CPA's to produce "comprehensive financial planning reports" Runs on MSDOS and they have sold 400 units Syntelligence: Underwriting Advisor System Lending Advisor System. Delivered on IBM 30 and 43 series with connections to PC/AT's. Nikko Securities Co and Fujitsu: (under development) a system for selecting stocks for investment. Daiwa Securities: Placed a system to provide investment councelling into operation last month. Yamaichi Securities: developing AI based investment products in collaboration with Hitachi and Nippon Univak. Nomura Securities: This is Japan's largest stock broker and they embarking on a broad- based AI R&D program. The total revenue for financial expert systems is five million dollars. In 1985, financial applications were five percent of all epxert systems. In 1986, it is expected to be twenty percent. One in five of large financial institutions have applied expert systems. __________________________________________________________________________ Japan Watch: Nomura research Institute has developed a tool called DORSAI for assisting in the production of expert systems in PROLOG. Hitachi's Energy Research Institute have developed a system for proving theorems at a high speed. Hitachi applied for a patent and it employs the "Connection Graph technique." Hitachi will use this system for VLSI logic design, factory automation, real-time failure diagnostics, chemical compound synthesis, and hardware applications. 60 percent of Japanese corporations are beginning to utilize AI or are studying such a move. 28 percent of Japanese hardware, software companies and heavy computer users have plans to enter into AI while 32 percent are currently involv ed in AI. 52 percent of the companies with plans to enter AI expressed an interest in expert systems. __________________________________________________________________________ Micro Trends: Discussion of Borlands's Turbo prolog including reactions from Arity, Quintus Prolog, GOld Hill, CIGNA. __________________________________________________________________________ News: Amoco Corporation and IntelliCorp announced a joint venture to market AI products for molecular biology. The first new product will be Strategene. Boeing Computer Services and Carnegie Federal Systems Corproation will be working together on a Rome Air Development center contract to develop a "new engineering environment." Carnegie Federal Systems will support TRW in developing AI software for tactical mission planning and resource allocation functions. Qunintus Computer Systems has over 270 users with 170 of them using Quintus on the workstaiton. Revenues for 1985 were $2.1 million dollars with an 18 percent profit margin. Rapport, a DBMS may soon be available for Symbolics machines. It will run not only in single user mode but as a multi-user file server. UC Berkeley is developing a RISC based LISP machine with multiple processors (SPUR project). The Aquarius project involves using separate processors for numeric, Prolog and LISP processing, each optimized for its specific rule. Frank Spitznogle who formerly was President and chief operating officer of Lisp Machines is now President and chief operating officer of Knowledge Systems of Houston Texas. They will be applying AI to oil and gas industry. Their first product will be an exploration-potential evaluation consultant. Thomas Kehler is now chairman and CEO of Intellicorp. ------------------------------ Date: Thu, 10 Apr 86 11:33:07 est From: John McLean Subject: Lucas on AI > From: Stanley Letovsky > At the conference on "AI an the Human Mind" held at Yale early in > March 1986, a paper was presented by the British mathematician John > Lucas. He claimed that AI could never succeed, that a machine was in > principle incapable of doing all that a mind can do. His argument went > like this. Any computing machine is essentially equivalent to a system > of formal logic. The famous Godel incompleteness theorem shows that for > any formal system powerful enough to be interesting, there are truths > which cannot be proved in that system. Since a person can see and > recognize these truths, the person can transcend the limitations of the > formal system. Since this is true of any formal system at all, a person > can always transcend a formal system, therefore a formal system can > never be a model of a person. Stanley Letovsky tries to refute this argument by showing that a formal description that describes Lucas' beliefs may have unprovable assertions that Lucas nevertheless believes. > What is critical to realize, however, is that the Godel sentence > for our model of Lucas is not a belief of Lucas' according to the model. > The form of the Godel sentence > G: not(provable(G)) > is syntactically distinct from the form of an assertion about Lucas' > beliefs, > believes(p,t) > Nothing stops us from having > believes(G,t) > be provable in the system, despite the fact that G is not itself > provable in the system. This view of what G must look like is too restrictive. Note that the Godel sentence for a system of first order arithmetic is an assertion in number theory ("There is no integer such that..."). The fact that the assertion numeralwise represents an assertion about provability takes a great deal of showing. Similarly, the Godel sentence for our model of Lucas may be an assertion about what Lucas believes. Since Lucas is going to have beliefs about his beliefs and what is provable in the system, it's not hard to believe that we can construct a self-referential sentence G such that Lucas believes G at t but believe(G,t) is not a theorem. This is particularly plausible since there is a strong connection between what Lucas believes and what is provable in the system. In particular, believes(believes(x,y),t) will be provable iff Lucas believes that he believes x at y. But it is plausible to assume that Lucas believes that he believes x at y iff he believes x at y, i. e., iff believes(x,y) is provable. In other words, the belief predicate is a provabilility predicate for the system restricted to statements about beliefs. To fill this out, note that we will probably have that if believes("not(believes(x,y))",t) then not(believes(x,y)) since if Lucas believes that he doesn't believe p, then he doesn't believe p. Now consider G: believes("not(G,t)",t). If our system is consistent and such a G exists, G is not provable. If G were provable, then not(G) would also be provable given our observation since G is a statement about belief. I believe that it is possible to construct such a sentence G, but this does not imply that we can't dismiss Lucas. Lucas' argument is unconvincing since there is no reason to believe that for any formal system, I can see and recognize the Godel sentence for that system. Godel sentences for a particular system are long and complicated. Hence, there is no reason to believe that Lucas surpasses every formal system. In fact, it is clear that there is at least one formal system that can recognize as true a sentence that Lucas can't. Consider the system that contains one axiom: "is a sentence that Lucas will never recognize as true when appended to its own quotation" is a sentence that Lucas will never recognize as true when appended to its own quotation. The system recognizes the sentence as true since it's an axiom; Lucas doesn't. John McLean mclean@nrl-css ...!decvax!nrl-css!mclean ------------------------------ Date: Thu, 10 Apr 86 11:19:13 EST From: tes%bostonu.csnet@CSNET-RELAY.ARPA Subject: Computer Consciousness Informal talk on computer consciousness: The whole family of questions like "Can computers feel emotion?" and "Is it possible for a computer to be conscious?" define a loaded, emotionally-charged subject. Some people (especially "artistic" folks, in my experience) give an immediate emphatic "NO!" to these questions; other people (many Science-is-The-Answer-to-Everything sorts) devise computational models that parallel what we know about physical brain structure and conclude "yes, of course"; and other folks remain somewhere in the middle or profess "it's too complicated - I don't know." My main beef with some physicalist or reductionist opinions is the *assumption* that nothing except physical events exist in the universe, and that a physical or functional description of a system describes its essence entirely, and therefore if the human brain's neural interactions are simulated by some machine then this machine is for all intents and purposes equivalent to a human mind. To me, the phenomenological red that I perceive when looking at an apple is OBVIOUSLY real, as is my consciousness. It is ridiculous to conclude that consciousness and phenomenological experiences do not exist simply because they cannot be easily described with mathematics or the English language. My main beef with immediate emphatic "NO"s is that it may reflect an emotional fear of examining "taboo" territories, perhaps because such inquiry threatens the Meaning of Life or the sovereignty of the human mind.(There is no need to expound on how much suffering this attitude has brought upon our ancestors throughout history.) To find out that the human mind is "just this" or "just that" would significantly alter certain worldviews. The possibilites that are left to me are either that 1) Consciousness "emerges" from the functionality of the brain's neural interactions (if this is true, then it would be entirely possible, in principle, for a computer program with the same functionality to generate consciousness), 2) There is a dualism of the mental and the physical with mysterious interactions between the two realms, and 3) Other possibilities which no one has thought of yet. Now the first two may seem ridiculous, and I have no idea how to prove or disprove them, but they remain *possibilities* for me because they are not yet disproven. The physicalist proposal, on the other hand, is proven wrong (or rather its absolute universality is proven wrong) by the simplest introspective observation. I am not campaigning for a ceasing of all brain research or cognitive science; these sciences will continue to yield useful information. But I hope that these researchers and their fans do not delude themselves into thinking that the only aspect of the universe which exists is the aspect that science can deal with. Tom Schutz CSNET: tes@bu-cs ARPA: tes%bu-cs@csnet-relay UUCP: ...harvard!bu-cs!tes ------------------------------ End of AIList Digest ******************** From vtcs1::in% Sat Apr 12 06:43:07 1986 Date: Sat, 12 Apr 86 06:43:03 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #81 Status: R AIList Digest Saturday, 12 Apr 1986 Volume 4 : Issue 81 Today's Topics: Bibliography - Technical Reports #4 ---------------------------------------------------------------------- Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Technical Reports #4 %R CBM-TR-109 %T The Role of World Knowledge in Planning %A N.S. Sridharan %A C.F. Schmidt %A J.L. Goodson %D 1/82 %I Rutgers University, Department of Computer Science %K AI09 common-sense %X Common-sense planning demands a rich variety of world knowledge. We have examined here the view that world knowledge can be structured to form the interface between a hierarchy of action types and a hierarchy of types of objects. World knowledge forming this interface includes not only the traditional statements about preconditions and outcomes of actions, but also the normal states of objects participating in the actions and normative actions associated with the objects. Common-sense plans are decomposed into goal-directed, preparation, and the normative components. This has heuristic value and may serve to simplify the planning algorithm. The algorithm invokes world knowledge for goal customization, action specification, computation of preconditions and outcomes, object selection, and for setting up subgoals. %R CBM-TR-110 %I Rutgers University, Department of Computer Science %D 5/80 %T An Experimental Transformation of a Large Expert Knowledge %A R.N. Goldberg %A S.M. Weiss %K internist AI01 AA01 %X An experiment is described in which a significant part of the INTERNIST knowledge base for diagnosis in internal medicine is translated into an EXPERT model. INTERNIST employs the largest and broadest knowledge base of all the medical consultation systems which have been developed in recent years. EXPERT is a general system for designing consultation models. The translated model shows reasonable competence in the final diagnostic classification of 431 test cases. There are differences in the internal representation and reasoning strategies of the two systems. However, when a knowledge base has been encoded in a relatively uniform manner, this experiment demonstrates the feasibility of transfer of knowledge between large-scale expert systems. %R CBM-TR-111 %I Rutgers University, Department of Computer Science %D 6/80 %T A Process for Evaluating Tree-Consistency %A J.L. Goodson %X General knowledge about conceptual classes represented in a concept hierarchy can provide a basis for various types of inferences about an individual. However, the various sources of inference may not lead to a consistent set of conclusions about the individual. This paper provides a brief glimpse at how we represent beliefs about specific individuals and conceptual knowledge, discusses some of the sources of inference we have defined, and describes procedures and structures that can be used to evaluate agreement among sources whose conclusions can be viewed as advocating various values in a tree partition of alternate values. %R CBM-TR-112 %I Rutgers University, Department of Computer Science %D 9/80 %T "A Methodology for the Construction of Natural Language Front Ends for Medical Consultation System %A V. Ciesielski %D 1/82 %K AI01 AI02 AA01 %X A methodology for constructing natural language front ends for Associational Knowledge type (AK-type) medical consultation systems is described. AK-type consultation systems use associational knowledge of the form "if A and B and C then conclude D with a weight of w" to perform diagnostic reasoning. It is shown that the knowledge needed to "understand" patient description is not the associational knowledge in the consultation system but rather knowledge of structural relations and the way they are expressed in surface language. The two main structural relations involved are: (1) ATTRIBUTE of OBJECT = VALUE. Surface forms of this relation are variants and augmentations of the template "The X of Y is V". (2) OBJECT have-component COMPONENT. Surface forms of this relation are variants and augmentations of the template "The X has/contains/includes Y". This kind of knowledge can be represented in the Attribute-Component/Structured Object (AC/SO) package which was developed as part of this research. The AC/SO package is given a definition of the @u(concept) "PATIENT" for a disease area and the corresponding lexicon. %R DCS-TR-118 %I Rutgers University, Department of Computer Science %D 9/82 %T Transformational Programming--Applications to Algorithms and Systems %A Robert Paige %K AA08 %X Transformational programming is a nascent software development methodology that promises to reduce programming labor, increase program reliability, and improve program performance. Our research centers around a prototype transformational programming system called RAPTS (Rutgers Abstract Program Transformation System), developed during the past several years at Laboratory for Computer Science Research. Experiments in RAPTS with algorithm derivations are expected to lead to pragmatic applications to algorithm design, program development, and large system construction. %R DCS-TR-115 %I Rutgers University, Department of Computer Science %D 4/82 %T A Survey of Research in Strategy Acquisition %A R. Keller %D 7/82 %X This paper surveys literature in the area of strategy acquisition for artificial and human problem solving systems. A unifying view of the term "strategy" is suggested which places strategies along a continuum from abstract to concrete. Major concerns of strategy acquisition research are described, including (i) strategic component learning, (ii) strategy applicability recognition, (iii) strategy customization and (iv) strategy transformation. Various researchers' approaches to these issues are reviewed and open problems are discussed. %R DCS-TR-114 %I Rutgers University, Department of Computer Science %D 3/82 %T The Control of Inferencing in Natural Language Understanding %A Abe Lockman %A David Klappholz %K AI02 %X The understanding of a natural language text requires that a reader (human or computer program) be able to resolve ambiguities at the syntactic and lexical levels; it also requires that a reader be able to recover that part of the meaning of a text which is over and above the collection of meanings of its individual sentences taken in isolation. %X The satisfaction of this requirement involves complex inferencing from a large database of world-knowledge. While human readers seem able to perform this task easily, the designer of computer programs for natural language understanding faces the serious difficulty of algorithmically defining precisely the items of world-knowledge required at any point in the processing, i.e., the problem of @i[controlling inferencing]. This paper discusses the problems involved in such control of inferencing; an approach to their solution is presented, based on the notion of determining where each successive sentence "fits" into the text as a whole. %R DCS-TR-113 %I Rutgers University, Department of Computer Science %D 4/82 %T Consistent-Labeling Problems and Their Algorithms: Part II %A B. Nudelo %D 10/82 %K AI14 AI10 AI03 inter-variable compatibility %X A new parameter is introduced to characterize a type of search problem of broad relevance in Artificial Intelligence, Operations Research and Symbolic Logic. This paramater, which we call inter-variable @b[compatibility] is particularly important in that complexity analyses incorporating it are able to capture the dependence of problem complexity on search order used by an algorithm. Thus compatibility-based theories can provide a theoretical basis for the extraction of heuristics for choosing good search orderings - a long-sought goal for such problems, since it can lead to significant savings during search. We carry out expected complexity analyses for the traditional Backtrack algorithm as well as for two more recent algorithms that have been found empirically to be significant improvements, Forward Checking and word-wise Forward Checking. We extract compatibility-based ordering-heuristics from the theory for Forward Checking. Preliminary experimental results are presented showing the large savings that result from their use. Similar savings can be expected for other algorithms when heuristics taking account of inter-variable compatibilities are used. Our compatibility-based theories also provide a more precise way of predicting which algorithm is best for a given problem. %A B. Nudel %T Understand Consistent-Labelling Problems and Their Algorithms and Their Algorithms: Part I %R DCS-TR-112 %D (forthcoming) %I Rutgers University, Department of Computer Science %K AI14 AI10 AI03 %R DCS-TR-109 %I Rutgers University, Department of Computer Science %D 12/81 %T Equations - The "Improved Constraint Satisfaction Algorithms using Inter-Variable Compatibilities" %A B. Nudel %K consistent labeling AI03 %X This report addresses the problem of improving algorithms for solving @b(consistent-labeling) (also called @b(constraint-satisfaction) problems. The concept of @b(compatibility) between variables in such problems is introduced. How to obtain compatibilities analytically and empirically is discussed, and various compatibility-based heuristics (as well as some useful but less effective non compatibility-based heuristics) are developed to improve a version of the Waltz algorithm which was found best of a set of consistent-labeling problem algorithms tested by Haralick [5]. Empirical results with these heuristics are very encouraging, with over an order of magnitude improvement in performance with respect to the basic algorithm on a set of randomly generated consistent-labeling problems. %R DCS-TR-107 %D 10/81 %I Rutgers University, Department of Computer Science %T Note on Learning in MDS Based on Predicate Signatures %A C.V. Srinivasan %K AI06 AI04 %X This note illustrates a simple learning scheme in the context of two examples. In the first example the system learns the distinguishing features of the letters in the English alphabet, where each letter is described in terms of a relational system of features. In the second example the domain is a set of family relationships. In this case the system identifies invariant properties like "father.father=grandfather" that exist in the domain. .sp 1 In both examples the system first creates an abstraction of the given set of relations and uses the abstraction to identify invariant (or distinguishing) features of the given set of relations. The abstraction scheme is based on the concept of "predicate signatures" that is described in the note. .sp 1 The method is a general one. It can be used to identify large classes of invariant (or distinguishing) features of sets of objects where each object is described in terms of a set of relations that hold true for the object. %R DCS-TR-106 %I Rutgers University, Department of Computer Science %D 10/81 %T Knowledge Representation and Problem Solving in MDS %A C. V. Srinivasan %K AI11 %X This work presents a new approach for using a first order theory to generate procedures for solving goal satisfaction problems without using general theorem proving. The core of the problem solving system has three basic components: an inferencing mechanism based on @u(residues), a control structure for "means-end" analysis that uses @u(natural deduction), and a generalization scheme that is based on the structure of statements in the domain theory itself. .sp 1 The work represents a beginning in the development of knowledge based systems that can generate their own problem solving programs, evolve with experience and adapt to a changing domain theory. %R DCS-TR-95 %D 10/80 %I Rutgers University, Department of Computer Science %T A Mini-Max Problem %A W.L. Steiger %K AI03 %X Determine an algorithm, better than complete enumeration, for the following problem: given a non-negative integer matrix, permute the entries in each column independently so as to minimize the largest row sum. This problem had arisen in determining an optimal scheduling for a factory work force. %R DCS-TR-92 %D 4/80 %T Average Case Behavior