Date: Mon 13 Jun 1988 15:37-EDT From: AIList Moderator Nick Papadakis Reply-To: AIList@AI.AI.MIT.EDU Us-Mail: MIT Mail Stop 38-390, Cambridge MA 02139 Phone: (617) 253-2737 Subject: AIList Digest V7 #30 To: AIList@AI.AI.MIT.EDU Status: RO AIList Digest Tuesday, 14 Jun 1988 Volume 7 : Issue 30 Today's Topics: AI Languages Further Observations on the Fredkin Masters Open Construal of Induction The Definition of Information ---------------------------------------------------------------------- Date: Fri, 10 Jun 88 15:03:21 +0100 From: mcvax!swivax!vierhout@uunet.UU.NET (Paul Vierhout) Subject: AI Languages AIlanguage features: old: procedure-data equivalence less old: nondeterminism, 'streams' ,unification,OPS5 pattern matching, shell-like: ability to specify frames and/or rules, and possibly control promises: abstract models of cognitive tasks like the Interpretation Models of Breuker and Wielinga (SWI-UvA, Amsterdam) for knowledge acquisition, or the six generic tasks of Chandrasekaran (Ohio State Univ.). Not at all an exhaustive list; shouldn't an AIlanguage ideally exhaustively offer all features currently available ? ------------------------------ Date: 11 Jun 88 07:33:43 GMT From: mtune!ihnp4!utah-cs!gr.utah.edu!uplherc!sp7040!obie!wsccs!dharve y@att.att.com Subject: AI Languages In article <19880527050431.8.NICK@MACH.AI.MIT.EDU>, Pat Hayes writes: > Date: Mon, 16 May 88 13:29 EDT > From: hayes.pa@Xerox.COM > Subject: Re: AIList Digest V7 #1 > In-reply-to: AIList Moderator Nick Papadakis 's > message of Sat, 14 May 88 21:50 EDT > To: AIList@AI.AI.MIT.EDU > cc: hayes.pa@Xerox.COM > > I was fascinated by the correspondence between Gabe Nault and Mott Given in > vol7#1, concerning "an artifical intelligence language ... something more than > lisp or xlisp." Can anyone suggest a list of features which a programming > language must have which would qualify it as an "artificial intelligence > language" ? > > Pat Hayes You wanted an opinion, so here it is! There probably isn't such a thing as list of 'must' features for an AI language! There are things I would 'like' it to have though. The first is that it must allow the action to proceed anywhere (ie any Lisp function (procedure), and Prolog predicate, or any FORTH word). The second is that it must be at an abstract level that allows me to forget about what the computer is doing and concentrate on what I am doing. Obviously, there is no such language in existence that does this yet. So we latch onto anything that even closely resembles this ideal. Take your pick, Smalltalk, Lisp, Prolog. One person's dream is bound to be someone else's nightmare! dharvey @ wsccs (David A Harvey) I am responsible for Nobody, and Nobody is responsible for me. Behind the Mormon Curtain. ------------------------------ Date: 11 Jun 88 16:06:46 GMT From: hyper-sun1.jpl.nasa.gov!cracraft@jpl-elroy.arpa (Stuart Cracraft) Subject: Further Observations on Fredkin Masters Open Now that there has been a "settling in" period, when people have been able to digest the recent news, it seems advisable to put it in perspective. A brief retrospective: Two new chess machines played in an Eastern U.S. chess tournament against 18 masters and a few experts and class players. While the performance of the machines, especially one of them which came in 2nd in the tournament with a performance rating of over USCF 2500, is laudable, an inspection of the games reveals that many of the machine's human opponents sacrificed pawns and exchanges needlessly. This style of play against an unknown opponent (the machine) by players would seem to indicate a level of contempt that is generally self-defeating. Discomforting is the fact that several of these players had faced very powerful computer programs earlier in their careers, almost always scoring a plus. The players who indicate contemptuousness end up taking unordinary risks and generally underestimating most of their opponent's moves; this lowers their quality of play and greatly degrades their performance. So, while I