of the Alpha-Beta Tree Pruning Algorithm %A George Shrier %D 1/82 %I Rutgers University, Department of Computer Science %K AI03 %R DCS-TM-15 %I Rutgers University, Department of Computer Science %D 3/81 %T Some Experiments in Abstraction of Relational Characteristics %A R.M. Keller %A D.J. Nagel %K AA09 AI01 AI04 %X Two experiments performed in knowledge-based inference are discussed in this paper. The experiments are directed at abstracting structural regularities and patterns inherent in a database of binary relations. A novel graph representation to facilitate abstraction is used in approaching some classical problem areas. This representation is compact and powerful, and an efficient algorithm has been developed to help control the exhaustive nature of certain types of inductive problems. .sp 1 One area of experimentation concerns the discovery of intensionally definable re lations in a family database. Another is the recognition of alphabetic characters using directional relations defined for points on a grid. Within a test bed system, KBLS, a scheme for computing abstractions is briefly summarized, and implications for future extensions are discussed in light of experimental results. %R DCS-TM-16 %I Rutgers University, Department of Computer Science %D 3/83 %T Solving the Plane Geometry Problem by Learning %A Liben Xu %K AI01 AA13 AI14 %X The top-down technique for solving a geometry problem is described. The top-down method uses "general rules," they are obtained by learning. This report focuses on general heuristics to obtain the general rules for solving a geometry problem. %R DCS-TR-89 %D 5/80 %I Rutgers University, Department of Computer Science %T Parts I, II, III of KNOWLEDGE BASED LEARNING SYSTEMS DS + CVS = A Proposal for Research CVS = An Intro. to the Meta-Theory & Logical Foundations %A D. Sandford %K AI01 AI04 %X Current state of the art experience in designing domain specific, intelligent, automated problem solving systems argues convincingly that: Firstly, large amounts of what is known as domain dependent or domain specific knowledge is crucial to achieving acceptable efficiency in realistic problem solving situations; and secondly that the task of implementing such systems "from scratch" is such a formidable one that it has impeded experimental research into the nature and role of domain specific knowledge in problem solving. .sp 1 This project is directed towards attaining an understanding of the processes and types of organizations required for an automated system to be able to learn for itself the relevant domain dependent knowledge from its experience with the domain. The research is based on a meta-theory of knowledge based learning systems, systems that can discover domain knowledge and use it to solve problems in a domain. The research project will employ both experimentation with implemented systems and theoretical analysis of systems. The goals are to shed light on both the detailed mechanisms by which domain dependent knowledge increases search efficiency, and to understand the type of innate biases that an automated system needs, to be able to analyze a domain and discover the appropriate domain knowledge. The research is based on a meta-theory of systems that are both knowledge based systems and learning systems. .sp 1 The research focuses on two kinds of systems: Systems that can build and use empirical theories of domains, and systems that use Axiomatic theories and theorem proving. The nature of domain knowledge and ways of using it in both these systems are investigated. ------------------------------ End of AIList Digest ******************** From vtcs1::in% Sun Apr 13 00:44:26 1986 Date: Sun, 13 Apr 86 00:44:22 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #82 Status: R AIList Digest Saturday, 12 Apr 1986 Volume 4 : Issue 82 Today's Topics: Bibliography - Technical Reports #5 ---------------------------------------------------------------------- Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Technical Reports #5 %R DCS-TR-90 %I Rutgers University, Department of Computer Science %D 5/80 %T Knowledge-based learning, an Example %A C. V. Srinivasan %D 1/82 %K AI04 %X How may a machine "learn" from examples of situations that are presented to it? What may constitute the "knowledge" of a set of such situations? How should the examples be presented to the machine? Are there general principles which a machine can use to acquire the knowledge automatically by examining the examples presented to it, and to use the knowledge so obtained to solve problems in a domain? These are the general concerns of my research. %R CBM-TR-138 %D 5/84 %T Hardware Fault Diagnosis & Expert Systems %A Allen Ginsberg %D 5/84 %K AI01 AA04 AA21 %I Rutgers University, Department of Computer Science %X Recent research in Expert Systems has begun to deal with problem domains that do not fit into the "classification problem" mode, the latter being the sort of problems that have been most amenable to Expert System technology. Hardware Fault Diagnosis(HFD) is an example of such a problem domain. Problems in HFD typically involve a "localization" problem as a component, i.e., @i[where] is the location of the fault? This paper takes a critical look at some current work in HFD, viz. Genesereth, Davis, with a view towards determining the differences between classification and localization problems that are likely to necessitate new approaches to knowledge representation and acquisition if Expert Systems are to be successful in such a domain. %R CBM-TR-139 %I Rutgers University, Department of Computer Science %D 5/84 %T Localization Problems and Expert Systems %A Allen Ginsberg %K AI01 %X Expert systems approaches to problem solving have recently had enormous success and influence in the field of AI. The most successful of these systems tend to deal with a certain kind of problem type which have been called "classification problems." Very recently, we have seen the emergence of a number of expert systems that deal with a different category of problem, a category that I will call "localization problems." The purpose of this paper is to characterize this class of problems, contrast it with the classification problem category, give some examples of localization problems, and suggest some new avenues for expert system research dealing with problems in this category. %R CBM-TR-140 %I Rutgers University, Department of Computer Science %D 5/84 %T Investigations in the Mathematical Theory of Problem Space Representations and Problem Solving Methods %A Allen Ginsberg %K AI09 %X In this paper I address the issue of how a system that has the ability to do problem solving and planning - in the sense of being in possession of generalized schemas or templates for carrying out these activities - can know whether a particular type of planning or, if you will, problem solving strategy, is a "good" one to employ in solving problems in a particular domain? It seems to me that, in general, in order to make such judgements in a reasonable fashion a problem solver must either be in possession of some general theoretical facts concerning the nature and structure of problem types, i.e., a theory of problem types, or at the very least, have been programmed by someone having such a theory. This paper is a step in the direction of constructing such a theory. .sp 1 The structure of the paper is as follows. First I discuss the nature of problem solving and planning in general, and give a preliminary description of a particular planning template. Next I describe and illustrate a mathematical framework within which one can formulate problem representations. Finally I deal with the question of what facts about the structure of a problem representation are relevant to the determination of whether or not the aforementioned planning template is applicable to the problem at hand. %R CBM-TR-141 %D 5/84 %I Rutgers University, Department of Computer Science %T Representation & Problem Solving: Theoretical Foundations %A Allen Ginsberg %X The word "representation" and its cognates is probably the most popular word in AI today. If anything qualifies as "the fundamental assumption of AI," it is probably the view that intelligence is essentially the ability to construct and manipulate symbolic @i[representations] of some "reality" in order to achieve desired ends. Furthermore, probably every researcher in AI would agree that the key to AI's success lies with the general area known as "knowledge representation." This point of view has been buttressed not only by the failures of early "general purpose" AI systems, but much more so by the recent success of expert systems. The philosophy behind the expert systems approach is one that has, rightfully come to infect the entire field of AI: intelligence essentially depends upon the ability to @i[represent] and store a potentially vast amount of knowledge in ways that enable it to be easily accessed and utilized in the performance of various tasks. The key concept here is @i[representation]. %X Given the fact that AI has come to embrace these doctrines, and the likelihood that there is a good deal of truth in them, it is incumbent upon us to examine their foundations, for better or for worse. It would be nice to have answers to questions such as What is a representation?, When are two or more representations representations of the same or different real world situations?, What are the ways in which representations can be "manipulated?" It would be even nicer if the answers to such questions were provided by a general formal theory of representation. In this paper I attempt to lay some of the groundwork for such a theory, with emphasis on the role of representation in problem solving. %R CBM-TR-142 %D 5/84 %T A Model for Automated Theory Formation for Problem Solving Systems" %A A. Ginsberg %X The goal of this paper is to contribute towards the understanding and eventual mechanization of the processes whereby an @i[intelligent] problem solver @i[learns] to improve its performance in a given task domain by formulating and using @i[theories] regarding that domain. In order to achieve this goal it is necessary for us, as designers of such a system, to have a fairly good idea of a) the various sorts of knowledge that are required for a problem solver to acquire new knowledge that will hopefully improve performance, and of b) how each of these types or sources of knowledge comes into play in this process. In this paper I give an abstract description of the domains of knowledege required for theory formation, and also illustrate the ideas with a concrete example. The type of system contemplated in this paper incorporates ways of structuring background knowledge that are natural and will, I believe, prove to be useful in designing self-improving AI programs. %X In this paper I address the issue of how a system that has the ability to do problem solving and planning - in the sense of being in possession of generalized schemas or templates for carrying out these activities - can know whether a particular type of planning or, if you will, problem solving strategy, is a "good" one to employ in solving problems in a particular domain? It seems to me that, in general, in order to make such judgements in a reasonable fashion a problem solver must either be in possession of some general theoretical facts concerning the nature and structure of problem types, i.e., a theory of problem types, or at the very least, have been programmed by someone having such a theory. This paper is a step in the direction of constructing such a theory. %X The structure of the paper is as follows. First I discuss the nature of problem solving and planning in general, and give a preliminary description of a particular planning template. Next I describe and illustrate a mathematical framework within which one can formulate problem representations. Finally I deal with the question of what facts about the structure of a problem representation are relevant to the determination of whether or not the aforementioned planning template is applicable to the problem at hand. %R CBM-TR-143 %D 5/84 %I Rutgers University, Department of Computer Science %T A Knowledge Representation Framework for Expert Control of Interactive Software Systems %A Apte, C. %A S. Weiss %K AI01 AA08 %X Expert problem solving strategies in many domains make use of detailed quantitative or mathematical techniques coupled with experiential knowledge about how these techniques can be used to solve problems. In many such domains, these techniques are available as part of complex software packages. In attempting to build expert systems in these domains, we wish to make use of these existing packages, and are therefore faced with an important problem: how to integrate the existing software, and knowledge about its use, into a practical expert system. We define a framework of a @i[hybrid model] for representing problem solving knowledge in such domains. A hybrid model consists of a @i[surface] and a @i[deep] model. The surface model is the production rule-based expert subsystem that is driven by domain specific control and interpretive knowledge. The deep model is the existing software, reorganized as necessary for its interpretation by the surface model. We present an outline of a specialized form-based system for acquisition and representation of expert knowledge required for this hybrid modeling. %R CBM-TR-144 (THESIS) %I Rutgers University, Department of Computer Science %D 9/84 %T A Framework for Expert Control of Interactive Software Systems %A C.V. Apte %K AI01 AA08 %X Expert problem-solving strategies in many domains require the use of detailed mathematical techniquers coupled with experiential knowledge about how and when to use the appropriate techniques. In many of these domains, such techniques are made available to experts in large software packages. In attempting to build expert systems for these domains, we wish to make use of these existing packages, and are therefore faced with an important problem: how to integrate the existing software, and knowledge about its use, into a practical expert system. The expert knowledge is used, in dynamic selection of appropriate programs and parameters, to reach a successful goal in the problem-solving. This kind of expert problem-solving is achieved through two interacting bodies of knowledge; problem domain knowledge, and knowledge about the programs that comprise the software package. %X This thesis describes the framework of a @i[hybrid expert system] for representing problem-solving knowledge in these domains. This hybrid system may be characterized as consisting of a @i[surface] model and a @i[deep] model. The surface model is a production-rule based expert subsystem that consists of heuristics used by an expert. The deep model is a collection of methods, each parameterized by a set of controlling and observed parameters. The method and their results are reasoned about using their parameter sets. The existing software is reorganized as necessary to map it into the deep model structure of a hybrid system. This framework has evolved out of an effort to build an expert system for performing well-log analysis (ELAS - @i[Expert Log Analysis System]). A generalized expert-system building methodology based upon principles drawn from ELAS is introduced. The use of @i[method-abstractions] in assembling a hybrid system is discussed. The notion of @i[worksheet-reasoning] is defined, and discussed. %R CBM-TR-145 (THESIS) %D 10/84 %T Shift of Bias for Inductive Concept Learning %A Paul E. Utgoff %K AI04 %X We identify and examine the fundamental role that bias plays in inductive concept learning. Bias is the set of all influences, procedural or declarative, that causes a concept learner to prefer one hypothesis to another. Much of the success of concept learning programs to date results from the program's author having provided the learning program with appropriate bias. To date there has been no good mechanical method for shifting from one bias to another that is better. Instead, the author of a learning program has himself had to search for a better bias. The program author manually generates a bias, from scratch or by revising a previous bias, and then tests it in his program. If the author is not satisfied with the induced concepts, then he repeats the manual-generate and program-test cycle. If the author is satisfied, then he deems his program successful. Too often, he does not recognize his own role in the learning process. .sp 1 Our thesis is that search for appropriate bias is itself a major part of the learning task, and that we can create mechanical procedures for conducting a well-directed search for an appropriate bias. We would like to understand better how a program author does about doing his search for appropriate bias. What insights does he have? What does he learn when he observes that a particular bias produces poor performance? What domain knowledge does he apply? .sp 1 We explore the problem of mechanizing the search for appropriate bias. To that end, we develop a framework for a procedure that shifts bias. We then build two instantiations of the procedure in a program called STABB, which we then incorporate in the LEX learning program. One, called "constraint back propagation" uses analytic deduction. We report experiments with the implementations that both demonstrate the usefulness of the framework, and uncover important issues for this kind of learning. %R CBM-TR-146 %I Rutgers University, Department of Computer Science %D 5/85 %T A Framework for Representation of Expertise in Experimental Design for Enzyme Kinetics %A Von-Wun Soo %A Casimir A. Kulikowski %A David Garfinkel %K AA10 AI01 %X In this paper, we present part of our current research on expert systems in enzyme kinetics. Because of the richness and diversity of the problem solving knowledge required in this domain, we have found it to be an excellent vehicle for studying issues of knowledge representation and expert reasoning in AI. Biochemical experimental design, the focus of this paper, is a major problem solving activity of the enzyme kineticist that has not been explored by expert systems researchers. Their problem solving expertise can usually be described as the application of a sequence of methods. In designing a complicated biochemical experiment, the experimenter has several methods to choose from at any stage. These methods are represented as computer programs which can be organized into a hierarchy. This paper proposes a structure for these problem solving methods and an expert consultation system for experimental design. .sp 1 We have found that problem solving expertise in experimental design can be divided into three phases. In the first phase, we deal with problems of selecting the experimental methods that satisfy an experimenter's goal, given certain postulated models. The experimental conditions and optimal design points can be derived if the model is given and the goal and the assumptions of the optimal design criterion are satisfied. In the second phase, we deal with the problems of preparing an enzyme assay. The interactions among experimental conditions and other influencing factors must be carefully controlled so that the correct concentration of a given species can be calculated. In the third phase, we face the problem of analyzing and interpreting the experimental data and recommending further refinement of the experiment. ------------------------------ End of AIList Digest ******************** From vtcs1::in% Sun Apr 13 06:41:00 1986 Date: Sun, 13 Apr 86 06:40:53 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #83 Status: R AIList Digest Sunday, 13 Apr 1986 Volume 4 : Issue 83 Today's Topics: Bibliography - Recent Articles #1 ---------------------------------------------------------------------- Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Recent Articles #1 %J ComputerWorld %D FEB 10, 1986 %V 20 %N 6 %P 44+ %K AI01 AT10 AA15 University of Calgary GA04 Synerlogic %X "Synerlogic, Inc. has joined with the University of Calgary (Alberta) in a project to develop an expert system to assist in converting subject matter knowledge into computer-based training courseware." %T Martinizing %J Datamation %D FEB 15, 1986 %P 19 %N 4 %V 32 %K AT14 AT13 AT12 AA08 James Martin Knowledge Ware %X "I wish to correct a serious error in your article on 'building a Better Program' (Oct. 1 p 42). The TI Tools do not, as you state, have 'levels of integration far in advance of DDI.' Exactly the opposite is true. A brief session with the DDI tools and the Ti too ls would reveal immediately that the TI tools cannot compare in richness and functionality with the DDI tools. The DDI tools use artificial intelligence techniques and are a generation beyond TI. THE DDI tools already have extensive use in MIS organizations. The direct implementation of my own implementations i s in DDI. On Dec. 1 DDI changed its name to KnowledgeWare to reflect the AI knowledge base of its tools. This is as far as I am aware, the first practical application of AI techniques for automating the planning, analysis and design of systems. James Martin Tuppeny House Tuckerstown, Bermuda (begin section by editors) James Martin appears to be confused about the 'tools' to which user sources are referring in the story. The sources compared the relative merits of (to date) unannounced tools under development at TI and DDI - into software productivity tools currently used in MIS organizations. Though Martin has consulted with TI on the use of his methodology and is entitled to his opinion, company sources say that he isn't in a position to effectively comment on it upcoming tools or to compare them with those of DDI. We'll have to wait for the marketplace to do that. -- ED" %T World Watch %J Datamation %P 60 %D FEB 15, 1986 %P 19 %N 4 %V 32 %K GA01 India Institute for New Generation Technology %X "India hopes to curry favor with Japan's Institute for New Generation Computer Technology. So, a few of India's premier fifth-generation researchers may soon be packing their bags for a trip to India." %A John R. Dixon %T Will Mechanical Engineers Survive Artificial Intelligence %J Mechanical Engineering %D FEB 1986 %P 8+ %V 108 %N 2 %K AA05 AT14 %X Raj Reddy stated 'In the twenty-first century, much of what mechanical engineers now do will be done by machines' The rest of the editorial, discusses whether this is a reality. %A Howard K. Dicken %T Turning Micros Into Mavens %J High Technology %D MAR 1986 %P 71 %V 6 %N 3 %K AI01 H01 Expertelligence Macintosh AT16 Migent Software Intellicorp Enrich Transform Logic AA15 %X Expertelligence, which sells expert-system shells, Lisp and Prolog for the Macintosh had fiscal 1985 revenue of $834,000. Losses were $411,000. Intellicorp had 1985 sales of $8.7 million dollars with a loss of $724,000. Migent Software has purchased an expert system for interfaces to user software from Transform Logic. The software is called Enrich and sells for $595.00 %A Stanley Aronoff %A Glyn F. Jones %T From Data to Image to Action %J IEEE Spectrum %D DEC 1985 %V 22 %N 12 %P 45-52 %K AI06 Mult-Spectral Scanner Landsat AA03 forestry crop yields Cropcast AI01 %X discusses various aspects of the hardware for image processing. Crop forecasting now achieve 97% accuracy with 95% accuracy three months prior to harvest. They state that expert systems will combined with image processing to create a new generation of information systems. %T Adept, Kawasaki in Japan Accord %J Electronic News %D FEB 10, 1986 %P 45 %V 32 %N 1588 %K AI07 GA01 GA02 AT16 AI06 %X Adept has licensed Kawasaki Heavy industries to manufacture and sell its robotics line in Japan. Adept estimated it will receive one million dollars in the next three years. This also includes the AdeptVision systems. Adept has shipped more than $500,000 worth of robotics equipment ot Kawaski since last September %T Notes: Software and Services %J ComputerWorld %D JAN 27, 1986 %P 33 %V 20 %N 4 %K LogicWare MProlog Revelations Control Data Corp Cyber H04 T02 AT16 %X Logicware and Revelations Research have joined efforts to put a version of MProlog on the Control Data Corp's Cyber 205 %A Eric Bender %T DBMS tools: Not Natural Yet %D JAN 20, 1986 %P 19+ %V 20 %N 3 %K Ashton-Tate H01 AA09 AI02 Clout Lotus Development Human Access Language Brodie Associates %X Interviews with various people about natural language and data base management systems, particularly for micros. Of note, David Hull of Ashton Tate said that although they are evaluating natural language systems, they have not seen any that deliver the benefits that they think their clients want %T New Products %D JAN 20, 1986 %P 85 %V 20 %N 3 %K Experience in Software Idea Generator H01 AI01 %X Experience in Software, Inc. announced the Idea Generator a tool to help the user solve problems. It costs $195.00 and runs on the IBM PC. %A Steven Burke %T Arity/Prolog Tools Assist in Creating AI-Based Software %J InfoWorld %D JAN 20, 1986 %P 14 %V 8 %N 3 %K Unitek Technology Arity Dr. vance Giboney Arthur Young and Company Peter Gabel Darryl Rubin Kim Frazier AI01 AA06 AA08 GA04 H01 T01 T02 %X Unitek Technologies is using Arity's tools to enhance accounting sofware. Knowledgeware is using it to automate writing computer code which is being developed in conjunction with Arthur Young and Company. %A Alice LaPlante %T Talking with your Computer %J InfoWorld %V 8 %N 2 %D JAN 13, 1986 %P 25-26 %K Digital Equipment Corporation DECtalk AI05 AI01 %X general discussion of applications of uses of voice input and output systems. DEC says that 90 percent of its customer's use DECTALK for telephone applications; it expects that its next generation system will have voice recognition and voice synthesis as part of an expert system. %A Keith Thompson %T Q&A is Fun, Useful Business System %J InfoWorld %V 8 %N 2 %D JAN 13, 1986 %K AI02 AA15 AA09 H01 AT17 %X review of Q and A, which is a database and word processor claiming to be based on artificial intelligence. It has a natural language interface to the data base. It uses AI to tell where the address is in a letter automatically to print out the envelope. It received a rating of 9.0 out of 10 with very good in performance and excellent in documentation, ease of learning, ease of use, error handling, support and value. %A Barbara Robertson %T The AI Typist: Writing Aid is Fast and Easy, But Bug Plagued %J InfoWorld %V 8 %N 2 %D JAN 13, 1986 %P 35 %K AT17 AT03 H01 AA15 %X AI Typist is a word processing system for IBM PC's that "uses artificial intelligence to provide a real-time typist." The program scans a dictionary looking for character-by-character matches while typing. It highlights characters at the point it finds a mismatch. For example, if a user types appearing, highlighting appears as one types the second a since ape matches a word in the dictionary. It doesn't correct the spelling nor allow the user to look at the dictionary. It also had bugs in the basic word processing capability. It received a 2.4 out of 10 with unacceptable ratings under performance and value, poor in documentation, satisfactory in error handling and very good under ease of learning, ease of use and support. %T TI Introduces PC Scheme Lisp Device %J InfoWorld %V 8 %N 2 %D JAN 13, 1986 %P 51 %K T01 H01 AT02 %X TI has introduced PC Scheme for $95.00 which runs on IBM PC's and TI Instruments PC's. It has a compiler. %T Advertisement %J Unix/World %V 11 %N 11 %D DEC 1985 %P 56 %K Silogic Knowledge WorkBench AT01 AI01 AI02 AA09 T03 %T For the Record %J Unix/World %V 11 %N 11 %D DEC 1985 %P 10 %K Flexible Computer NASA Johnson Space Center Unix %X "NASA's Johnson Space Center, Dallas, has purchased a massively parallel Flex/32 Computer from Flexible Computer Corp. for its Artificial Intelligence section, which is responsible for evaluating fifth generation computing systems for AI development and applications." %T Review of Introduction to Robotics by Arthur J. Critchlow %J BYTE %V 11 %N 1 %D JAN 1986 %P 57-60 %K AI07 AT07 %T Software Notes %J ComputerWorld %D JAN 13, 1986 %V 20 %N 2 %P 25+ %K Inference Corp NASA Symbolics AT16 AI01 AA08 %X "Inference Corp. and the National Aeronautics and Space Administration have agreed to develop jointly a software development workstation design Inference's expert system technology." It will use Symbolics 3600 and assist both in reuse of code and generation of new code. %A Edward Warner %T Gold Hill, Intel Developing LISP for Multimicroprocessors %J ComputerWorld %D JAN 13, 1986 %V 20 %N 2 %P 26 %K T01 H03 %X Intel announced an Agreement with Gold Hill to develop and market jointly a Common Lisp Computer Intel's HyperCube IPSC. %T Executive Report/Expert Systems %J ComputerWorld %D JAN 13, 1986 %V 20 %N 2 %P 43-62 %K T01 T02 T03 H02 AI01 AA04 AA08 AA21 AT08 O02 Sterling Wentworth PlanMan Fountain Hills Software semiconductor Travellers Insurance Teknowledge %X Half of all Fortune 500 companies actively pursue expert system development. Fountain Hills Software sells Fair Cost, a cost modelling program for semiconductor components Sterling Wentworth Corpo offers Planman, a financial panning expert system targetted at tax advisors. Ion Technology Services, markets Diagnostic Troubleshooter an expert system for the maintenance of specialized semiconductor equipment The expert system market is worth 75 million with government and research efforts account for as much as two thirds of this. Fortune 500 companies efforts make up most of the rest of the market. Custom Development Life Span for Expert Systems compiled by Arthur D. Little in developing 30 "large-scale strategic knowledge-based systems, typically for Fortune 500 companies:" .TS tab(~); c c c c l l l l. Phase~Duration~Level of Effort~Cost Proof of Concept:~4 to 6 months~1 to 2 man years~$150,000 to $400,000 Demonstration~4 to 6 months~1 to 2 man years~$150,000 to $400,000 Prototype~12 to 18 months~8 to 12 man years~$1.2 to 2.4 million Total Resource Comm~20 to 30 months~10 to 16 man years~$1.5 to 3.2 milion .TE Travelers Insurance developed a successful expert system to help diagnose failures on IBM 8100 controllers. It had 70 rule programs and was done with Teknowledge's M-1 system. %T SuperShorts %J ComputerWorld %D JAN 13, 1986 %V 20 %N 2 %P 108 %K AI01 AA20 T03 H02 AT16 %X Lisp Machine, Inc. and the Process Management Divison of Honeywell announced that they will work to bring artificial intelligence to the process control market. As part of that effort, they will work together to interface PICON with the Honeywell control system TDC3000. [In Applied Artificial Intelligence, it was reported that this interfacing was already accomplished and is running at one site. LEFF] %T GM Delco to Fund Cognex Vision Systems %J Electronic News %D JAN 6, 1986 %V 32 %N 1583 %P 58 %K AI06 AA04 AT16 %X Delco Electronics has agreed to give COGNEX $500,000 to develop an engineering prototype of a machine vision system for automatically inspecting the placement of surface-mounted devices on printed circuit boards. Cognex's Checkpoint 1100 system was reported to have achieved measurements accurate to within 2 mils within 99.8 percent of the test cases. %A James Fallon %T Racal Electronics, Norsk Data to End 2-Year AI Joint Venture %J Electronic News %D JAN 13, 1986 %V 32 %N 1584 %P 29 %K AT16 GA03 %X A 1.44 million joint investment to develop an artificial system was terminated since the project was delayed and the market for that particular product no longer existed. %T Plessey to Develop Speech-Input CPU %J Electronic News %V 31 %N 1582 %D DEC 30, 1985 %P 8 %K Alvey Edinburgh University Imperial College University of Loughborough AI05 GA03 H03 %X The Alvey Directorate has selected Plessey as a prime contractor in a 19.88 million dollar project to develop a system that receives human speech and displays the words on the screen. No vocabulary size or response time was given for the proposed system. It will use parallel processing %A Peggy Watt %T Scanner Puts Text On-Line %J ComputerWorld %D Dec 30, 1985/JAN 6, 1986 %V 19 %N 52 %P 1+ %K Dest Corporation AI06 AT02 %X DEST Company announced PCSCAN which is a system that recognizes the type faces in most business documents. It costs $3000.00. The optical reader equipment supports 300 dpi printers. The optical reader alone is $1995.00. The software inserts appropriate formatting codes for such things as tabs, paragraphs and page breaks. %A J. Mostow %T Forword: What is AI? And What Does It Have to Do with Software Engineering? %J IEEE Transactions on Software Engineering %V SE-11 %N 11 %D NOV 1985 %P 1253-1256 $K AA08 %A R. Balzert %T A Fifteen Year Perspective on Automatic Programming %J IEEE Transactions on Software Engineering %V SE-11 %N 11 %D NOV 1985 %P 1257-1267 %K AI08 SAFE AA08 Insformation Sciences Institute GIST RSL TRW symbolic evaluation software maintenance POPART PADDLE %X SAFE was a system that took up to a dozen informal sentences that specified a piece of software and produced a formal specification. GIST is a formal specification language that attempted to minimize the translation from the way people think about processes to the way they write about them. They developed a prototype of a system to convert GIST to natural language and they have a joint effort underway with TRW to design a system to convert RSL specifications to natural language. They also developed a system to symbolically evaluate GIST specifications. They also have a natural language behavior explainer. %T New Products/Microcomputers %J ComputerWorld %D FEB 24, 1986 %V 20 %N 8 %P 89 %K T01 H02 Practical Artificial Intelligence VAX DS-32 AP-10 %X Practical Artificial Intelligience has announced the DS-32 and AP/10 which are attached processors for the IBM personal computer and Digital Equipment VAX designed to support artificial intelligence. The DS-32 costs $2700 and the AP/10 costs $6000 %T Ben Rosen's Ansa: Will it Ever be Another Lotus? %J Business Week %D MAR 3, 1986 %P 92-95 %V 2935 %K Paradox SRI AA09 H01 %X discusses the founding and prospects for Paradox, a data base system with artificial intelligence features %A Mary Petrosky %T Expert Software Aids Large Systems Design %J Infoworld %V 8 %N 7 %P 1+ %K AI01 AA08 H01 AT02 AT03 %X Knowledge-Ware announced Information Engineering Workstation that provides tools for data flow diagrams and action diagrams. It runs on IBM PC/AT's and cost $7500.00. I could not find an explanation of where AI was used, in spite of the title of the article. %T Expert System Moves Into Military Cockpit %J Electronics %V 58 %N 51 %P 15 %D DEC 23, 1985 %K AA18 Air Force Wright Aeronautical Laboratory Threat Expert Analysis System AI01 %X The Air Force's Wright Aeronautical Laboratory, Wright-Patterson Air Force Base had set a deadline of January 10 for the Threat Expert Analysis System, a system that would warn pilots of enemy threats and recommend possible responses. %T Device Mixes Images from Eight Cameras %J Electronics %V 58 %N 51 %P 76 %D DEC 23, 1985 %K Pattern Processing Technologies Framesplitter AI06 %X Framesplitter is a system that combines the input from several solid state video cameras into a single composite image. This system allows a system to gain a 360 degree view while only processing one image. %A Clifford Barney %T Language Boils Down to Boolean Expressions %J Electronics %V 58 %N 51 %P 25-26 %D DEC 23, 1985 %K G. Spencer-Brown Wittgenstein Bertrand Russel Laws of Form Advanced Decision Systems Air Force pictorial logic canonical forms Losp Symbolics AI10 AI14 AA18 H02 T01 T02 %X Losp is a system based on the "Laws of Form" which was developed by G. Spencer-Brown a British Mathematician who studied with Bertrand Wittgenstein. The system was developed by Advanced Decision Systems and will be put to use in an Air Force project on pictorial logic. The language is being microcoded to run on a Symbolics work station. Lisp and Prolog will be translated to LOSP ------------------------------ End of AIList Digest ******************** From vtcs1::in% Mon Apr 14 00:51:02 1986 Date: Mon, 14 Apr 86 00:50:57 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #84 Status: R AIList Digest Sunday, 13 Apr 1986 Volume 4 : Issue 84 Today's Topics: Bibliography - Recent Articles #2 ---------------------------------------------------------------------- Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Recent Articles #2 %A Clifford Barney %T Expert Systems Makes it Easy to Fix Instruments %J Electronics %V 58 %N 51 %D DEC 23, 1985 %P 26 %K AI01 AA04 AA21 Ada Lockheed Missiles and Space Lockheed Expert System %X Lockheed Missiles and Space has developed a generic expert system to assist in repairing and calibrating 55000 instruments. This system has been used successfully on a Hewlett-Packard 6130C digital voltage source. The epxert system was written in ADA. The system is being applied to 20 different systems including signal-switching and computer aided design. %A Robert T. Gallagher %T French Make Retools to Fight the Japanese %J Electronics %V 58 %N 51 %P 26-28 %D DEC 23, 1985 %K AA05 GA03 RTC La Radiotechnique Compelec cathode-ray tube AI07 %X RTC La Radiotechnique Compelec has converted a cathode-ray tube to robotics. Robots are being used to place the luminescent materials on the tube screens, testing, and placement in packing materials. The areas requiring manual work are fitting the shadow masks on to the CRT's frames and the final test where tuning is done. %A H. Berghel %T Spelling Verification in Prolog %J SIGPLAN Notices %V 21 %N 1 %D JAN 1986 %P 19-27 %K T02 %X describes a system to check words against table and if mispelled to suggest possible correct spellings. %A D. Brand %T On Typing in Prolog %J SIGPLAN Notices %V 21 %N 1 %D JAN 1986 %P 28-30 %K T02 %T Advertisement %J BYTE %D JAN 1986 %V 11 %N 1 %P 348 %K T01 T02 T03 H01 AT01 %X Price List on AI type products from The Programmers SHOP 800-421-8006 128B Rockland Street, Hanover MA 02339 .TS tab(~); l l l. EXSYS~PCDOS~$359 INSIGHT 1~PCDOS~$95 INSIGHT 2~PCDOS~$449 APES~~$359 ADVISOR~~$949 ES Construction~~$100 ESP~~ $845 Expert CHOICE~~$449 GC LISP (Large Model)~~$649 Compiler and LM Interpreter~~$1045 TLC LISP~CPM-86~$235 ~MSDOS Waltz LISP~CPM~$149 ~MSDOS ExperLisp~~$439 IQ LISP~~$155 TRANSLISP-PC~~$75 BYSO~~$125 MuLISP-86~$199 ARITY PROLOG ~Compiler~$1950 ~MSDOS~$495 MPROLOG~PCDOS~~$725 PROLOG-1~~$359 PROLOG-2~~$1849 MicroProlog~~$229 Prof. Micro Prolog3~~$359 .TE %A Hugh Aldersey-Williams %T Computer Eyes Turn to Food %J High Technology %D JAN 1986 %P 66-67 %V 6 %N 1 %K Vision Systems International Nello Zuech Roger Brook strawberry citrus juice mixed vegetables AI06 %X At the University of Florida, Gainseville, they are working on a vision system to pick citrus fruit when it is ripe. Arthur D. Little is working on a sytem that would determine whether the mixture in a package of mixed vegetables contains the correct proportion of different ingredients. %T International Robomation Gets Two Million in Orders %J Electronic News %D MAR 3, 1986 %P 46 %V 32 %N 1591 %K AI07 AA04 Chrysler AT&T Hewlett-Packard Zenith SMD printed-circuit board solder paste %X Orders included $750,000 from Chrysler for a surface mounted device inspection system, $400,000 from HP, $270,000 for inspection of through-hole components, $400,000 from AT&T and $305,000 from Zenith for high-through put SMD inspection %T Asahi to Market Lincoln Inspector %J Electronic News %D MAR 3, 1986 %P 46 %V 32 %N 1591 %K AA04 AI06 Lincoln Laser GA01 GA02 %X Asahi Optical Company has agreed to market Lincoln Laser Co's line of automatic optical inspection systems in Japan. Lincoln Laser plans to manufacture the equipment in Japan. First year sales are projected to be 40 systems valued at approximately $16.7 million. %A Peggy Watt %T Expert System: Boeing AI Academy Schools In-House Talent %J ComputerWorld %D MAR 3, 1986 %V 20 %N 9 %P 1+ %K AI01 AT18 AA18 Janusz S. Kowalik connectors AA04 space station. %X U. S. Department of Defense announced in 1981 that artificial intelligence will be a requisite in defense contract bids in the late 1980's. Boeing Computer Services established an Artificial Intelligence Support Center which graduates associates after a year of training including developing a project of relevance to Boeing. The system acommodates 20 people in two classes scheduled each year and receives inquiries from 40 people a year out of 106,000 total employees of Boeing. They are developing an expert system for process specs for connector assemblies. It recommends actions in about 60% of the situations it encounters. It runs in Prolog on a DEC VAX. They are developing an expert system to monitor space-station cabin environment changes. Also developed are systems for airplane part design maintenance and diagnosis. Another helps determine air resistance assists the aerodynamicist in defining and evaluating parameters. %T Top of the News %J ComputerWorld %D MAR 3, 1986 %V 20 %N 9 %P 1+ %K Kurzweill Applied Intelligence Voice Writer AI05 %X "Kurzweil Applied Intelligence, Inc.'s Voice Writer, a voice-recognition word processing device that will handle discrete, noncontinuous speech at up to 60 words per minute, is on track for a third-quarter introduction, inventor Raymond Kurzweil disclosed last week." It will support between 5000 and 10000 words and will cost under $20,000. %T Borland Enters AI Arena with Turbo Prolog Development Tool %J ComputerWorld %D MAR 3, 1986 %V 20 %N 9 %P 14 %K Turbo-Prolog Borland International T02 H01 %X Turbo-Prolog costs $99.95. It has an incremental compiler that generates native code and linkable object modules compatible with the IBM MS-DOS linker. It includes a full screen editor, pull-down menus, graphical and text-based windows. It will be available April 25. The next version of Turbo Pascal will be able to exchange information with Turbo Prolog. The system runs at 100,000 LIPS. %T New Products/Microcomputers %J ComputerWorld %D MAR 3, 1986 %V 20 %N 9 %K OPS-83 Production Systems Technologies T03 H01 %X Production Systems Technologies, Inc. has announced that OPS83 is now available for use on the IBM PC. It costs $1950.00 %T New Products/Systems and Peripherals %J ComputerWorld %D MAR 3, 1986 %V 20 %N 9 %K Maxvideo Minvideo Datacube Multibus Addgen-1 frame store DSP Systems AI06 %X Datacube Inc. has introduced Minivideo, a real time image processing subsystem, and has added three modules to its MaxVideo product line. Minvideo-10 and Minvideo 7 are 8-bit 512 by 512 and 384 by 512 boards for Intel Multibus or IIbx based computers. X DSP system has announced a FRAME STORE that can store a snap shot of 50 Mhz data. It can store up to 32K 16 bit words %A Gadi Kaplan %T Industrial Electronics %J IEEE Spectrum %V 23 %N 1 %D JAN 1986 %P 61-64 %K Fujitsu General Electric process control Foxboro Farot-M6 AI06 AA05 AI07 counterweights %X GE has developed a system that can weld using inert gas at 40 mm per second or about twice the rate of any other system. It uses a vision system. General Electric has developed an expert system tool called GEN-X. Foxboro announced controllers with 200 rules. Japanese manufacturers last year made 50,000 industrial robots valued at 1.2 billion. Fujitsu expects to sell $2.1 billion in Japanese industrial robots and $4 billion in 2000 years. Japanese auto manufacturers buy 40 percent of the robots produced. Farot M6 robots made by Fujitsu have two arms which can be worked in coordination. Fujitsu has eliminated the need for counterweights and can place components with 30 micrometer accuracy at speeds up to 2 meters per second. %A Mark A. Fischeti %A Glenn Zorpette %T Power and Energy %J IEEE Spectrum %V 23 %N 1 %D JAN 1986 %K AA04 AI01 Westinghouse Electric Corporation nuclear power Babcock and Wilson EG&G Idaho reactor %X "Westinghouse Electric Corporation of Pittsburgh, PA offers the Genaid diagnostic software package to monitor changing conditions in power plant generators, analyze them, and warn plant operators of potential trouble." EG&G Idaho of Idaho Falls has a Reactor Safety Assessment system which "processes large amounts of data from a nuclear power plant during an emergency, makes diagnoses, and outliens the consequences of subsequent actions. After final refinements, this expert system program is to go on line this year at he Nuclear Regulatory Commison's Operations Center in Washington Center. The system was developed for use with Babcock and Wilcox Pressurized-water reactors and will be adapted for use with other reactors." [In Spang-Robinson report, they indicated that the Japanese are putting major amounts of money into expert systems for nuclear reactor operations. See my summary for more info. LEFF ] %A Richard Brandt %T Micromechanics: The Eyes and Ears of Tomorrow's Computers %J BusinessWeek %D MAR 17, 1986 %P 88-89 %N 2937 %K AI07 AI06 signature verification Novasensor Schlumberger diabetes insulin Clini-Therm Corporation NEC Solartron Transducer Hiroshi Tanigawa %X Micromechanics, the making of mechanical sensors completely out of semiconductors, is a $250,000,000 business. Europe is increasing its market share. The most widely used devices are pressure sensors with a silicon chip with a hole etched nearly through it leaving a thin membrane. Hitachi sells about a million of such sensors per year which it sells