think the performance of the 2nd-runner and its predecessor is quite good, I also feel that players will be "on-guard" even more so in the future and that this incident does not mean a lessening of human chess; rather, it is a call to arms so that we may all regard our opponent with more respect. Stuart ------------------------------ Date: 12 Jun 88 20:24:11 GMT From: tness7!tness1!flatline!erict@bellcore.bellcore.com (j eric townsend) Subject: Re: Further Observations on Fredkin Masters Open In article <6998@elroy.Jpl.Nasa.Gov>, Stuart Cracraft writes: > This style of play against an unknown opponent (the machine) by players > would seem to indicate a level of contempt that is generally self-defeating. Now maybe this is already done, and my ignorance will get me a swift boot to the ego, but... Why not play these games double-blind? Not being an avid chess player, there may be lots of reasons that this wouldn't work that don't occur to me at this moment... Problems: 1. Players able to tell by "style" that they're playing a computer(?) I know this is true for most any computerized wargame/combat simulation. 2. Undue discomfort for the human players by not being able to see *any* opponent ever. Always playing a "black-box" human flunkie/motorized chess piece mover. (See benefit #1) Benefits: 1. Above discomfort would be spread equally against all opponents this was the point in the first place: create an equal level of discomfort for those playing humans. (Discomfort could be nonexistant for each therefore equal.) 2. Um.. Um.. Oh well. It was just a thought... -- Know Future Skate UNIX or go home, boogie boy... J. Eric Townsend ->uunet!nuchat!flatline!erict smail:511Parker#2,Hstn,Tx,77007 ..!bellcore!tness1!/ ------------------------------ Date: 10 Jun 88 19:53:38 GMT From: Venugopala R. Dasigi Subject: Construal of Induction Path: mimsy!venu From: venu@mimsy.UUCP (Venugopala R. Dasigi) Newsgroups: comp.ai.digest Subject: Re: construal of induction Message-ID: <11908@mimsy.UUCP> Date: 10 Jun 88 19:53:36 GMT References: <19880609224213.9.NICK@INTERLAKEN.LCS.MIT.EDU> Reply-To: venu@mimsy.umd.edu.UUCP (Venugopala R. Dasigi) Organization: U of Maryland, Dept. of Computer Science, Coll. Pk., MD 20742 Lines: 43 In an earlier article Raul Valdes-Perez writes: >The concept of induction has various construals, it seems. The one I am >comfortable with is that induction refers to any form of ampliative >reasoning, i.e. reasoning that draws conclusions which could be false >despite the premises being true. This construal is advanced by Wesley >Salmon in the little book Foundations of Scientific Inference. Accordingly, >any inference is, by definition, inductive xor deductive. ^^^^^^^^^^^^^^^^^^^^^^^ > >I realize that this distinction is not universal. For example, some would >distinguish categories of induction. I would appreciate reading comments >on this topic in AILIST. I think it was Charles Sanders Peirce who made the distinction between three types of resoning: induction, deduction and abduction. (Also, Harry Pople's famous paper on "The Mechanization of Abductive Logic," Proc. IJCAI, 1973, pp 147-152 mentions this. Consider the following three possible components of reasoning: 1. A --> B 2. A 3. B (e.g., 1. All beans in this bag are white. 2. This bean is from this bag. 3. This bean is white.) Deduction involves inferring 3 from 1 and 2. Induction involves inferring 1 from 2 and 3. Abduction invloves inferring 2 from 1 and 3. (This was the way Peirce characterized the three types of logic.) Now, my point is abduction also involves drawing conclusions which could be false despite the premises being true, but that is not commonly construed as a type of induction. Accordingly, I am not comfortable with the statement that any inference is inductive XOR deductive (exclusive, all right, but not necessarily exhaustive). I admit I have to read Salmon's book, though. --- Venu Dasigi -- Venugopala Rao Dasigi ARPA: venu@mimsy.umd.edu CSNet: venu@umcp-cs/venu@mimsy.umd.edu UUCP: {allegra,brl-bmd}!mimsy!venu@uunet.uu.net US Mail: Dept. of CS, Univ. of Maryland, College Park, MD 20742-3255 ------------------------------ Date: 10 Jun 88 18:05:04 GMT From: USENET NEWS Subject: definition of information Path: linus!mbunix!bwk From: bwk@mitre-bedford.ARPA (Barry W. Kort) Newsgroups: comp.ai.digest Subject: Re: definition of information Summary: And now for something completely different. Keywords: Yes, but what difference does it make? Message-ID: <34059@linus.UUCP> Date: 10 Jun 88 18:05:03 GMT References: <19880609224803.1.NICK@INTERLAKEN.LCS.MIT.EDU> Sender: news@linus.UUCP Reply-To: bwk@mbunix (Barry Kort) Distribution: world Organization: IdeaSync, Inc., Chronos, VT Lines: 21 Bruce Nevin asks: >Can anyone point me to a coherent definition of information respecting >information content, as opposed to merely "quantity of information"? I quote the following from Stewart Brand's book, _The Media Lab_: In 1979 anthropologist-philosopher Gregory Bateson offered another definition of "information": "Any difference which makes a difference." He said, "The map is not the territory, we're told. Very well. What is it that gets from the territory onto the map?" The cartographer draws in roads, rivers, elevations--things the map user is expected to care about. Data, signal ("news of a difference") isn't information until it means something or does something ("makes a difference"). The definition of information I kept hearing at the Media Lab was Bateson's highly subjective one. That's philosophically heartwarming, but it also turns out there's a powerful tool kit lurking in the redefinition. --Barry Kort ------------------------------ Date: Sat, 11 Jun 88 13:57:20 PDT From: Bob Riemenschneider Subject: Re: definition of information => It is often acknowledged that information theory has nothing to say => about information in the usual sense, as having to do with meaning. => ... => => Can anyone point me to a coherent definition of information respecting => information content, as opposed to merely "quantity of information"? => => Bruce Nevin => bn@cch.bbn.com Actually, much the same formalization applies to "real" information. See R. Carnap and Y. Bar-Hillel, "An Outline of a Theory of Semantic Information", Technical Report 247, Research Laboratory of Electronics, MIT, October 1952. (Reprinted in Y. Bar-Hillel, _Language and Information_, Addison-Wesley, 1964.) J. Hintikka, "On Semantic Information", in: J. Hintikka and P. Suppes (eds.), _Information and Inference_, Reidel, 1970. for starters. I'm not sure what you mean by `respecting information content', but this approach *is* based on analysis of the logical consequences of messages. -- rar ------------------------------ Date: 12 Jun 88 13:49:00 EDT From: Nahum (N.) Goldmann Subject: def of info In AIList Digest V7 #26 Bruce Nevin asks: >>Can anyone point me to a coherent definition of information respecting >>information content, as opposed to merely "quantity of information"? Having analyzed several dozens of various definitions I, of course, havily favour my own (see N. Goldmann, Online Research and Retrieval... ISBN 0-8306-1947-X, Chapter 2). I believe it is about "meaning" as opposed to "statistics". Greetings Nahum Goldmann acoust@bnr.ca (613)763-2329 ------------------------------ Date: Mon, 13 Jun 88 09:43:03 EDT From: "Bruce E. Nevin" Subject: Re: definition of information My understanding is that Carnap and Bar-Hillel set out to establish a "calculus of information" but did not succeed in doing so. Communication theory refers to a physical system's capacity to transmit arbitrarily selected signals, which need not be "symbolic" (need not mean or stand for anything). To use the term "information" in this connection seems Pickwickian at least. "Real information"? Do you mean the Carnap/Bar-Hillel program as taken up by Hintikka? Are you saying that the latter has a useful representation of the meaning of texts? Bruce Nevin bn@cch.bbn.com ------------------------------ End of AIList Digest ********************