at ten dollars a piece. Millar Instruments puts such sensors at the end of a blood pressure monitor to take readings inside a blood vessel. Researchers at MIT are working on a system that will translate nerve impulses into controls for prosthetics. The MIT team anticipates the first tests on humans with three years. There are devices with a set of diving boards for measuring accelerations. IBM is using such a device in a pen to detect the hand motions in writing a signature. This data is analyzed to determine if there is a forger. Texas Instruments is perfecting a silicon chip with one million mirrors for use in optical computing. %T AI to Dominate Optics Symposium %J Electronics %D MAR 3, 1986 %P 70 %V 59 %N 9 %K George Gilmore AI06 evidencing AA18 %X Discussion of the Society of Photooptical Instrumentation Engineers Symposium on Optics symposium on Applications of Artificial Intelligence III. %A Alice LaPlante %T Stock Market Finds AI Attractive Buy %J InfoWorld %V 8 %N 9 %D MAR 3, 1986 %K Teknowledge Harvey Newquist Intellicorp AT16 %X Discussions of public offerings of Teknowledge's new public offering. %A Ivars Peterson %T Computing Art %J Science News %V 129 %P 138-140 %N 9 %D MAR 1, 1986 %K Richard Diebenkorn Frank Lloyd Wright Architecture grammar art architecture Russell Kirsch Joan Marvin Minsky AA25 %X Using a grammar, scientists have developed grammars for Frank Lloyd Wright's architecture and Richard Diebenkorn's "Ocean Park" canvasses. These have been used to develop works that appeared to by the author. Diebenkorn when shown the works said "I looked and felt immediate recognition." %A Scott Mace %T Microrim Team To Study Data Management %J Infoworld %D March 10, 1986 %V 8 %N 10 %K AA09 %X Microrim is setting a R&D group to exploit what it calls a 'potentially revolutionary' technology for making database management easier. [Microrim makes RBASE database products for microcomputers and CLOUT, a natural language interface.] %A Karen Sorensen %T Scientific Application for Expert System in Works %J Infoworld %D March 10, 1986 %V 8 %N 10 %K gas chromatography Award Software AI01 AA02 Award Software C H01 %X Award Software is developing an expert system for making identifications of chemical substances. It is designed for use with gas chromatography. They are using C to develop the software. %T Infomarket %J Infoworld %D March 10, 1986 %V 8 %N 10 %K H01 T03 Intelligent Machine Co. Knowledge Oriented Language Knowol Rock Mountain Medical Software HouseCall AI01 AA01 %X Intelligent Machine Co is advertising The Knowledge Oriented Language for $39.95. HouseCall is a home medical system which can make over 400 diagnoses. It costs $49.95 and runs on IBM PC's and Apples %A Daniel R. Pfau %A Barry A. Zack %T Understanding Expert System Shells %J Computerworld Focus %D February 19, 1986 %V 20 %N 7A %K T03 %P 23-24 %A Girish Parikh %T Restructuring Your Cobol Programs %J Computerworld Focus %D February 19, 1986 %V 20 %N 7A %P 39-42 %K AI01 AA08 Cobol-SF %A Elisabeth Horwitt %T LISP Systems Tied to SNA %J ComputerWorld %D MAR 10, 1986 %V 20 %N 10 %P 1 %K Symbolics H02 AA06 CICS IBM %X Symbolics introduced a product to allow their Symbolics 3600's to communicates via SNA. They also provide an interface to use CICS to access VSAM files. The hardware + software costs $17,900 for the first Symbolics and $4900 for each additional Symbolics or IBM. %A Eric Bender %T The Concerted Kurzweil Effort %J ComputerWorld %D MAR 10, 1986 %V 20 %N 10 %P 33+ %K Voice Writer AI05 %X describes demonstration of Kurzweill's add on for the IBM PC to do speech recognition. Kurzweill will be selling a Voice Writer which will handle 5000 words and allow eight users. It uses parallel processing to accept dictation at 60 words per minute. %T New Products/Microcomputers %J ComputerWorld %D MAR 10, 1986 %V 20 %N 10 %P 81 %K AA08 H01 %X P-Cube Corb has annoucned Mansys/IRM a "knowledge-based" system to help assess the quality of the procedures and processes within an information systems department. It costs $1800 and runs on IBM PC's. %T Spin-Offs %J IEEE Spectrum %D MAR 1986 %V 23 %N 3 %P 17 %K Color Systems Technology colorization AA25 %X describes the system used to color old movies. %A Ernest W. Kent %A Michael O. Shneier %T Eyes for Automatons %J IEEE Spectrum %D MAR 1986 %V 23 %N 3 %P 37 %K Honeywell Navy AI06 AI07 cleaning agriculture printed circuit-board propeller CAD/CAM Control Automation Interscan Odetics range images Environmental REsearch Institute AA18 AA19 Automatix Advanced Vision Systems ITMI Marketing Corp Analog Devices Automation Intelligence %X At Honeywell, they are using a vision system to identify missing leads in printed-circuit boards. It uses four videocamera's 90 degrees apart to capture light reflections. There are 200 companies offering a product or service related to machine-vision. 50 of these offer complete systems. At a Navy ship-building system, a video system inspects propellors and compares the results against the CAD/CAM database to see if it was made possible. Autonomous mobile robots are under commercial development for materials transport, commercial cleaning, and construction. Total sales for machine vision systems have double in each of the last two years. %A Mark A. Fischetti %T A Review of Progress at MCC %J IEEE Spectrum %D MAR 1986 %V 23 %N 3 %P 76-82 %K AA04 VLSI-CAD reconvergent fanout problem H02 LDL AI10 T02 AA09 %X In the VLSI-CAD, area, they are using 81 LISP machines. They developed a module editor which lays out circuitry graphically. They have developed an algorithm for solving the reconvergent fanout problem. Discusses the knowledge-base that is supposed to contain "common-sense" They have developed a test application that will help IC chip designers. The database group is developing a system to compile large logic systems on disks %A Glenn Zorpette %T Robots for Fun and Profit %J IEEE Spectrum %D MAR 1986 %V 23 %N 3 %P 71-75 %K AA25 AI07 Survival Research Labs %X discusses various robots that are part of art shows or used for entertainment. Survival Research Labs puts on demonstrations where large mobile robots destroy props, animal carcasses or one another. %T Gould Acquires Vision Systems Unit %J Electronic News %V 32 %N 1592 %D MAR 10, 1986 %P 14 %K Gould Automated Intelligence Opti-Vision AI06 AT16 %X Gould has acquired the VisionSystems division of Automated Intelligence. This division makes the Opti-Vision system. ------------------------------ End of AIList Digest ******************** From vtcs1::in% Sun Apr 13 06:41:10 1986 Date: Sun, 13 Apr 86 06:41:04 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #85 Status: R AIList Digest Sunday, 13 Apr 1986 Volume 4 : Issue 85 Today's Topics: Bibliography - Recent Articles #3 ---------------------------------------------------------------------- Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Recent Articles #3 Definitions D BOOK24 The World Yearbook of Robotics Research and Development, 1985\ %I Gale Research Corporation\ %D 1985 D MAG14 Computer-Aided Design\ %V 17\ %N 9\ %D NOV 1985 D MAG15 Theoretical Computer Science\ %V 39\ %N 2-3\ %D AUG 1985 D BOOK25 Analysis of Concurrent Systems\ %E B. T. Denvir\ %E W. T. Harwood\ %E M. I. Jackson\ %E M. J. Wray\ %S Lecture Notes in Computer Science\ %V 207\ %I Springer-Verlag\ %C Berlin-Heidelberg-New York\ %D 1985 D MAG16 Soviet Journal of Computer and Systems Sciences\ %V 23\ %N 4\ %D JUL-AUG 1985 D MAG17 International Journal of Man-Machine Studies\ %V 23\ %N 5\ %D NOV 1985 D MAG18 Cybernetics and Systems\ %V 16\ %N 1\ %D 1985 __________________________________________________________________________ %T Intelligent Robots and Computer Vision %I SPIE -- The International Society for Optical Engineering %D September 16-20, 1985 %N 579 %C Cambridge, MA %E David P. Casasent %K AT15 AI07 AI06 %A David Nitzan %T Development of Intelligent Robots: Achievements and Issues %B BOOK24 %K AI07 %A Ray Basey %T Training for the Introduction of Robots New Technology and Control Systems, Operation and Maintenance %B BOOK24 %K AI07 AT18 %A H. H. Rosenbrook %T Social and Engineering Design of a Flexible Manufacturing System %B BOOK24 %K AI07 AA05 O05 %A Igor Aleksander %T Extension of Robot Capabilities Through Artificial Vision: A Look into the Future %B BOOK24 %K AI07 AI06 %T The World Directory of Robotics and Development Activities %B BOOK24 %K AI07 AT19 %X info on robotics research in 26 countries, list of groups %T A Guide to Grant Awarding Bodies %B BOOK24 %K AI07 AT19 %A Phillippe Coiffet %T Robot Technology: Modeling and Control %V 1 %I Prentice-Hall %D 1982 %K AI07 AT15 %A Philippe Coiffet %T Robot Technology: Interaction with the Environment %V 2 %I Prentice-Hall %D 1983 %K AI07 AT15 %A Jean Vertut %A Philippe Coiffet %T Robot Technology: Teleoperation and Robotics: Evolution and Development %V 3A %I Prentice-Hall %D 1986 %K AI07 AT15 AT20 %A Jean Vertut %A Philippe Coiffet %T Teleoperations and Robotics: Applications and Technology %V 3B %I Prentice-Hall %D 1985 %K AI07 AT15 %A F. L. Hote %T Robot Components %V 4 %I Prentice-Hall %D 1983 %K AI07 AT15 %A Michel Parent %A Claude Laureau %T Robot Technology: Logic and Programming %I Prentice-Hall %D 1985 %V 5 %K AI07 AT15 %T Robot Technology: Decision and Intelligence %I Prentice-Hall %D (not yet published) %K AI07 AT15 %V 6 %A Alain Liegeois %T Robot Technology: Performance and Computer-Aided Design %I Prentice-Hall %D 1985 %K AI07 AT15 AA05 %V 7 %A John Haugeland %T Artificial Intelligence: The Very Idea, 1985 %I MIT Press %D 1985 %K AT15 %A H. J. De Man %A I. Bolsens %A E. vanden Meersch %A J. van Cleynenbreugel %T DIALOG: An Expert Debugging System for MOS VLSI Design %J IEEE Transactions on Computer-Aided Design %D JULY 1985 %V CAD-4 %N 3 %P 303-311 %K AI01 AA04 %A Michael A. Rosenman %A John S. Gero %T Design Codes as Expert Systems %J MAG14 %P 399-409 %K AA05 AI01 %A Hitoshi Furuta %A King-Sun Tu %A James T. P. Yao %T Structural Engineering Applications of Expert Systems %J MAG14 %P 410-19 %K AA05 AI01 %A Mary Lou Maher %T HI-RISE and Beyond: Directions for Expert Systems in Design %J MAG14 %P 420-427 %K AA05 AI01 %A A. D. Radford %A J. S. Gero %T Towards Generative Expert Systems for Architectural Detailing %J MAG14 %P 428-435 %K AA05 AI01 %A David C. Brown %T Failure Handling in A Design Expert System %J MAG14 %P 436-442 %K AA05 AI01 %A Daniel R. Rehak %A H. Craig Howard %T INterfacing Expert Systems with Design Databases in Integrated CAD Systems %J MAG14 %P 443-454 %K AA05 AI01 %A Anna Hart %T Knowledge Elicitation: Issues and Methods %J MAG14 %P 455-462 %A John S. Gero %T Bibliography of Books on Artificial Intelligence with Particular Reference to Expert Systems and Knowledge Engineering %J MAG14 %P 463-464 %K AI01 AT09 %A D. Kapur %A P. Narendran %A M. S. Krishnamoorthy %A R. McNaughton %T The Church-Rosser Property and Special Thue Systems %J MAG15 %P 123-134 %K AI14 %A C. Bohm %A A. Berarducci %T Autoamtic Snythesis of Type Lambda-Programs on Term Algebras %J MAG15 %P 135-154 %K AI14 AA08 %A M. W. Bunder %T An Exptension of Klop's Counterexample to the Church-Rosser Property to Lambda-Calculus with Other Ordered Pair Combinators %J MAG15 %P 337 %K AI14 %A M. Rodriguez artalejo %T Some Questions About Expessiveness and Relative Completeness in Hoare's Logic %J MAG15 %P 189-206 %K AA08 %T The Functions of T and Nil in Lisp %J Software Practice and Experience %V 16 %N 1 %D JAN 1986 %P 1-4 %K T01 %A R. Milner %T Using Algebra for Concurrency-Some Approaches %B BOOK25 %P 7-25 %K AA08 %A H. Barringer %A R. Kuiper %T Towards the Hierarchical, Temproral Logic, Specification of Concurrent Systems %B BOOK25 %P 157-183 %K AA08 %A R. Koymans %A W. P. Deroever %T Examples of a Real-Time Temporal Logic Specification %B BOOK25 %P 231-251 %K AA08 %A V. S. Medovyy %T Translation from a Natural Language into a Formalized Language as a Heuristic Search Problem %J MAG16 %P 1-9 %K AI02 AI03 %A M. K. Valiyev %T On Temporal Dependencies in Databases %J MAG16 %P 10-17 %K AA09 %A Z. M. Kanevskiy %A V. P. LItvinenko %T Minimization of the Average Duration of a Discrete Search Procedure %J MAG16 %P 126-129 %K AI03 %A A. S. Yuschenko %T The Problem of Dynamic Control of Manipulators %J MAG16 %P 139 %K AI07 %A I. Vessey %T Expertise in Debugging Computer Programs - A Process Analysis %J MAG17 %P 459-494 %K AA08 AI08 %A J. H. Boose %T A Knowledge Acquisition Program for Expert Systems Based on Personal Construct Psychology %J MAG17 %P 495-526 %K AI01 %A E. J. Weiner %T Solving the Containment Problem for Figurative Language %J MAG17 %P 527-538 %K AI02 %A R. R. Yager %T Explantory Models in Expert Systems %J MAG17 %P 539-550 %K AI01 %A T. Munakata %T Knowledge-Based Systems for Genetics %J MAG17 %P 551-562 %K AI01 AA10 %A Ronald R. Yager %T On the Relationship of Methods of Aggregating Evidence in Expert Systems %J MAG18 %P 1-22 %K AI01 %A Ronald R. Yager %T Strong Truth and Rules of INference in Fuzzy Logic and Approximate Reasoning %J MAG18 %P 23-64 %K AI01 O04 %A Witold Pedrycz %T Structured Fuzzy Models %J MAG18 %P 103 %K O04 ------------------------------ End of AIList Digest ******************** From vtcs1::in% Mon Apr 14 00:51:42 1986 Date: Mon, 14 Apr 86 00:51:37 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #86 Status: R AIList Digest Monday, 14 Apr 1986 Volume 4 : Issue 86 Today's Topics: Queries - String Reduction & Imagen Support, Logic & Linguistics - Michael Moss Collection, Speech - Expert Conversationalist, Brain Theory - Comments on Kort's Article ---------------------------------------------------------------------- Date: 10 Apr 86 22:04:26 GMT From: allegra!mit-eddie!think!harvard!seismo!mcvax!ukc!sjl@ucbvax.berkeley.edu (S.J.Leviseur) Subject: String reduction Does anybody have any references to articles on string reduction as a reduction technique for applicative languages (or anything else)? They seem to be almost impossible to find! Anything welcome. Thanks sean sjl@ukc.ac.uk sjl@ukc.uucp sjl%ukc@ucl-cs.edu ------------------------------ Date: Fri, 11 Apr 86 17:28:25 EST From: "Srinivasan Krishnamurthy" <1438@NJIT-EIES.MAILNET> Subject: Vendor Support: SYMBOLICS I need vendor support information for IMAGEN Laser Printer on SYMBOLICS 3640. Any pointers will be greatly appreciated. Please send mail to the address given below: Net: Srini%NJIT-EIES.MAILNET@MIT-MULTICS.ARPA Thanks. Srini. ------------------------------ Date: 10 April 1986 2355-EST From: Es Library@A.CS.CMU.EDU Subject: E&S Library news [Forwarded from the CMU bboard by Laws@SRI-AI.] ** The CMU Library system has purchased the science library of the late Michael Moss in England. The collection consists of over 11 thousand volumes, mostly in Logic, Linguistics and Philosophy of Science and of Language. After many months of efforts on the part of a number of people in the administration and in the CS and Philosophy departments, the collection will be shipped from England later this week. The Moss Collection will substantially strengthen the library services to the newly created Philosophy Department and Program in Computational Linguistics, and will complete, and go beyond, the rebuilding of the Logic collection, of which much was vandalized a few years ago. [...] -- Daniel Leivant ------------------------------ Date: 11 Apr 86 00:06:36 GMT From: tektronix!uw-beaver!ssc-vax!bcsaic!michaelm@ucbvax.berkeley.edu (michael maxwell) Subject: Expert conversationalist :-) One of the problems in AI, specifically in the field of natural language, has been the problem of endowing an artificially intelligent program with the ability to converse in an intelligent manner, observing such principals of conversation as turn taking, empathy with the conversational partner, etc. The problem has been solved! I recently saw in action an expert conversationalist at a local store. The AI program was cleverly disguised as a stuffed bird. However, when you talk to it, it talks back (in a bird language; doubtless an English language program will soon be out, as the chances for making large profits would seem to be much greater.) Before you dismiss this as a simple case of an electronic box that beeps when it detects a sound, let me tell you about some of its capabilities. First, it demonstrates true turn-taking abilities. It does *not* simply listen for sounds and beep back; rather, it waits until you are done talking, and then responds. Second, it does *not* simply beep back; rather, it tailors its response to you. If you talk in an excited voice, it responds in an excited voice; if you talk calmly, it uses a much more subdued response. It tailors both its pitch and speed of speech to your mood as well. Genuine empathy! Think of the possibilities; you could hook it up in place of your phone answering machine to respond to all the carpet cleaning, chimney sweeping, and donate-to-charity-X calls that you get! More relevant to this net, you could hook up your favorite implementation of Eliza + a speech generation device, and have a true Rogerian psychologist at your beck and call. I think I'll buy some stock in this company... -- Mike Maxwell Boeing Artificial Intelligence Center ...uw-beaver!uw-june!bcsaic!michaelm ------------------------------ Date: 9 Apr 86 03:08:09 GMT From: tektronix!orca!tekecs!mikes@ucbvax.berkeley.edu (Michael Sellers) Subject: Re: Computer Dialogue and *bigtime misinformation* (kind of long) > Joseph Mankoski writes a thought provoking article on whether > survival logic in NASA computers has any connection to human > survival instincts wired into to our brains from birth. I would have to agree there is *some* connection, but I would put it on the level of the survival instincts of planaria or hydra, not (even infant) humans. We are much, much more complex than that. > I have been pondering this question myself. [...] > > Joseph asks for a theory of feelings. As it happens, I just wrote > a brief article on the subject, which may or may not be suitable > for publication after editorial comment and revision. Just for > the hell of it, let me append the article and solicit comments > from netters interested in this topic. Okay, here goes. Its real difficult for me to keep this from becoming a big flame-out (this is the second time I've tried to respond; this time I won't kill it before it posts). While I'm sure Barry meant well (unless this is just badly written satire -- which I would find hard to believe), the following article is basically a big pile of misinformation and what seems to be idle conjecture. The "acknowledgements" at the end only serve to give this article a legitimacy it does not deserve by claiming nonspecific sources and thanking PhD types (what are these folks doctors *of*, Barry?). This isn't meant as a personal flame; its just that I've seen so much hype/misinformation/ crapola about the brain/mind/AI recently that when I saw this I couldn't keep quiet. It is possible, I suppose, that a large part of neuroscience has completely turned around in the last six months or so...but I doubt it. If this is true, please excuse my comments as the ravings of an old-worlder. Of course, I'd have to see your sources before believing you. I'd be *glad* to refer you to mine. For brevity (ha!), I haven't re-posted Barry's entire article; nor have I noted/flamed all the things I found objectionable. Some of the assertions in this article, however, could not be ignored. > ==================== Article on Feelings ======================== > > A Simplified Model of the Effects of Perceived Aggression > in the Work Environment > > Barry Kort > > Copyright 1986 > > Introduction > > [...] > > The effects that I wish to investigate are not the > behavioral responses, but the more fundamental internal body > sensations or somatic reactions which lie behind the > subsequent behavioral response. [...] > > A Model of Nature of Aggressive Behavior > > It has been said that civilization is a thin veneer. > Underneath our legacy of some 5000 years of civilization > lies our evolutionary past. Deep within the human brain one > can find the vestiges of our animal nature-the old mammalian > brain, the old reptilian brain. Of principal interest here > are two groups of structures responsible for much of our > "wired-in" instincts. Not quite instincts. The basal ganglia, which make up most of what is sometimes called the old mammalian brain (itself enclosing what some call the R-complex, or reptilian brain) mainly govern biological drives and needs such as hunger, thirst, sex, etc. The main governor of these is the hypothalamus, working in tandem with the pituitary. The thalamus, amygdala, and several other nuclei also contribute to these drives, and have some part in our emotional responses including fear, anger, happiness, nervousness, etc. But these are not insticts nor instinctive. > The cerebellum is responsible for much of our risk-taking, > self-gratifying drives, including the aggressive sex drives. > It is the cerebellum that says, "Go for it! This could be > exciting! Damn the torpedoes, full speed ahead." I couldn't believe this when I first read it. I would still like to believe that Barry mistyped or misread this. The cerebellum (the wrinkled-looking thing that hangs under the back of the cerebrum) play *ABSOLUTELY NO PART* in our rational, cognitive, or emotive behavior!!! What it does do is play a major role in coordinating complex motor actions, such as tying your shoes or dancing the foxtrot (especially, it seems, learned and often repeated actions such as these, as opposed to one-time actions like climbing a tree). I can't imagine where you got this piece of information, Barry. It sounds like it came out of Nat'l Enquirer University. The aforementioned hypothalamus does play a large part in assertive or aggressive action, though this is mediated by the frontal and parietal portions of the cortex and the amygdala and caudate nuclei (in case you wanted to know :-). > The limbic system, on the other hand, is responsible for > self-protective behaviors. The limbic system perceives the > threats to one's safety or well-being, and initiates > protective or counter measures. The limbic system says, > "Hold it! This could be dangerous! We'd better go slow and > avoid those torpedoes." This rates most of my paragraph above. I've never seen anything about cautious behaviors arising in the limbic system, though I know of no reason why some components of such behavior couldn't begin there. The level of behavior suggested here is way to complex for this stage of the processing. Cognitive overlays of our internal biochemical states make up the majority of what we perceive as emotional states/responses. > Rising above it all resides the neocortex or cerebrum. This > is the "new brain" of homo sapiens which is the seat of > learning and intelligence. It is the part that gains > knowledge of cause and effect patterns, and overrules the > myopic attitude of the cerebellum and limbic system. > -> Occasionally, the cerebral cortex is faced with a novel > | situation, where past experience and learning fail to > | provide adequate instruction in how to proceed. In that > | case, the usual patterns of regulation are ineffective, > | and the behavioral response may revert back to the more > | primitive instincts. | This is an interesting piece of conjecture, and one I've not seen recently. It doesn't seem to likely, however, since we have (evolutionarily) paid dearly for our enlarged cortices. Why would we throw out all our observational & computational power just because a situation doesn't match any previously encountered? This would seem to be a marvelous lack of a very valuable resource. It is likely that when the perceived danger or novelty of a situation is *too* great that all our finely-tuned observational and learning powers are thrown out the window in favor of old tired-and-true methods, but this is not as general as is stated here. > [...] > > Somatic Reactions to Stress > > When an individual is presented with an unusual situation, > the lack of an immediately obvious method of dealing with it > may lead to an accumulation of stress which manifests itself > somatically. For instance, first-time jitters may show up > as a knotting of the stomach (butterflies), signaling fear > (of failure). A perceived threat may cause increased heart > rate, sweating, or a tightening of the skin on the back of > the neck. (This latter phenomenon is commonly known as > "raising of one's hackles," which in birds, causes the > feathers to stand up in display mode, warning off the > threatening invader.) Teeth clenching, which comes from > repressing the urge to express anger, leads to a common > affliction among adult males-temporal mandibular joint > (TMJ). Leg shaking and pacing indicate a subliminal urge to > flee, while cold feet corresponds to frozen terror (playing > 'possum). All of these are variations on the > fight/flight/freeze instincts mediated by the limbic system. > They often occur without our conscious awareness. These are also manifestations of the activation of the sympathetic nervous system, probably by the release of epinephrine (adrenalin) into the blood- stream. This can occur with a variety of different emotions, and is much less specific than we are led to believe here. (The use of analogies from biology and the use of an acronym also bug me in this context, since they also seem to lend legitimacy to what is a not very well thought out supposition.) All of these are the result of bloodflow being directed away from non-vital areas (digestive tract, extremities -- butterflies and cold feet) and toward more vital areas (head and muscles -- facial flush, leg shaking, etc) in addition to other secondary effects of the adrenaline (increased heart/respiration rate, sweating, skin tightening). > [...] A person's awareness of and > sensitivity to such somatic feelings may affect his mode of > expression. The somasthetic cortex is the portion of the > brain where the body stresses are registered, and this > sensation may be the primary indication that a stressor is > present in the environment. A challenge for every > individual is to accurately identify which environmental > stimulus is linked to which somatic response. The somasthetic [portion of the] cortex does more than register body stresses. This is the area where *all* sensory input for the body surfaces is perceived. While stressors in the environment can have somatic effects, these do not have a one-to-one (or even a few-to-a few) correspondence with the area of the body or the type of response given. *ALL* stress, if it is bad enough, will effect your body (I sometimes get the "runs" when things get REAL bad), but this effect is not likely to be consistently manifested in one part of your body or with one single reaction. > Somatic responses such as those outlined above are > intimately connected with our expressed feelings, which > usually are translated into some behavioral response along > the axis from aggressive to assertive to politic to > nonassertive to nonaggresive. This is incomplete at best. It is unrealistic to limit the translation of somatic effects into one spectrum of behavioral states/effects, and vastly oversimplifying the situation as well (some oversimplification is inevitable, but not to the extent that you lose all informational content of the thought!). > The challenge is to find and > effectuate the middle ground between too much communication > and too little. The goal of the communication is to > identify the cause and effect link between the environmental > stressor and the somatic reaction, and from the somatic > reaction to the behavioral response. The challenge is all > the more difficult because the most effective mode and > intensity of the communication depends on the maturity of > the other party. This sounds to me for all the world like a paragraph off of the back of a badly researched pop-psych book. I'm somewhat of theoretic conservative; I don't like to see new and wild theories/models thrown around without proper thought and research behind them. While the sentiment here seems to be good, the assumptions and assertions are a mishmash of misinformation, hopeful conjecture, and psych 101. > Acknowledgements > > The original sources for the ideas assembled in this paper > are too diffuse to pinpoint with completeness or precision. > However, I would like to acknowledge the influence of so > many of my colleagues who took the time to contribute their > ideas and experiences on the subject matter. I especially > would like to thank Dr. John Karlin, Dr. R. Isaac Evan, and > Dr. Laura Rogers who helped me shape and test the models > presented here. Like I said, who are these folks, and what sort of feedback did they give you? While I'm at it, is this article being published? If so, where, and what editor let it pass by?! > Comments are invited. > > --Barry Kort ...ihnp4!houxm!hounx!kort Well, you asked. I'd be more than happy to hear any comments to my comments, and/or to view any sources anyone has. I have them in abundance myself. None of this has been intended as a personal flame. I am just speaking out against what is a glaring example of some of the half-baked theories being slung around today. If you want to attack *my* assertions, go ahead (I'm sure there's room for everybody :-). All personal flames will be sent directly to /dev/uranus without comment. My address is ...ihnp4(etc)!tektronix!tekecs!mikes Mike Sellers "The strength and weakness of youth is that it cannot see its own strength and weakness." ------------------------------ End of AIList Digest ******************** From vtcs1::in% Mon Apr 14 00:51:58 1986 Date: Mon, 14 Apr 86 00:51:52 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #87 Status: R AIList Digest Monday, 14 Apr 1986 Volume 4 : Issue 87 Today's Topics: Philosophy - Wittgenstein & Computer Consciousness ---------------------------------------------------------------------- Date: 9 Apr 86 00:18:00 GMT From: pur-ee!uiucdcs!uiucdcsp!bsmith@ucbvax.berkeley.edu Subject: Re: Natural Language processing You are probably correct in your belief that Wittgenstein is closer to the truth than most current natural language programming. I also believe it is impossible to go through Wittgenstein with a fine enough toothed comb. However, there are a couple of things to say. First, it is patently easier to implement a computer model based on 2-valued logic. The Investigations have not yet found a universally acceptable interpretation (or anything close, for that matter). To try to implement the theories contained within would be a monumental task. Second, in general it seems that much AI programming starts as an attempt to codify a cognitive model. However, considering such things as grant money and egos, when the system runs into trouble, an engineering-type solution (ie, make it work) is usually chosen. The fact that progress in AI is slow, and that the great philosophical theories have not yet found their way into the "state of the art," is not surprising. But give it time--philosophers have been working hard at it for 2500 years! Barry Smith ------------------------------ Date: 8 Apr 86 00:32:00 GMT From: ihnp4!inuxc!iubugs!iuvax!marek@ucbvax.berkeley.edu Subject: Re: Natural Language processing Interestingly enough, similar sentiments to your endorsment of L.W. are strongly voiced with respect to Charles Sanders Peirce, by semioticians. >From what I can surmise about Pericean thought, their thrust (or, trust) appears questionable. I am not implying that this necessarily casts a pall on the Vienna School, but my present inclination is to read the Dead Greats for inspiration, not vindication or ready-made answers. -- Marek Lugowski Indiana U. CS Dept. Bloominton, Indiana 47405 marek@indiana.csnet -------- ``I mistrust all systematizers and avoid them. The will to a system is a lack of integrity'' -- Friedrich Nietzsche (``Twilight of the Idols, or How One Philosophizes with a Hammer'') ``Onwards, hammerheads, bright and dangerous, we're big and strong and we're sure of something'' -- Shriekback (``Oil and Gold'') ------------------------------ Date: Sat, 12 Apr 86 23:06:42 est From: Nigel Goddard Reply-to: goddard@rochester.UUCP (Nigel Goddard) Subject: Re: computer consciousness In article <8604110647.AA25206@ucbvax.berkeley.edu> "CUGINI, JOHN" writes: > >Thought I'd jump in here with a few points. > ... > >3. Taking up the epistemological problem for the moment, it >isn't as obvious as many assume that even the most sophisticated >computer performance would constitute *decisive* evidence for >consciousness. Briefly, we believe other people are conscious >for TWO reasons: 1) they are capable of certain clever activities, >like holding English conversations in real-time, and 2) they >have brains, just like us, and each of us knows darn well that >he/she is conscious. Clearly the brain causes/supports >consciousness and external performance in ways we don't >understand. A conversational computer does *not* have a brain; >and so one of the two reasons we have for attributing >consciousness to others does not hold. > It is not just having a brain (for which most of us have no direct evidence anyway), but having a head, body, mouth, eyes, voice, emotional sensitivity and many other supporting factors (no one of which is *necessary*, but the more there are there the better the evidence). I guess a brain is necesary, but were one to come across a brain with no body, eyes, voice, ears or other means for verifying its activity, would one consider it to be conscious ? Personally I think that the only practical criterion (i.e. the ones we use when judging whether this particular human or robot is "conscious") are performance ones. Is a monkey conscious ?. If not, why not ? There are people I meet who I consider to be very "unconscious", i.e. their stated explanations of their motives and actions seem to me to completely misunderstand what I consider to be the *real* explanations. Nevertheless, I still think they are conscious entities, and the only way I can rationalize this paradox is that I think they have the ability to learn to understand the *real* reasons for their actions. This requires an ability to abstract and to make an internal model of the self, which may be the main factors underlying what we call consciousness. Nigel Goddard ------------------------------ Date: 8 Apr 86 09:57:10 GMT From: hplabs!qantel!lll-lcc!lll-crg!styx!lognet2!seismo!ll-xn!topaz!harvard !h-sc1!pking@ucbvax.berkeley.edu Subject: Re: Computer Dialogue In all this discussion of "feelings," "survival instinct," and "consciousness," one point is being overlooked. That is, can you really say that a behavioral reaction (survival instinct) is a feeling if the animal or computer has no consciousness? Joseph Mankoski asked whether or not one could say that the shuttle's computers were displaying a form of "programmed survival instinct." I think that the answer is yes. This does not mean that shuttle missions were aborted because the computer wanted to save itself. Biologists, however, are quick to point out that cats run away from dogs not because they want to save themselves, but because the sight of a dog triggers a cat's flight (abort) mechanism. The net effect of the cat's behavior is to increase its chances of survival, but the cat (and the shuttle's computer) has no "desire to survive." But we, as humans, DO have a desire to survive, don't we? When faced with danger, we do everything in our power to avoid it. The difference is that we are conscious of our attempts to avoid danger, even if we do not understand them. "Why did you run away from that snake," someone might ask. "To escape possible injury," we rationalize. The more truthful answer, however, is "It just happened -- it was the first thing that came to mind." But what of the sensation of fear that comes over us in such situations? "Fear" is just a name we have given to the sensation of anxiety coupled with avoidance-behavior. For the most part, we are observers of our own behavior (and our own thoughts, for that matter: introspection). Sure, we have control over our instinctual tendencies, but not as much as we would like to think. Witness the acrophobic "unable" to climb a fire-escape. Why would courage be such an envied quality if it weren't so hard to defeat one's instinctual (intuitive) reactions. Unfortunately, gut-feeling tendencies can backfire, as in the case of drug addiction. In this case, the emotional mind sets the goal ("get drugs") and the rational mind does what it can to get satiate the emotional mind despite knowledge of the damage being done. Phobias aren't so desirable either. What I'm getting at is that "desires" and "feelings" are how we experience the state of our mind, just as colors are the way we experience light frequency and pain is the way we experience tissue damage. To say a computer has feelings is incorrect unless the computer is AWARE of its behavior. You could possibly say that the shuttle's computer aborted the mission to prevent it's own death (i.e. it felt fear) if one of the sensory inputs to the computer was the fact that it was entering the abort- state. The same argument could be made for consciousness. That to be conscious is to be aware of one's own thought process and state of mind (a sixth sense?). Computers (and Barry Kort's gigantic telephone switching system) are not conscious. While they receive input from the various "senses" (telephone exchanges, disk- drives, users), they receive no information about themselves. One could say that a time-sharing system that monitor its own status is "conscious" but this is a very limited consciousness, since the system cannot construct an abstract world-model that would include itself, a requirement for personal identity. If a computer could compile sensory information about itself and the world around it into an abstract model of the "world," and then use this model to interact with the world, then it would be conscious. Further, if it could associate pieces of its model to words, and words to a grammar, then it could communicate with people and let us know "what it's like to be a computer." ------- I would appreciate any reactions. Paul King UUCP: {seismo,harpo,ihnp4,linus,allegra,ut-sally}!harvard!h-sc4!pking ARPA: pking@h-sc4.harvard.EDU BITNET: pking@harvsc4.BITNET ------------------------------ Date: 9 Apr 86 23:18:21 GMT From: decvax!linus!philabs!cmcl2!seismo!ll-xn!cit-vax!trent@ucbvax.berkeley .edu (Ray Trent) Subject: Re: Computer Dialogue In article <1039@h-sc1.UUCP> pking@h-sc1.UUCP (paul king) writes: >"consciousness," one point is being overlooked. That is, can you >really say that a behavioral reaction (survival instinct) is a >feeling if the animal or computer has no consciousness? Please define this concept of "consciousness" before using it. Please do so in a fashion that does not resort to saying that human beings are mystically different from other animals or machines. Please also avoid self-important definitions. (e.g. consciousness is what humans have) >is to increase its chances of survival, but the cat (and the >shuttle's computer) has no "desire to survive." The above request also applies to the term "desire". >difference is that we are conscious of our attempts to avoid ... >"It just happened -- it was the first thing that came to mind." Huh? This pair of sentences seems to say that your definition of "consciousness" is that consciousness is "the first thing that [comes] to mind." I don't think that split second decisions are a good measure of what most people call consciousness. > [two paragraphs that seem to reinforce the idea that > consciousness has much to do with "gut-level reactions" and > "instincts"] >What I'm getting at is that "desires" and "feelings" are how we My definition of these concepts would say that they "are" the actions that a life process take in response to certain stimuli. >tissue damage. To say a computer has feelings is incorrect >unless the computer is AWARE of its behavior. You could possibly No, to say that a computer has self-awareness is to say that it is AWARE of its feelings. Unless, of course, this is yet another self-defined concept. >say that the shuttle's computer aborted the mission to prevent >it's own death (i.e. it felt fear) if one of the sensory inputs >to the computer was the fact that it was entering the abort- >state. [reductio ad absurdum(sp?)] You could possibly say that a human entered abort mode (felt fear) if one of its sensory inputs was the fact that it was entering abort mode (feeling fear). >telephone switching system) are not conscious. While they receive >input from the various "senses" (telephone exchanges, disk- >drives, users), they receive no information about themselves. One Telephone systems receive no inputs about themselves? What about routing information derived from information the system has about its own damaged components? >the system cannot construct an abstract world-model that would >include itself, a requirement for personal identity. Here is a simple program to construct an abstract world-model that includes the machine: main() { printf("I think, therefore I am.\n"); } Try to convince me that humans do something fundamentally different here. (seriously) >If a computer could compile sensory information about itself and >the world around it into an abstract model of the "world," and >then use this model to interact with the world, then it would be >conscious. Further, if it could associate pieces of its model to >words, and words to a grammar, then it could communicate with >people and let us know "what it's like to be a computer." I give as example the relational database program. It collects sensory information about the world into an abstract model of the "world" and then uses this model to interact with the world. Is it therefore conscious? I don't think so. (how self-referential of me) If fact, I will go further...such a program associates pieces of its model to words and words into a grammer, and with the appropriate database, could indeed let us know "what it's like to be a computer," but I don't think that most people would call it conscious. >I would appreciate any reactions. Ask, and you shall receive. -- ../ray\.. (trent@csvax.caltech.edu) "The above is someone else's opinion only at great coincidence" ------------------------------ Date: 13 Apr 86 17:25:09 GMT From: dali.berkeley.edu!regier@ucbvax.berkeley.edu (Terrance P. Regier) Subject: Re: Computer Dialogue trent@csvax.caltech.edu writes: > Here is a simple program to construct an abstract world-model > that includes the machine: > > main() > { > printf("I think, therefore I am.\n"); > } > > Try to convince me that humans do something fundamentally > different here. (seriously) ^^^^^^^^^ Descartes' famous assertion was the result of a period of admirably honest introspection: After allowing himself to doubt the veracity of his beliefs, senses, etc., he found that some things (well, at least one thing) CANNOT be doubted. I think, therefore I am. Your admittedly concise and elegant program fails to capture the integrity and awareness of self implicit in the statement. It is closer in spirit to an involuntary burp. -- Terry ------------------------------ End of AIList Digest ******************** From vtcs1::in% Mon Apr 14 04:59:48 1986 Date: Mon, 14 Apr 86 04:59:43 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #88 Status: R AIList Digest Monday, 14 Apr 1986 Volume 4 : Issue 88 Today's Topics: Seminars - DADO/TREAT: Parallel Execution of Expert Systems (UTexas) & Inverse Method of Establishing Deducibility (SRI) & Perspectives, Prototyping, and Procedural Reasoning (CMU) & Improving Planning Efficiency (Rutgers) & Anaphora: Events and Actions (UPenn), Conference - AI Impacts at FAA, Date Change & Discourse Analysis & AI and Automatic Control ---------------------------------------------------------------------- Date: Wed, 9 Apr 86 10:06:20 CST From: Rose M. Herring Subject: Seminar - DADO/TREAT: Parallel Execution of Expert Systems (UTexas) University of Texas Computer Sciences Department COLLOQUIUM SPEAKER: Daniel Miranker Columbia University TITLE: DADO & TREAT: A Sytem for the Parallel Execution of Expert Systems DATE: Thursday, April 10, 1986 PLACE: TAY 3.144 TIME: 11:00-12:00 noon The development of expert computer programs has moved out of the research lab and into a quickly developing commercial field. The development of computer architectures that are better suited for executing these programs has recently come into the forefront of computer architecture research. Indeed, a new term, fifth generation computers, has been coined to describe these ar- chitectures. This talk will describe the architecture and software systems of a recently completed parallel computer, the DADO machine, designed to accelerate expert systems written in produc- tion system form. The talk will also describe a new production system matching algorithm that, although motivated by the algo- rithmic requirements of parallel computing, has been shown to be better than the RETE match (the currently accepted best produc- tion system algorithm), even in a sequential environment. COFFEE AT 10:30 in TAY 3.128 ------------------------------ Date: Thu 10 Apr 86 11:23:55-PST From: LANSKY@SRI-AI.ARPA Subject: Seminar - Inverse Method of Establishing Deducibility (SRI) WHAT IS THE INVERSE METHOD? Vladimir Lifschitz (VAL@SAIL) Stanford University 11:00 AM, MONDAY, April 14 SRI International, Building E, Room EJ228 (new conference room) In 1964, the same year when J. A. Robinson introduced the resolution rule, a Russian logician and philosopher, Sergey Maslov, published his four-page paper, "An Inverse Method of Establishing Deducibility in Classical Predicate Calculus". Maslov's method is based on a major discovery in proof theory which has remained largely unnoticed by logicians. The method does not require that the goal formula be written in clausal or even prenex form, and there may exist a possibility of applying it to non-classical systems (e.g., modal). Computer programs based on the inverse method are reported to be comparable, in terms of efficiency, to those using resolution. The inverse method has been also applied to solving new special cases of the decision problem for predicate logic, and it can serve as a uniform approach to solving almost all known solvable cases. In this talk I explain the idea of the inverse method on a simple example. Note to visitors: SRI now has stricter security rules and won't allow people to just walk up to the AIC. If you have any problems being admitted, please call either me (Amy Lansky -- x4376) or Margaret Olender (x5923). ------------------------------ Date: 10 April 1986 1536-EST From: Betsy Herk@A.CS.CMU.EDU Subject: Seminar - Perspectives, Prototyping, and Procedural Reasoning (CMU) Speaker: David A. Evans, Dept. of Philosophy, CMU Date: Wednesday, April 23 Time: 11:30 - 1:00 Place: 5409 Wean Hall Title: Perspectives, prototyping, and procedural reasoning In the special task of developing a consultation and tutoring facility for the CADUCEUS expert system, it is necessary to identify several perspectives over detailed diagnostic information, which can be organized into meta-level knowledge structures that reflect explicit procedures, contexts, and pragmatics, associated with the task of explaining and justifying diagnostic inferences. Such structures offer concrete interpretations of notions such as prototypes (taken from cognitive science) and suggest constraints that can be exploited in controlling discourses and procedural reasoning. ------------------------------ Date: 10 Apr 86 13:18:38 EST From: PRASAD@RED.RUTGERS.EDU Subject: Seminar - Improving Planning Efficiency (Rutgers) Machine Learning Colloquium REAPPR: Improving planning efficiency via local expertise and reformulation Bresina, J.L., Marsella, S.C., and Schmidt, C.F. Rutgers University 11 AM, April 22, Tuesday #423, Hill Center Abstract We discuss planning within the problem reduction paradigm. Within this paradigm, a key issue is handling subproblem interactions. We point out the advantages of problem reduction over goal reduction (which characterizes most previous planning systems). We introduce an implemented planning system - REAPPR - which extends the problem reduction paradigm to capture and efficiently utilize expert planning knowledge. The features of REAPPR include: (i) potential parallelism, (ii) local control information, (iii) flexible problem reduction, and (iv) reformulations. ------------------------------ Date: Fri, 11 Apr 86 12:19 EST From: Tim Finin Subject: Seminar - Anaphora: Events and Actions (UPenn) Forwarded From: Ethel Schuster on Thu 10 Apr 1986 at 20:22 TOWARDS A COMPUTATIONAL MODEL OF ANAPHORA IN DISCOURSE: REFERENCE TO EVENTS AND ACTIONS Ethel Schuster Abstract When people talk or write, they refer to things, objects, events, actions, facts and/or states that have been mentioned before. Such context-dependent reference is called anaphora. In general, linguists and researchers working in artificial intelligence have looked at the problem of anaphora interpretation as that one of finding the correct antecedent for an anaphor--that is, the previous words or phrases to which the anaphor is linked. Lately, people working in the area of anaphora have suggested that in order for anaphors to be interpreted correctly, they must be interpreted by reference to entities evoked by the previous discourse rather than in terms of their antecedents. This work describes the process of dealing with anaphoric language when the reference is to events and actions. It involves four issues: (i) what aspects of the discourse give evidence of the events and actions the speaker is talking about, (ii) how actions and events are represented in the listener's discourse model, (iii) how to identify the set of events and actions as possible choices, and (iv) how to obtain the speaker's intended referent to an action or event from a set of possible choices. Anaphoric forms that are used to refer to actions and events include sentential-it, sentential-that pronominalizations as well as do it, do that, and do this forms. Their interpretations can be many and because of that, they cannot be understood only on linguistic grounds but on models of the discourse. So, I will concentrate on developing the four previously mentioned issues along with other mechanisms that will provide us with better tools for the successful interpretation of anaphoric referents to actions and events in discourse. April 16, 1986 11 am Moore 129 (Faculty Lounge) Advisor: Bonnie Webber Committee: Tim Finin, Chair Aravind Joshi Ellen Prince (Linguistics Dpt.) Tony Kroch (Linguistics Dpt.) Candy Sidner (BBN) ------------------------------ Date: 11 Apr 86 23:55:32 GMT From: hplabs!sdcrdcf!burdvax!blenko@ucbvax.berkeley.edu (Tom Blenko) Subject: Conference - AI Impacts at FAA, Date Change ARTIFICIAL INTELLIGENCE IMPACTS WORKSHOP presented by AMERICAN COMPUTER TECHNOLOGIES, INC. >>> June 11-13, 1986 <<<-- NOTE change of date FAA Technical Center Atlantic City Airport, New Jersey This is a mildly-technical workshop for marketing, planning and manufacturing professionals who are interested in artificial intelligence. Workshop emphasizes marketing data, competitive analyses, planning information, financials, opportunities and contraints, etc., from a world-wide survey of businesses and governments involved in AI. Information can be obtained from: American Computer Technologies, Inc. 237 Lancaster Avenue, Suite 255 Devon, PA 19333 Attn: Ms. Carol Ward (215) 687-4148, and/or Ms. Pat Watts of the Federal Aviation Administration Technical Center: (609) 484-6646. (This information is being posted for a friend: please respond to the address given above). ------------------------------ Date: 10 Apr 86 16:55:09 GMT From: decvax!mcnc!akgua!ganehd!anv@ucbvax.berkeley.edu (Andre Vellino) Subject: Conference - Discourse Analysis First Ad Hoc Conference on Discourse Analysis April 24-25, 1986 138 Tate Hall University of Georgia Athens, Georgia Thursday, April 24 9 a.m. Rainer Bauerle (University of Tubingen) "Nominalizations, Event Anaphora, and Order of Events in a DRT-framework" 10.30 a.m. Coffee Break 11 a.m. Nirit Kadmon (University of Massachusetts, Amherst) "Maximal Collections, Specificity, and Discourse Anaphora" 12.30 p.m. Lunch Break 2 p.m. Hans Kamp (University of Texas, Austin) "Plural Anaphora and Plural Determiners" Friday, April 25 9 a.m. Craige Roberts (University of Massachusetts, Amherst) "Modal Subordination and Pronominal Anaphora in Discourse" 10.30 a.m. Coffee Break 11 a.m. Michael Covington (University of Georgia, Athens) "Modelling Implicature with Defeasible Logic" 12.30 p.m. Lunch Break 2 p.m. Barbaree Partee (University of Massachusetts, Amherst) "Nominal and Temporal Anaphora" Advanced Computational Methods Center University of Georgia Athens, Georgia 30602 For further information contact Marvin Belzer (404) 542-5110 ------------------------------ Date: 11 Apr 1986 19:22:05 EST From: ALSPACH@USC-ISI.ARPA Subject: Conference - AI and Automatic Control Dr. Andrews National Aeronautics & Space Administration Ames Research Center San Jose, CA Dear Dr. Andrews: Per your note to AI-LIST on April 1, regarding the synergism between the fields of artificial intelligence and automatic control, I would like to bring your attention to the American Control Conference to be held in Seattle from June 18-20 this year. The American Control Conference is sponsored by the American Automatic Control Council, which is a council consisting of member organizations which include the AIAA, AICHE, ASME, IEEE, ISA, and SCS. The ACC is the U.S. representative to IFAC (the International Federation of Automatic Control). In addition, other engineering societies, such as Automation Engineers, participate. This is the largest conference on control held in the United States, and is multidisciplinary. It has been held for a number of years. Looking at this year's program, it is clear that your idea of exploring the common ground between control and artificial intelligence is already seriously in progress. Out of 68 sessions, there are seven sessions whose major themes are artificial intelligence and control, or robotics and control. First, on Wednesday A.M., there is a session on Artificial Intelligence in Process Control. The Chairman is R. Moore of LISP Machines, Inc., and a number of national and international experts are talking about this very interesting topic. In parallel with this session on Wednesday A.M., there is a session entitled Robotics that explores many aspects of robotics control. The Chairman of this session is Jason Speyer from the University of Texas at Austin, and it will be co-chaired by M. Railey from the University of Akron. On Wednesday P.M., there is a session entitled Artificial Intelligence Applications in Sensor Fusion and Command and Control. This session is chaired by Dr. S. Brodsky, Sperry Corporation, and addresses some very interesting work in the area of artificial intelligence applications to sensor fusion and command and control. Typical papers from this session include J. Flynn of DARPA on "Carrier Based Threat Assessment", J. Delaney of Stanford talking on "Multisensor Report Integration Using Blackboards", and M. Grover and M. Stachnick of Advanced Decision Systems discussing "Overlooked and Unconventional AI Techniques for Command and Control". A number of other very interesting papers are in this session. On Thursday A.M., there is a session on 4D Aircraft Guidance and Expert Traffic Management, which is chaired by A. Chakravarty of Boeing Commerical Airplane Company and co-chaired by R. Schwab, also of Boeing. An exemplar paper in this session is "Time-Based Air Traffic Management Using Expert Systems" by L. Tobias and J. Scoggins of NASA Ames Research Center. Running in parallel on Thursday A.M., is a specialist session on Direct Drive Robot Arms. This is chaired by J. Slotine of Massachusetts Institute of Technology and co-chaired by H. Asada of Kyoto University, Japan. Another general session on Artificial Intelligence is to be held on Thursday P.M., chaired by J. Birdwell from the University of Tennessee and co-chaired by G. Allgood, Oak Ridge National Laboratory. A number of excellent papers include: "Domains of Artificial Intelligence Relevant to Systems", by J. Birdwell and J. Crockett, University of Tennessee, and J. Gabriel of Argonne National Laboratory; "Knowledge Representation by Scripts in an Expert Interface" by J. Larsson and P. Persson of Lund Institute of Technology; and "An Expert System to Control a Fusion Energy Experiment" by R. Johnson, et al., from Lawrence Livermore Laboratories. On Friday A.M., there is a session on Aerospace and Robotics Applications of Nonlinear Control, chaired by F. Fadali, University of Nevada-Reno and co-chaired by T. Dwyer, University of Illinois. In parallel on Friday A.M., there is a session on Robot Tracking Control chaired by George Saridis of Rensselaer Polytechnic Institute. On Friday P.M., there is a session on Multitarget Tracking and Data Association chaired by C. Chong, Advanced Information & Decision Systems, and co-chaired by M. Shensa, Naval Ocean Systems Center. This discusses an area that is ripe for artificial intelligence applications and, for example, includes a paper entitled "An Expert System for Surveillance Automation" by R. Mucci of BBN Laboratories. Also in parallel with this Friday P.M. session is one on Robot Control chaired by J. Garbini, University of Washington, and co-chaired by C. Nachtigal of Kistler Morse Company. In addition to these sessions, there are a number of papers on artificial intelligence, expert systems and robotics applications scattered throughout a number of other sessions in the program. Also, of interest to people who are interested in the AI List information, there is a one-day tutorial workshop on Monday, June 16, preceding the conference, entitled "Intelligent Control System Design and Analysis". The purpose of this workshop is to introduce control systems engineers and engineering managers to the possibility of using intelligent systems during the design and analysis of control systems. Participants will learn the techniques for building expert systems and will see examples of their use in control system design. This tutorial workshop will be taught by Guy Beale of Vanderbilt University and Charles Buenzli of Gilbarco-Exxon. On Tuesday, June 17, another tutorial workshop will be taught by Roger Brockett of Harvard University and Robert M. Goor of General Motors Research Laboratory. The topic of this workshop will be "Modeling and Control of Robotic Manipulators". The General Chairman for the conference is Dr. Ed Stear, who is Associate Dean of Electrical Engineering at the University of Washington and Head of the Washington Technology Center. It may also be of interest to this community that one of the plenary speakers is Dr. Robert Rankine, Brigadier General, U.S. Air Force and Head of Air Force SDI activities. He will discuss some of the control challenges associated with the SDI Program and with the proposed new hypersonic trans-atmospheric vehicles. All in all, for someone interested in the merging of the fields of artificial intelligence, expert systems and automatic control, this is an excellent conference to attend. There is also a great social program planned for the evenings to allow informal discussions among the attendees. Also, Expo '86 is only a few miles up the road in Vancouver, British Columbia, for those interested in attending this activity before or after the conference. To obtain information regarding registration, please contact the office of Dagfinn Gangsaas, BMAC, P.O. Box 3707, MS 33-12, Seattle, WA 98124, (206) 241-4348. Preliminary programs may be obtained by sending a request to me via Arpanet, c/o ALSPACH (at) USC-ISI or mailing a request to D. L. Alspach, ORINCON Corporation, 3366 N. Torrey Pines Ct., Suite 320, La Jolla, CA 92037. Sincerely, Daniel L. Alspach Program Chairman 1986 American Control Conference BBN Laboratories. ------------------------------ End of AIList Digest ******************** From vtcs1::in% Wed Apr 16 00:49:04 1986 Date: Wed, 16 Apr 86 00:48:58 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #89 Status: RO AIList Digest Tuesday, 15 Apr 1986 Volume 4 : Issue 89 Today's Topics: Bibliography - Recent Articles #4 ---------------------------------------------------------------------- Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Recent Articles #4 %A Richard L. Wexelblat %T Editorial %J SIGPLAN Notices %V 21 %N 3 %D MAR 1986 %P 1 %K AI03 AA17 Queens Problem H03 ADA %X SIGPLAN is having a contest to determine the best solution for the N<=8 Queens problem using concurrency in ADA in a substantive manner. Deadline for submissions is June, 1986. %A Scott Mace %T Ansa Upgrades Paradox, Drops Copy Protection %J InfoWorld %V 8 %N 11 %D MAR 17, 1986 %P 3 %K Paradox AA09 H01 %X Ansa software announced upgrades to its system and has dropped copy protection %T Resources %J InfoWorld %V 8 %N 11 %D MAR 17, 1986 %P 17 %K AI01 H01 Cahners Publishing Company Users Group Expert Systems Strategies %X The New York IBM PC Users Group has announced a special interest group for Expert Systems. The meetings are held March 25. For more info contact NYPC, Suite 614, 30 Wall Street, 10005 (212) 533 NYPC. Cahners Publishing Company produces a newsletter called Expert Systems Strategies. Charter rate is $207, regular rate is $247. Address is Cahners Publishing Co. P. O. Box 59, New Town Branch Boston, MA 02258 (617) 964 3030 %A Cornelius Willis %T The Problems with AI %J InfoWorld %V 8 %N 11 %D MAR 17, 1986 %P 20 %K Insight 1 Level Five Research AI01 T03 H02 AT14 AT12 %X The director of marketing for Level Five Research writes that the reason that artificial intelligence has not been embraced by corporate MIS directors is that the charges for Lisp machines and knowledge engineers are way too high. He also claims that his product, Insight 1, is "the most widely used knowledge engineering tool in the world." %A Peggy Watt %T Ansa Move Woos Corporate Users %J ComputerWorld %D MAR 17, 1986 %V 20 %N 11 %P 1+ %K H01 AA09 Paradox Ashton-Tate AT04 %X Ansa Software's Paradox has been approved by only about two dozen large-account evaluators. In November and December, they sold 1,449 copies against 13,156 copies of Ashton Tate's DBASE products. In January, it was 813 copies of Paradox against 6,154 copies of DBASE. Ansa is investigating the micro-to-mainframe data base file exchange. %A Douglas Barney %T AI-Based Financial Systems Allocates Assets Based on Goals %J ComputerWorld %D MAR 17, 1986 %V 20 %N 11 %P 6 %K H02 AI01 AA06 First Financial Planner Services Plan Power Xerox personal planning %X discussion of First Financial Planner Services, Plan Power which is an expert system that performs personal financial planning. It has 6000 rules. %A Elisabeth Horwitt %T AI Integration Gets a Shot in the ARm as Vendors Link Products %J ComputerWorld %D MAR 17, 1986 %V 20 %N 11 %P 47+ %K Harvey Newquist Symbolics H03 SNA Texas Instruments Explorers Gould LMI SUN Apollo AT16 %X Symbolics has announced a link to IBM mainframes via SNA. Flavors Technology brought out a high speed bus to bus link between Lisp Machine, Inc. machines or Texas Instruments Explorer's. and Gould, Inc. superminicomputers. It costs $36,000. Texas Instruments plans to integrate their Explorer with SUN and Apollo. %A Eric L. Schwartz %A Bjorn Merker %T Computer-Aided Neuroanatomy: Differential Geometry of Cortical Surfaces and an Optimal Flattening Algorithm %J IEEE Computer Graphics and Applications %D MAR 1986 %V 6 %N 3 %P 36-44 %K AA10 AI08 %X describes the mapping of the visual field on the visual cortex of the monkey %T Apollo, TI to TIE Network, Workstation %J Electronic News %V 32 %N 1593 %D MAR 17, 1986 %P 18+ %K SUN LMI Flavor Common LISP Compact Lisp Machine H02 AT02 AT16 %X [Much of the material in this article was reported recently elsewhere in AILIST; only new stuff is in this abstract] Apollo will be selling a $3,500 Common Lisp. The link between Apollo and TI will be made using Apollo's Open Systems Tookit and TI's Flavor package. Apollo also hopes to use TI's single chip LISP machine. LMI's marketing director said that it is has always been the position of his company that Lisp machines and LISP cannot survive alone. He predicted that alignment of TI, SUN and Apollo will not affect LMI. Furthermore, he predicts that the single chip LISP machine development effort at TI will take at least 12 months. %A Richard H. McSwain %A Robert W. Goutld %T Taking the Fatigue Out of Fracture Surface Analysis %J Metal Progress %D MAR 1986 %V 129 %N 4 %K AA05 Metallurgy Failure Analysis AI06 striation Fourier Transform %X Describes use of the Fourier transform method to analyze the fracture surface of a material failing from fatigue. [When a material is repeatedly subjected to changes in stress, it may fail from fatigue. This is even true when the maximum load is well below the limit which would cause failure if it was applied in a steady state condition. When this happens, a characteristic striation appears on the fracture surface. This can be viewed with Scanning Electron Microscopy or even with the naked eye or magnifying glass. LEFF] %A Barry Meier %T Robot Subs Begin to Surface as Versatile Exploration Tools %J Wall Street Journal %D MAR 7, 1986 %V 78 %N 46 %P 19 %K Deep Ocean Engineering Company International Submarine Engineering Ltd. AA03 AI07 AA18 AA19 GA04 %X Describes some uses and research therein for robot submarines: Canadian Oceanographers will use one to hunt for oil below the icecap. It will be dropped through a hole in the ice 1000 miles from the North Pole. It will then navigate in a grid like pattern in a ten square mile area mapping the bottoms. Sonar will help the sub avoid icebergs. The thing is being build by International Submarine Engineering Ltd. .sp 1 R&D exists in applications of submarines to prospecting, repair of oil installations, perform rescue and recovery missions, and engage in spying. Work is done on developing sensors based on sound, AI systems to help robots react to currents and fiber optics for exchange of data with mother ships. Deep Ocean Engineering Company has developed a system of sensors to detect the weight and composition of objects under water. They are using tones to inform the operator of what the robot has in its arms. They are also perfecting AI to the point that submarines can be free-swimming. NASA is funding some of this work since they hope to apply the results to space travel. %T Advertisement %J BYTE %D APR 1986 %P 284 %V 11 %N 4 %K Solution Systems TransLisp T01 H01 AT01 %X Lisp for IBM PC for only $75.00 It is a 230+ function subset of Common Lisp and has MSDOS interface and graphics. (Solution Systems also sells the BRIEF editor %T Star Wars Divides A Campus %J BusinessWeek %D March 10, 1986 %P 82-86 %N 2936 %K Carl Hewitt MIT %X Discusses reactions of MIT people to SDI funding Carl Hewitt has decided to apply for SDI funding. The AI Lab at MIT received 55 percent of its 8 million dollar budget from the defense department. %A Michael Lesk %T Writing to be Searched: A Workshop on Document Generation Principles %J SIGIR Forum %V 19 %N 1-4 %D WINTER 1986 %P 9-14 %K Cucumber Information Knowledge Systems AI02 A08 AA14 %X "It is now possible to design full-text retrieval systems that accept conventional docuements and questions in natural English, and then retrie ve documents ofr passages from documents that probably answer the questions." Cucumber Information Systems and Knowledge Systems, Inc. sell such systems. A high degree of grammatical variation does not seem important to produce natural effects in short paragraphs (as evidenced by Karen Kukich's stock market report generator)" "Syntax is much less important for retrieval than semantics; you need to know what the words mean more than you need to know their relationship." "Editing manuals to make them suitable for machine translation, requiring simple translation, has turned out to make them better in the original language as well." %A Susanne M. Humphrey %T Automated Classification and Retrieval Program: Indexing Aid Project %J SIGIR Forum %V 19 %N 1-4 %D WINTER 1986 %P 16-17 %K AA14 AI02 AA01 %X Lister Hill Center of the National Library of Medicine is developing this system to generate indices consistent with those normally used by MEDLINE. They are using a frame based system. %A Frank Tansey %T Guru's Power Cuts Out the Competition %J Infoworld %V 8 %N 12 %D MAR 24, 1986 %P 14 %K AI01 AA09 AA06 university administration residency T03 H01 AT17 %X This is a review of GURU, an expert system tool that interfaces with MDBS's Knowledge Man. It supports up to 3000 rules, forward and backward chaining, inexact reasoning. The system also includes a text processor, graphics, spreadsheet, graphics and telecommunications. The system received a rating of 5.8 out of 10 with very good for performance and ease of use, satisfactory for documentation and value, It takes 1700 pages of documentation to describe the system. .sp 1 As of much interest as the review itself are the two systems that were two expert systems developed using GURU described in this review. The first was a system to assist in determining residency status of students for the California Universities for the purpose of determining tuition. The final expert system was judging cases with the experience of a person with six months to one year in evalulating such matters. The system was already able to impress people in the field with only fifty rules. They also wrote an expert system to do personal financial planning. This took 300 rules and embodied the entire expertise of the person writing the software. .sp 1 [I read elsewhere that MDBS has sold $6,000,000 of these packages since they came out. They cost $3,000 each. MDBS is known for Knowledge Man, probably the most powerful relational data base for micros. Keep in mind that InfoWorld tends to downgrade systems if they weren't written so as to be used by people lacking knowledge or aptitude for computers and thus most readers of AILIST would have a higher opinion of the package than 5.8 out of 10. LEFF ] %T Chairman Resigns From Automatix %J Electronic News %D Mar 24, 2986 %V 32 %N 1594 %K AT16 AT11 AI07 %X Philippe Villers resigned as chairman of robotics maker Automatix. Automatix has yet to turn a profit and lost $5,594,000 in 1985 and $14,193,000 in 1984. %A Michael Bucken %T Symbolics Starts VAR Program for 36-BIT Processing Systems %J Electronic News %D Mar 24, 2986 %V 32 %N 1594 %K AA04 H02 AT16 %X Symbolics has signed its first VAR contract with ICAD which is developing an engineering design software package. 40 percent of Symbolics customers are using the system for applications other than artificial intelligence. The system has sold about 2000 processors. %A Criag Stedman %T Management Seeking GCA Robotics Group %J Electronic News %D Mar 24, 2986 %V 32 %N 1594 %K Industrial Systems Group AT16 AI07 %X The management of the robotics division of GCA Corporation is trying to arrange a leveraged buyout. The division has lost 10 to 15 million dollars on sales of about $35 million. %A Tony Baer %T Finding the Titanic %J Mechanical Engineering %V 108 %N 3 %D MAR 1986 %K Jason Angus control chattering ARGO submersible salvage underwater AI06 AI07 %X One of the problems in underwater vision is backscattering from the light source of suspended particles. A good way of fighting this problem is to mount the light source away from the camera. The new lighting system on the Angus has yielded readable images of areas about as large as a city block. A system called Jason is being developed that will mount in the ARGO submersible. This system will be self-propelled and have its own manipulator arm. However, it is NOT going to need artificial intelligence [Emphasis mine, Leff] %A J. Houseley %T Getting a Grip on Sensors %J IEEE Spectrum %V 23 %N 4 %D APR 1986 %P 8 %K tactile sensors AI07 AT12 AT13 %X This is a comment by an article by Paolo Darlo and Danilo De Rossi of August 1985 on the subject of using tactile sensors in gripping objects in robotics. In a human being picking up an egg, the human being would apply enough force to prevent the weight of the egg from deflecting it. In gripping a hammer, friction between the hammer is used. There is a comment on the role of learning in applying the right amount of force to adjust for the change when the hammer impacts the nail. (There is also a response by the author. %T Advertisement %J Byte %D MAR 1986 %V 11 %N 3 %K AT03 H01 T03 AA18 AT01 Thunderstone Corporation Clarity Software Comprehension Logic-Line %X Add for Thunderstone Corporation's Logic Line 1 ($250), Logic-Line 2 ($400.00) and Comprehension ($75.00) for the IBM PC It is not clear from the advertisement what LOGIC-LINE1 and LOGIC-LINE2 actually do. Comprehension is supposed to enable a person to diagnose their weakness in a given discipline. Some quotes from this advertisement: "Our success has effectively stompted the mortal spit out of the brain damaged geeks whose rancid cells have been polluting the gene pool of legitimate AI professionals." "LOGIC-LINE1, a major breakthrough in sub-cognitive mathematics, distills the DNA/RNA like analog to any writer's thought processes. It allows you to search any textbase for actual concepts and inference patterns unique to that writer. In other words, even though Einstein may never have had a single thought about ecology, you can apply his thinking patterns to solving ecological problems!" "And at its highest level? You just might use Thunderstone tools to save the free world, again. That's right: Again! LOGIC-LINE 2 began with the mathematics of possibilistic analysis and recursion (developed by men like Alan Turing and Norbert Weiner) that directly led the Wellington College team to breaking the German naval codes in World War II." %A Melissa Calvo %T Japanese Firms Granted License by Compuserve %J InfoWorld %P 14 %V 8 %N 8 %D FEB 24, 1986 %K Network Information Forum Nissho Iwai Corporation machine translation AI02 %X Fujitsu announced an English to Japanese translator which works at 60,000 words per hour. Compuserve and Network Information Forum plan a database exchange which might use this translation software. ------------------------------ End of AIList Digest ******************** From vtcs1::in% Wed Apr 16 00:49:17 1986 Date: Wed, 16 Apr 86 00:49:11 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #90 Status: RO AIList Digest Tuesday, 15 Apr 1986 Volume 4 : Issue 90 Today's Topics: Bibliography - Recent Articles #5 ---------------------------------------------------------------------- Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Recent Articles #5 %A M. Celenk %A S. H. Smith %T A New Systematic Method for Color Image Analysis %R Tech. Rep EE 8509 %D DEC 1985 %I Stevens Institute of Technology Electrical Engineering and Computer Science Departments %K AI06 %A S. A. Friedberg %T Symmetry Evaluators %R TR134 (revised) %D JAN 1986 %I The University of Rochester Computer Science Department %K AI06 Hough transform %X $1.25 24 pages %A D. H. Ballard %A P. J. Hayes %T Parallel Logical Inference and Energy Minimization %R TR142 %D DEC 1985 %I The University of Rochester Computer Science Department %K connectionist H03 AI08 %X $1.50 34 pages %A J. A. Feldman %T Parallelism in High Level Vision %R TR146 %D JAN 1985 %I The University of Rochester Computer Science Department %K H03 AI08 AI06 %X 33 pages $1.50 %A J. Tenenberg %T Reasoning Using Exclusion: an Extension of Clausal Form %R TR147 %D JAN 1986 %I The University of Rochester Computer Science Department %K common sense reasoning AI10 AI11 %X 25 pages $1.25 %A D. H. Ballard %T Form Perception as Transformation %R TR148 %D JAN 1986 %I The University of Rochester Computer Science Department %K AI06 AI07 %X 34 pages $1.50 %A A. Basu %A C. M. Brown %T Algorithms and Hardware for Efficient Image Smoothing %R TR149 %D DEC 1984 %I The University of Rochester Computer Science Department %K AI06 H03 median mean filters %X 20 pages $1.00 %A B. Sarachan %T Experiments in Rotational Egomotion Calculation %R TR152 %D FEB 1985 %I The University of Rochester Computer Science Department %K AI06 %X 26 pages $1.25 [Seems to be a paper for a robot to determine if it got rotated. LEFF] %A G. W. Cottrell %T A Connectionist Approach to Word Sense Disambiguation %R TR154 %D MAY 1985 (PHD Thesis) %I The University of Rochester Computer Science Department %K AI02 AI08 %X 242 pages $7.25 %A J. A. Feldman %T Energy and the Behavior of Connection Models %R TR155 %D NOV 1985 %I The University of Rochester Computer Science Department %K H03 AI12 %X 41 pages, $1.75 %A D. Sher %T Template Matching on Parallel Architectures %R TR156 %D JUL 1985 %I The University of Rochester Computer Science Department %K H03 AI06 Fourier Transform WARP Butterfly %X 28 pages, $1.25 %A A. Bandopadhay %A J. Aloimonos %T Perception of Rigid Motion from Spatio-Temporal Derivatives of Optical Flow %R TR157 %D MAR 1985 %I The University of Rochester Computer Science Department %K AI06 %X 18 pages $1.00 [Seems to be another paper on getting a robot to tell whether somebody rotated it or not LEFF] %A J. Aloimonos %A A. Bandopadhay %T Perception of Structures from Motion: Lower Bound Results %R TR158 %D MAR 1985 %I The University of Rochester Computer Science Department %K AI06 %X 16 pages $1.00 %A J. Aloimonos %T One Eye Suffices: a Computational Model of Monocular Robot Depth Perception %R TR160 %D DEC 1984 %I The University of Rochester Computer Science Department %K AI06 optical flow depth perception orthographic perspective projection %X 16 pages $1.00 %A J. Aloimonos %A P. B. Chou %T Detection of Surface Orientation and Motion from Texture: 1. The Case of Planes %R TR161 %I The University of Rochester Computer Science Department %K AI06 Gibson %X 21 pages $1.25 %A Henry A. Kautz %T Toward a Theory of Plan Recognition %R TR162 %I The University of Rochester Computer Science Department %K AI09 %D JUL 1985 %X 15 pages $1.00 %A L. Shastri %T Evidential Reasoning in Semantic Networks: A Formal Theory and its Parallel Implementation %R TR166 %I The University of Rochester Computer Science Department %K H03 O04 %D SEP 1985 %X 256 pages $7.50 %A D. H. Ballard %A P. Gardner %A M. Srinivas %T Graph Problems and Connection Architectures %I The University of Rochester Computer Science Department %R TR167 %K H03 AI12 %D DEC 1985 %X 24 pages $1.25 %A A. Bandopadhay %T Constraints on the Computation of Rigid Motion Parameters from Retial Displacements %I The University of Rochester Computer Science Department %R TR168 %K AI07 AI06 %D OCT 1985 %X 77 pages, $2.75 [Seems to be another paper on getting a robot to tell whether somebody rotated it or not LEFF] %A A. Bandopadhay %A J. Aloimonos %T Perception of Structure and Motion of Rigid Objects %D DEC 1985 %I The University of Rochester Computer Science Department %R TR169 %K AI07 AI06 %X 55 pages $2.00 [Seems to be another paper on getting a robot to tell whether somebody rotated it or not LEFF] %A D. J. Litman %T Plan Recognition and Discourse Analysis: An Integrated Approach for Understanding Dialogues %D 1985 %R TR170 %I The University of Rochester Computer Science Department %K AI02 AI09 %X 197 pages $6.00 %A J. A. Feldman %A D. H. Ballard %A C. M. Brown %A G. S. Drell %T Rochester Connectionist Papers 1979-85 %D DEC 1985 %R TR172 %I The University of Rochester Computer Science Department %K AI12 AT21 %X no charge %A N. Murray %A E. Rosenthal %T On Deleting Links in Semantic Graphs %R TR 85-4 %I State University of New York at Albany, Computer Science Department %K predicate calculus path resolution AI11 %A S. Chaiken %A N. Murray %A E. Rosenthal %T An Application of $P sub 4$ Free Free Graphs in Theorem Proving %R TR85-8 %I State University of New York at Albany, Computer Science Department %K AI11 %X We describe the application of graphs that have no induced $P sub 4$ (4 vertex path) subgraphs to automatic theorem proving. The semantics of a propositional formula are expressed in terms of the maximal cliques in a $P sub 4$ free graph rather than by truth assignments. Arc sets of s-t paths in a series parallel network provide an equivalent formulation. We provide combinatorial foundations for Murray and Rosenthal's work on path resolution (e. g. TR84-1, TR 84-12 and TR 85-4) For any graph G, a c-block (resp d-block) is an induced subgraph H in G such that for all maximal cliques (resp maximal stable sets) C in G, C $int$ H is $PHI$ or is a maximal clique (resp. maximal stable set) in H. A full block is botha c-block and a d-block. Blocks are generalizations of substitution subgraphs which occur in Lovasz's work on perfect graphs. Theorem: If full block H is $P sub 4$-free then H must arise by substitution. Other properties in these blocks in arbitrary graphs and in $P sub 4$-free graphs are given. These constructs are instrumental in the development of several closely related inference rules collectively referred to as path resolution. Finally we show how semantics of $P sub 4$ graphs are generalized to blocking systems by Minty's painting lemma. This suggests possible generalization of path resolution to other combinatorial structures. %A M. Balaban %T Western Tonal Music - A New Domain for AI Research %R TR 85-10 %I State University of New York at Albany, Computer Science Department %K AI02 AA25 %A M. Balaban %T Knowledge Representation and Inferencing in a Musical Database %R TR 85-11 %I State University of New York at Albany, Computer Science Department %K frames AA25 AA14 T02 %A M. Balaban %T The Generalized Concept Formalism - A Frame and Logic Based Representation Model %R TR 85-20 %I State University of New York at Albany, Computer Science Department %K AA25 T02 %A Mira Balaban %T Foundations for Artificial Intelligence Research of Western Tonal Music %R TR 85-22 %I State University of New York at Albany, Computer Science Department %K AA25 %A M. Balaban %T CSM: An AI Approach to the Study of Western Tonal Music %R TR 85-24 %I State University of New York at Albany, Computer Science Department %K AA25 %A H. B. Hunt %A R. E. Stearns %T Distributive Lattices and the Complexity of Logics and Probability %R TR 85-28 %I State University of New York at Albany, Computer Science Department %K AI11 O04 %X Relationships between number of repetitions of variables in formulas and complexity of decision problems for the formulas. Applications to logic and probability: 1) Any reasonable propositional calculus with a reasonable implication operator has a coNP-hard logical Validy problem. This is true for very simple formulas involving or, and and a single occurrence of the implication operator 2) The set of theorems of the propositional calculus of classical implicative logic is coNP complete 3. Computing the probabilities of a joint event and a conditional event becomes "hard" almost immediately when the events E1 and E2 are not statistically independent %A H. B. Hunt %A R. E. Stearns %T Monotone Boolean Formulas, Distributive Lattices, and the Complexities of Logics, Algebraic Structures, and Computation Structures (Preliminary Report) %R TR85-29 %I State University of New York at Albany, Computer Science Department %K AI11 O04 %A Andrew Laine %A Seymour V. Pollack %T The Enhanced Wudma Image Processing %R WUCS-85-1 %I Department of Computer Science, Washington University %C St. Louis, Missouri %K AI06 %A S. E. Elnahas %A R. G. Jost %A J. R. Cox %A R. L. Hill %T Transmission Progressive of Digital Diagnostic Images %R WUCS-85-8 %I Department of Computer Science, Washington University %C St. Louis, Missouri %K AI06 AA01 %X Progressive transmission of digital pictures permits the receiver to construct an approximate picture first, then gradually improve the quality of reconstruction. %A James R. Slagle %A JOhn M. Long %A Michael R. Wick %A John P. Matts %A Arthur S. Leon %T Expert Systems in Medical Studies- A New Twist %R TR 86-3 %I University of Minessota, Department of Computer Science %D 1986 %K AA01 AI01 %A Robert M. Herndon, Jr. %A Valdis A. Berzins %T An Interpretive Technique for Evaluating Functional Attribute Grammars %R TR 86-5 %I University of Minessota, Department of Computer Science %R 1986 %A Robert M. Herndon, Jr. %A Valdis A. Berzins %T A Method for the Construction of Dynamic, Lazy Evaluators for Functional Attribute Grammars %R 86-6 %I University of Minessota, Department of Computer Science %R 1986 %A J. Schwartz %A M. Sharir %T Efficient Motion Planning Algorithms in Environments of Bounded Local Complexity %R 164 %I New York University, Courant Institute of Mathematical Sciences, Department of Computer Science %K AI07 %D JUN 1985 %A J. Schwartz %A M. Sharir %T Identification of Partially Obscured Objects in Two Dimensions by Matching of Noisy 'Characteristic Curves' %R 165 %I New York University, Courant Institute of Mathematical Sciences, Department of Computer Science %D JUN 1985 %K AI06 %A G. Landau %A U. Vishkin %T Efficient String Matching with k Mismatches %R 167 %I New York University, Courant Institute of Mathematical Sciences, Department of Computer Science %D JUN 1985 %X Give a text of length n, a pattern of length m and an integer k, we present an algorithm for finding all occurrences of the patterns in the text, with at most k mismatches running in O(k(mlogm + n) %A G. Landau %A U. Vishkin %T An Efficient String Matching Algorithm with k Differences for Nucleotide and Amino Acid Sequences %R 168 %I New York University, Courant Institute of Mathematical Sciences, Department of Computer Science %D JUN 1985 %X Algorithm to allow for optimal alignment of one sequence, the pattern of length m, with another longer sequence the text, of length n. These algorithms allow mismatches, deletions and insertions. If k is the maximum number of differences, then the time is O(k sup 2 n). %A R. Hummel %A A. Rojer %T Connected Component Labeling in Image Processing with MIMD Architectures %R 173 %I New York University, Courant Institute of Mathematical Sciences, Department of Computer Science %D SEP 1985 %K AI06 H03 %A S. Zucker %A R. Hummel %T Receptive Fields and the Representation of Visual Information %R 176 %I New York University, Courant Institute of Mathematical Sciences, Department of Computer Science %D SEP 1985 %K AI06 AI08 Gaussian retina %X Hypothesis that the receptive fields of the retina provide a suitable method for transmitting the image over the optic nerve which is a limited bandwidth channel. %A M. Landy %A R. Hummel %T A Brief Survey of Knowledge Aggregation Methods %R 177 %I New York University, Courant Institute of Mathematical Sciences, Department of Computer Science %D SEP 1985 %K AI04 %A G. Landau %A U. Vishkin %T Efficient String Matching with k Differences %R 186 %I New York University, Courant Institute of Mathematical Sciences, Department of Computer Science %D OCT 1985 %X If the mismatches considered are a single character mismatch, a superfluous character in the text or pattern, there exists an algorithm that runs in time O(m+k sup 2 n ) when the alphabet size is fixed and O(m log m + k sup 2 n) otherwise where m is length of pattern, k is the number of mismatches and n is the text. %A D. Leven %A M. Sharir %T On the Number of Critical Free Contacts of a Convex Polygonal Object Moving in 2-D Polygonal Space %R 187 %I New York University, Courant Institute of Mathematical Sciences, Department of Computer Science %D OCT 1985 %K AI07 %A J. Burdea %A H. Wolfson %T Automated Assembly of a Jigsaw Puzzle Using the IBM 7565 Robot %R 188 %I New York University, Courant Institute of Mathematical Sciences, Department of Computer Science %D NOV 1985 %K AI07 %A E. Davis %T Constraint Propagation on Real-Valued Quantities %R 189 %I New York University, Courant Institute of Mathematical Sciences, Department of Computer Science %D NOV 1985 %K AI03 %A N. S. Sridharan %T Representing Knowledge in Introduction using TAXMAN Examples %R LRP-TR-12 %D 11/81 %I Rutgers University, Department of Computer Science %R LRP-TR-13 %D 1/82 %T "A Computational Theory of Legal Argument" %A L. T. McCarty %A N. S. Sridharan %I Rutgers University Department of Computer Science %K AA24 tax %X The TAXMAN project is an experiment in the application of artificial intelligence to the study of legal reasoning and legal argumentation, using corporate tax law as an experimental problem domain. Legal concepts possess what is often termed "open-texture", that is, their definitions are subject to a continual process of construction and modification during the analysis of a contested case. We have developed a "prototype-plus-deformation" representation for the structure of such concepts, a representation which facilitates the formulation of several systematic methods of conceptual modification. We propose now to construct a cognitive model of the process of legal argument, using this representation. The research is aimed at developing explanations for the persuasiveness of certain strategies of legal argument, and at developing further the criteria of conceptual coherence, both task-specific and task-independent, which seem to constrain the space of plausible arguments. We emphasize not only the contributions of this research to Artificial Intelligence, but also the insights that may result for some of the fundamental issues in jurisprudence. %R LRP-TR-14 %D 9/82 %T "A Flexible Structure for Knowledge" %A N.S. Sridharan %K AA24 tax AI04 %I Rutgers University Department of Computer Science %X Concepts often dealt with in legal reasoning and argumentation are amorphous. For TAXMAN II, we have proposed in the past a Prototype and Deformation model for these amorphous concepts. In this model, a concept is represented as a structured space of exemplars, that is as a set of exemplars, structured by transformations and relationships among them. In this paper, the idea of representing a concept as a structured space of exemplars is extended; suggesting that all knowledge represented in a computer be organized as structured spaces and subspaces. Concepts are represented as spaces; concepts are also members of spaces. This duality is exploited to gain flexibility in the representation, that is, changes to the structure can be effected through computation. %R LRP-TR-15 %D 6/83 %T "Concept Learning by Building and Applying Transformations Between Object Descriptions" %A Donna Nagel %K AI04 analogy matching %I Rutgers University Department of Computer Science %X The Concept Learning presented here emphasizes the building of a transformation between an instance of a concept and another instance which is distinguished as a prototype of the concept. A recursive partial matcher is used to pinpoint components of structural object descriptions of the training instances for matching. Three procedures are described for inducing matches: building simple analogies, applying primitive transformations, and finding projections of the instances into domains of knowledge relevant to the concept being learned. This research is experimental in nature and directed at discovering flexible ways to define and represent concepts which are amorphous and open-textured. %R LRP-TR-16 %D 3/84 %T "EVOLVING SYSTEMS OF KNOWLEDGE" %A N.S. Sridharan %I Rutgers University Department of Computer Science %K AI01 %X The enterprise of developing knowledge-based systems, is currently witnessing great growth in popularity. The central unity of such programs is that they interpret knowledge that is explicitly encoded as @i[rules]. This paper is a statement of personal perspective by a researcher interested in fundamental issues in the symbolic representation and organization of knowledge. The discussion covers the nature of rules (Sec. 3), and methods of rule-handling (Sec. 4). The paper concludes with a discussion of how most concepts we use are open-textured and how they continually evolve with use (Sections 5,6,7). While rule-based programming comes with certain clear pay-offs, further fundamental advances in research is needed to extend the scope of tasks that can be adequately represented in this fashion. %R LRP-TR-17 %D 6/84 %T "Analogy with Purpose in Legal Reasoning from Precedents" %A S.Kedar-Cabelli %D 10/84 %I Rutgers University Department of Computer Science %K AA24 taxman tax AA04 %X One open problem in current artificial intelligence (AI) models of learning and reasoning by analogy is: which aspects of the analogous situations are relevant to the analogy, and which are irrelevant? It is currently recognized that analogy involves mapping some underlying causal network of relations between situations [Winston 82], [Gentner 83], [Burstein 83a], [Carbonell 83]. However, most current models of analogy provide the system with exactly the relevant relations, tailor-made to each analogy to be performed. As AI systems become more complex, we will have to provide them with the capability of automatically focusing on the relevant aspects of situations when reasoning analogically. These will have to be sifted from the large amount of information used to represent complex, real-world situations. .sp 1 In order to study these general issues, we are examining a particular case study of learning and reasoning by analogy: forming legal concepts by legal reasoning from precedents. This is studied within the TAXMAN II project, which is investigating legal reasoning using AI techniques [McCarty & Sridharan 82], [Nagel 83]. .sp 1 In this dissertation proposal, we will discuss the problem and a proposed solution. We examine legal reasoning from precedents within the context of current AI models of analogy. We then add a focusing capability. Current work on goal-directed learning [Mitchell 83a], [Mitchell & Keller 83], and explanation-based learning [Dejong 83] applies here: the explanation of how the precedent satisfies the intent of the law (i.e. its goals, or purposes) helps to automatically focus the reasoning on what is relevant. ------------------------------ End of AIList Digest ******************** From vtcs1::in% Tue Apr 15 06:51:27 1986 Date: Tue, 15 Apr 86 06:51:21 est From: vtcs1::in% (LAWS@sri-ai.ARPA) To: ailist@sri-ai.arpa Subject: AIList Digest V4 #91 Status: R AIList Digest Tuesday, 15 Apr 1986 Volume 4 : Issue 91 Today's Topics: Bibliography - Recent Articles #6 ---------------------------------------------------------------------- Date: WED, 10 JAN 84 17:02:23 CDT From: E1AR0002%SMUVM1.BITNET@WISCVM.WISC.EDU Subject: Recent Articles #6 %R CTA-TR-2 %D 8/80 %T Computability on Binary Trees - An Extended Abstract %A A. Yasuhara %A F. Hawrusik %A K.N. Venkataraman %D 1/82 %I Rutgers University %X We propose an effective method of computation on finite binary trees that is analogous to the effective computation on the natural numbers determined by the partial recursive functions. Not surprisingly, the method is LISP-like. A finitely axiomatizable theory is given that is shown to be just strong enough to represent the class of functions computable by this method. Several natural subclasses; of this class of functions are delineated and they are shown to be different from one another. %R CTA-TR-3 %D 3/81 %T Sub Classes of Programs for Computing on Binary Trees %A K.N. Venkataraman %D 1/82 %I Rutgers University %K T01 %X Several sub-classes of the deterministic regular programs that compute on binary trees are defined and relations of inclusion and inequality among these classes in terms of functions computable (by these programs) are established. Certain properties of these classes of programs are studied. In particular the sets recognized by these programs are characterized in terms of the domain and range of these programs. Most of the results that appear in this paper can easily be extended to programs computing on other recursively defined data structures. %R CTA-TR-4 %D 10/81 %T Decidability of the Purely Existential Fragment of the Theory of Term Algebras %A K.N. Venkararaman %X This thesis is concerned with the question of the decidability and the complexity of the decision problem for certain fragments of the theory of free term algebras. .sp 1 The existential fragment of the theory of term algebras is shown to be decidable by presenting a non-deterministic algorithm which given a quantifier free formula P, constructs a solution for P if it has one and indicates failure if there are no solutions. A detailed proof of the correctness of the algorithm is given. It is shown that the decision problem is in NP by proving that if a quantifier-free formula P has a solution then there is one that can be represented as a dag in space at most cubic in the length of P. The decision problem is shown to be complete for NP by reducing 3-SAT to that problem. It is also shown that the @ @ @-[o] hierarchy over a term algebra corresponds to the polynomial time hierarchy. .sp 1 The proof of the fact that the introduction of the selector functions into the first order language does not increase the complexity of the existential fragment of the theory is indicated. Thus it is established that the existential fragment of the theory of list structures in the language of NIL, CONS, CAR, CDR, = , @u[<] is NP-complete. .sp 1 It is shown that the equivalence of PB[;@u{<}] straight line programs is decidable follows easily from the decidability of the existential fragment of the theory of list structures. .sp 1 It is also shown that for any quantifier free formula P (in the language of a term algebra) there is an algorithm which given a recursive set S of cardinal numbers @u{<} @ @ @-[o], can decide whether or not the number of solutions of P is in S. %R ML-TR-1 %D 7/85 %T Purpose-Directed Analogy %A Smadar Kedar-Cabelli %I Rutgers University %X Recent artificial intelligence models of analogical reasoning are based on mapping some underlying causal network of relations between analogous situations. However, causal relations relevant for the purpose of one analogy may be irrelevant for another. We describe here a technique which uses an explicit representation of the purpose of the analogy to automatically create the relevant causal network. We illustrate the technique with two case studies in which concepts of everyday artifacts are learned by analogy. %R ML-TR-2 %D 8/85 %T Explanation-Based Generalization: A Unifying View %A T.M. Mitchell %A R.M. Keller %A S.T. Kedar-Cabelli %X The problem of formulating general concepts from specific training examples has long been a major focus of machine learning research. While most previous research has focused on empirical methods for generalizing from a large number of training examples using no domain-specific knowledge, in the past few years new methods have been developed for applying domain-specific knowledge to formulate valid generalizations from single training examples. The characteristic common to these methods is that their ability to generalize from a single example follows from their ability to explain why the training example is a member of the concept being learned. This paper proposed a general, domain-independent mechanism, called EBG, that unifies previous approaches to explanation-based generalization. The EBG method is illustrated in the context of several example problems, and used to contrast several existing systems for explanation-based generalization. The perspective on explanation-based generalization afforded by this general method is also used to identify open research problems in this area. %R RC-5882 %D February 1976 %A John Thomas %T A Method of Studying Natural Language Dialogue %I IBM Watson Research Center, User Interface Institute %K AI02 %R RC-10823 %D November 1984 %A John Thomas: %T Artificial Intelligence and Human Factors. %I IBM Watson Research Center, User Interface Institute %K AI08 %A Carbonell, Jaime %T Derivational analogy: a theory of reconstructive problem solving and expertise acquisition. %I Carnegie-Mellon University. Department of Computer Science. %R CMU-CS-85-115. %D 1985. %K Case-based reasoning. %A Kahn, Gary %A McDermott, John %T MUD: a drilling fluids consultant %I Carnegie-Mellon University. Department of Computer Science %R CMU-CS-85-116 %D 1985 %K Diagnostic systems, Knowledge acquisition AI01 AA03 AA21 %A Doyle, Jon %T Reasoned assumptions and Pareto optimality %I Carnegie-Mellon University. Department of Computer Science %R CMU-CS-85-121 %D 1985. %K Economic theory Group decision making Inference rules Non-monotonic reasoning AA11 %A David M. McKeown, Jr %A Pane, John F %T Alignment and connection of fragmented linear features in aerial imagery %I Carnegie-Mellon University. Department of Computer Science %R CMU-CS-85-122 %D 1985 %K Cultural features Feature extraction Image segmentation Region interpolation Spline approximation AI06 %A Dill, David %A Clarke, Edmund %T Automatic verification of asynchronous circuits using temporal logic %I Carnegie-Mellon University. Department of Computer Science %R CMU-CS-85-125 %D 1985 %K Circuit design Timing constraints AA04 AI11 %A Lehr, Theodore %T The implementation of a production system machine %I Carnegie-Mellon University. Department of Computer Science %R CMU-CS-85-126 %D 1985 %K Computer architecture OPS5 Performance improvement Production systems RISCF Rete algorithm AI01 %A Minton, Steven %T A game-playing program that learns by analyzing examples %I Carnegie-Mellon University. Department of Computer Science %R CMU-CS-85-130 %D 1985 %K Concept acquisition Constraint based generalization Forcing configurations Learning from examples Machine learning Tactical combinations Winning combinations AI04 AA17 %A Fox, Mark %A Wright, J. Mark %A Adam, David %T Experiences with SRL: an analysis of a frame-based knowledge representation %I Carnegie-Mellon University. Robotics Institute %R CMU-RI-TR-85-10 %D 1985 %K Knowledge representation languages %A Smith, Stephen %A Ow, Peng Si %T The use of multiple problem decompositions in time constrained planning tasks %I Carnegie-Mellon University. Robotics Institute %R CMU-RI-TR-85-11 %D 1985 %K Job shop scheduling %K Multi-agent planning systems %K Resource allocation AI10 %A Brost, Randy %T Planning robot grasping motions in the presence of uncertainty %I Carnegie-Mellon University. Robotics Institute %R CMU-RI-TR-85-12 %D 1985 %K Manipulators AI07 O04 AI09 %A Darlington, %A Field, A. %A Pull, H. %T The unification of functional and logic languages %I Imperial College of Science and Technology. Department of Computing %R Research report DOC 85/3 %D 1985 %K Functional programming Reduction Resolution AI10 %A Gregory, Steve %A Neely, Rob %A Ringwood, Graem %T Parlog for specification, verification and simulation %I Imperial College of Science and Technology. Department of Computing %R Research report DOC 85/7 %D 1985 %K PARLOG AI10 H03 %A Saint-Dizier, Patrick %T On syntax and semantics of adjective phrases in logic programming %I Institut National de Recherche en Informatique et en Automatique (INRIA) %R Rapport de recherche 381 %D 1985 %K AI10 %A Deransart, Pierre %A Maluszynski, Jan %T Relating logic programs and attribute grammars %I Institut National de Recherche en Informatique et en Automatique (INRIA) %R Rapport de recherche 393 %D 1985 %K Attribute dependency scheme Data flow analysis Logic programming AI10 %A Gazdar, Gerald %A Pullum, Geoffrey K %T Computationally relevant properties of natural languages and their grammars %I Stanford University. Center for the Study of Language and Information %R CSLI-85-24 %D 1985 %P 45 %K AI02 %A Fagin, Ronald %A Vardi, Moshe %T An internal semantics for modal logic: preliminary report %I Stanford University. Center for the Study of Language and Information %R CSLI-85-25 %D 1985 %P 24p %K AI10 %A Barwise, Jon %T The situation in logic - III: simulation, sets and the axiom of foundation %I Stanford University. Center for the Study of Language and Information %R CSLI-85-26 %D 1985 %A van Benthem, Johan %T Semantic automata %I Stanford University. Center for the Study of Language and Information %R CSLI-85-27 %D 1985 %A Sells, Peter %T Restrictive and non-restrictive modification %I Stanford University. Center for the Study of Language and Information %R CSLI-85-28 %D 1985 %A Abadi, Martin %A Manna, Zohar %T Nonclausal temporal deduction %I Stanford University. Department of Computer Science %R STAN-CS-85-1056 %D 1985 %P 17p %K Nonclausal resolution Propositional temporal logic AI10 AI11 %A Mason, Ian A %A Talcott, Carolyn L %T Memories of S-expressions: proving properties of Lisp-like programs that destructively alter memory %I Stanford University. Department of Computer Science %R STAN-CS-85-1057 %D 1985 %K Computations over memory structures Correctness proofs Robson copying algorithm AI11 AA08 %A Taubenfeld, G %A Francez, N %T Proof rules for communication abstractions %I TECHNION - Israel Institute of Technology. Department of Computer Science %R Technical report 332 %D 1984 %K Concurrent programming Deadlock Invariants Program verification %K Scripts AA08 %A Shmueli, O %A Tsur, S %A Zfira, H %T Rule supporting in PROLOG %I TECHNION - Israel Institute of Technology. Department of Computer Science %R Technical report 337 %D 1984 %K T02 %A Shapiro, Ehud %T A subset of Concurrent Prolog and its interpreter %I Weizmann Institute of Science. Department of Applied Mathematics %R CS83-06 %D 1983 %K T02 H03 %X "This is a revised version of technical report TR-003, ICOT-Institute for New Generation Computing Technology."; %A Shapiro, Ehud %A Takeuchi, Akikazu %T Object oriented programming in Concurrent Prolog %I Weizmann Institute of Science. Department of Applied Mathematics %R CS83-08 %D 1983 %K H03 T02 %A Harel, David %A Peleg, David %T Process logic with regular formulas %I Weizmann Institute of Science. Department of Applied Mathematics %R CS83-11 %D 1983 %A Hellerstein, L %A Shapiro, Ehud Y %T Implementing parallel algorithms in Concurrent Prolog %I Weizmann Institute of Science. Department of Applied Mathematics %R CS83-12 %D 1983 %K T02 H03 %X Summary/draft, August 1983 %A Manna, Zohar %A Pnueli, Amir %T How to cook a temporal proof system for your pet language %I Weizmann Institute of Science. Department of Applied Mathematics %R CS83-13 %D 1983 %K AA08 AI11 %A Harel, David %A Peleg, David %T On static logics, dynamic logics and complexity classes %I Weizmann Institute of Science. Department of Applied Mathematics %R CS83-15 %D 1983 %K AI11 %A Feldman, Yishai A %T A decidable propositional probabilistic dynamic logic %I Weizmann Institute of Science. Department of Applied Mathematics %R CS83-18 %D 1983 %K AI11 %A Barringer, Howard %A Kuiper, Ruurd %A Pnueli, Amir %T Now you may compose temporal logic specifications %I Weizmann Institute of Science. Department of Applied Mathematics %R CS84-09 %D 1984 %K AI11 %A Shapiro, Ehud Y %T The Bagel: a systolic Concurrent Prolog machine (lecture notes) %I Weizmann Institute of Science. Department of Applied Mathematics %R CS84-10 %D 1984 %K H03 T02 %A Peleg, David %T Concurrent dynamic logic %I Weizmann Institute of Science. Department of Applied Mathematics %R CS84-14 %D 1984 %K T02 H03 %A Mierowsky, Colin %T Design and implementation of flat Concurrent Prolog %I Weizmann Institute of Science. Department of Applied Mathematics %R CS84-21 %D 1984 %K H03 T02 %X Thesis (M.S.) %A Bloch, Charlene %T Source-to-source transformations of logic programs %I Weizmann Institute of Science. Department of Applied Mathematics %R CS84-22 %D 1984 %K AI10 %X Thesis (M.S.) %A Viner, Omri %T Distributed constraint propagation %I Weizmann Institute of Science. Department of Applied Mathematics %R CS84-24 %D 1984 %K H03 %X Thesis %A Peleg, David %T Concurrent program schemes and their logics %I Weizmann Institute of Science. Department of Applied Mathematics %R CS84-25 %D 1984 %K H03 T02 %A Lichtenstein, Orna %A Pnueli, Amir %T Checking that finite state concurrent programs satisfy their linear specification %I Weizmann Institute of Science. Department of Applied Mathematics %R CS84-26 %D 1984 %K AA08 %A Nygate, Yossi %T Python: a bridge expert on squeezes %I Weizmann Institute of Science. Department of Applied Mathematics %R CS84-27 %D 1984 %A Nixon, I. M. %T I.F.: an Idiomatic Floorplanner %I University of Edinburgh. Department of Computer Science %R CSR-170-84 %D 1984 %K VLSI AA04 %A Sannella, Donald %A Tarlecki, Andrzej %T On observational equivalence and algebraic specification %I University of Edinburgh. Department of Computer Science %R CSR-172-84 %D 1984 %A Prasad, K. V. S %T Specification and proof of a simple fault tolerant system in CCS %I University of Edinburgh. Department of Computer Science $R CSR-178-84 %D 1984 %K AA08 AI11 %A Blake, Andrew %T Inferring surface shape by specular stereo %I University of Edinburgh. Department of Computer Science %R CSR-179-84 %D 1984 %K AI06 %A Dolan, Charles %T Memory based processing for cross contextual reasoning: reminding and analogy using thematic structures %I University of California, Los Angeles. Computer Science Department %R CSD-850010 %D 1985 %X Thesis (M.S.) %A Hooper, Richard %T An application of knowledge-based systems to electronic computer-aided engineering, design, and manufacturing data base transport %I University of California, Los Angeles. Computer Science Department %R CSD-850011 %D 1985 %K AA05 AA04 %X Thesis (Ph.D.) %A Rendell, Larry %T Induction, of and by probability %I University of Illinois, Urbana-Champaign. Department of Computer Science %R UIUCDCS-R-85-1209 %D 1985 %K Conceptual clustering Inductive inference AI04 Noise management Probabilistic learning systems %A Rendell, Larry %T Genetic plans and the probabilistic learning system: synthesis and results %I University of Illinois, Urbana-Champaign. Department of Computer Science %R UIUCDCS-R-85-1217 %D 1985 %K Conceptual clustering AI12 AI04 %A Anderson, James W. %T Portable Standard LISP on the Cray %I Los Alamos National Laboratory %R LA-UR-84-4049 %D 1984 %K T01 H04 PSL %A Arnon, Dennis S. %T Supercomputers and symbolic computation %I Purdue University. Department of Computer Sciences %R CSD-TR-481 %D 1984 %K AI14 H04 %A J. Schwartz %T A Survey of Program Proof Technology %I New York University, Courant Institute, Department of Computer Sciences %D SEP 1978 %R 001 %K AA08 AI11 %A S. Stolfo %A M. Harrison %T Automatic Discovery of Heuristics for Non-Deterministic Programs %D JAN 1979 %I New York University, Courant Institute, Department of Computer Sciences %R 007 %K AI04 AI03 %A M. Sharir %T Algorithm Derivation by Transformations %D OCT 1979 %I New York University, Courant Institute, Department of Computer Sciences %R 021 %K AA08 %A A. Walker %T Syllog: A Knowledge Based Data Management Systems %D JUN 1981 %I New York University, Courant Institute, Department of Computer Sciences %R 034 Sciences %K AA09 %A J. Schwartz %A M. Sharir %T On the Piano-Movers Problem, I. Case of A Two Dimensional Rigid Polygonal Body Moving Amidst Polygonal Barriers %D OCT 1981 %I New York University, Courant Institute, Department of Computer Sciences %R 039 R1 Sciences %K AI07 %A J. T. Schwartz %A M. Sharir %T On the Piano Movers Problem, II General Techniques for Computing Topologic Properties of Real Algebraic Manifolds %D FEB 1982 %R 041 R2 %I New York University, Courant Institute, Department of Computer Sciences %K AI07 %A J. Schwartz %A M. Sharir %T On the Piano Movers Problem III Coordinating the Motion of Several Independent Bodies: The Special Bodies Moving Amidst Polygonal Bariers %D SEP 1982 %R 052 r3 %I New York University, Courant Institute, Department of Computer Sciences %K AI07 %A C. O'Dunlaing %A C. Yap %T The Voronoi Diagram Method of Motion-Planning: I. The Case of a Disc %D MAR 1982 %R 053 R4 %I New York University, Courant Institute, Department of Computer Sciences %K AI07 %A M. Sharir %A E. Azriel-Sheffi %T On the Piano Movers Problem IV Various Decomposable Two-Dimensional Motion Plannings Problems %D FEB 1983 %I New York University, Courant Institute, Department of Computer Sciences %R 058 R6 %K AI07 %A J. Schwartz %T Structured Light Sensors for 3-D Robot Vision %D MAR 1983 %R 065 R8 %I New York University, Courant Institute, Department of Computer Sciences %K AI06 AI07 %A C. Yap %T Complexity of Motion Coordination %R R12 %I New York University, Courant Institute, Department of Computer Sciences %K AI07 %A J. Schwartz %A M. Sharir %T On the Piano Movers Problem: V. The Case of a Rod Moving in Three Dimensional Space Amidst Polyhedral Obstacles %R 083 R13 %I New York University, Courant Institute, Department of Computer Sciences %D JUL 1983 %K AI07 %A R. Cole %A C. Yap %T Shape from Probing %R 104 R15 %I New York University, Courant Institute, Department of Computer Sciences %D OCT 1983 %K AI07 AI06 %A J. Schwartz %A M. Sharir %T Some Remarks on Robot Vision %R 119 R25 %I New York University, Courant Institute, Department of Computer Sciences %D APR 1984 %K AI07 AI006 %A C. Bastuscheck %A J. Schwartz %T Preliminary Implementation of a Ratio Depth Sensor %R 124 R28 %I New York University, Courant Institute, Department of Computer Sciences %D JUN 1984 ------------------------------ End of AIList Digest ********************