Date: Wed 15 Jun 1988 02:04-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 #33 To: AIList@AI.AI.MIT.EDU Status: RO AIList Digest Wednesday, 15 Jun 1988 Volume 7 : Issue 33 Today's Topics: Philosophy: Who else isn't a science? scope of ailist Me, Karl, Stephen, Gilbert Definition of Information representation languages ---------------------------------------------------------------------- Date: 13 Jun 88 13:07:50 GMT From: marsh@mitre-bedford.arpa (Ralph J. Marshall) Subject: Re: Who else isn't a science? In article <10785@agate.BERKELEY.EDU> weemba@garnet.berkeley.edu writes: > >Indeed, many modern dictionaries now give an extra meaning to the word >"intelligent", thanks, partly due to AI's decades of abuse of the term: >it means "able to peform some of the functions of a computer". > >Ain't it wonderful? AI succeeded by changing the meaning of the word. > >ucbvax!garnet!weemba Matthew P Wiener/Brahms Gang/Berkeley CA 94720 I don't know what dictionary you are smoking, but _MY_ dictionary has the following perfectly reasonable definition of intelligence: "The ability to learn or understand or to deal with new or trying situations." (Webster's New 9th Collegiate Dictionary) I'm not at all sure that this is really the focus of current AI work, but I am reasonably convinced that it is a long-term goal that is worth pursuing. ------------------------------ Date: 14 Jun 88 07:25:00 EDT From: "CUGINI, JOHN" Reply-to: "CUGINI, JOHN" Subject: scope of ailist As a somewhat belated response to the complaints about endless philosophizing, I offer the following quote from H.G. Wells, "The Future in America", written in 1906, after Wells had toured the states. He was writing specifically about Washington, providing some additional poignancy for those of us who work in the DC area, but perhaps it has wider pertinence: It is perhaps near the truth to say that this dearth of any general and comprehensive intellectual activity is due to intellectual specialization. The four thousand scientific men in Washington are all too energetically busy with ethnographic details, electrical computations or herbaria, to talk about common and universal things. They ought not to be so busy, and a science so specialized sinks halfway down the scale of sciences. Science is one of those things that cannot hustle; if it does it loses its connexions. In Washington some men, I gathered, hustle, others play bridge, and general questions are left a little comtemptuously, as being of the nature of "gas," to the newspapers and magazines. Philosophy, which correlates the sciences and keeps them subservient to the universals of life, has no seat there. My anticipated synthesis of ten thousand minds refused, under examination, to synthesize at all; it remained disintegrated, a mob, individually active and collectively futile, of specialists and politicians. John Cugini ------------------------------ Date: Tue, 14 Jun 88 08:18:37 -0400 (EDT) From: David Greene Subject: Re: Me, Karl, Stephen, Gilbert In AIList Digest V7 #29, Stephen Smoliar writes: > What have all those researchers who don't spend so much > time with computer programs have to tell us? I'm not advocating Mr. Cockton's views, but the limited literature breadth in many AI papers *is* self-defeating. For example, until very recently, few expert system papers acknowledged the results of 20+ years of psychology research on Judgement and Decision Making. It seems odd that AI people studying experts decision making would not reference behavioral/ performance research on human/ expert decision making. The works of Kahneman, Tversky, Hogarth and Dawes (to name some luminaries), all identify inherent flaws in human (including experts') judgement. These dysfunctional biases result in consistent suboptimal decision rules across many realistic conditions (setting aside debates on "optimality"). Yet, AI researchers and knowledge engineers attempt to produce fidelity to the expert and compare the resultant system to the experts performance. Is it a wonder that many ES's don't work in the field... Perhaps a broader literature/ research exposure could be advantageous to AI (or any field)... -David dg1v@andrew.cmu.edu Carnegie Mellon "You're welcome to use my oppinions, just don't get them all wrinkled..." ------------------------------ Date: Tue, 14 Jun 88 07:52:42 PDT From: golden@frodo.STANFORD.EDU (Richard Golden) Subject: Re: Definition of Information 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"? This question is really related to an earlier discussion concerned with viewing probability theory as a measure of belief. We can think of a knowledge structure as being represented by a probability distribution which assigns some "degree of belief" (i.e., a probability) to some set of events (i.e., a sample space). Let X be an event which occurs with probability p(X). Then clearly an equivalent "knowledge structure" which assigns some "degree of surprise" (i.e., -LOG[p(X)]) to some set of events (i.e., a sample space) may be constructed. The simple point which I am making is that the SAMPLE SPACE and the STRUCTURE OF ITS ELEMENTS is a necessary component of the definition of information in a technical sense and information CONTENT (for the most part) resides in this SAMPLE SPACE. Richard Golden (golden@psych) ------------------------------ Date: Tue, 14 Jun 88 10:42:12 bst From: Ian Dickinson Subject: Re: representation languages / otter:comp.ai.digest / vierhout@swivax.UUCP (Paul Vierhout) / writes: > 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 ? If you read the various papers from Chandresakaran's group, you will see that one of their central hypotheses is that you cannot define a single language that is usable to write all applications. Each generic task (and I think there are rather more than six) will have its own language, which specialises control and data structures to that task. In fact, they seem to anticipate a spectrum of languages each individually suited to *part* of the application being What we absolutely *must* avoid in defining representation languages of the future is the "good feature explosion" - ie adding new features like frames, rules, backwards, forwards and side ways inference, 12 inheritiance schemes, etc - simply because they could be useful in some circumstance. This is the route to the KEE's and ART's ++ of future representation schemes. Whilst I have no doubt that these systems are useful today, _I_ as an application developer want to see a representation system that is maximally small whilst giving me the power that I need. The philosophy I would like to see adopted is: o define conceptual representations that allow applications to be written at the maximum level of abstraction (eg generic tasks) o define the intermediate representations (frames, rules, sets ..) that are needed to implement the conceptual structures o choose a subset of these representations that can be maximally tightly integrated with the base language of your choice (which would not be Lisp in my choice) By doing this, we can not only help the application developer by giving her access to all of the abstraction power in the system, but also have a chance of getting the semantics of these systems properly understood and defined. ++ KEE and Art are registered trademarks. Ian. +---------------------+--------------------------+------------------------+ |Ian Dickinson net: All opinions expressed | |Hewlett Packard Labs ijd@otter.hplabs.hp.com are my own, and not | |Bristol, England ijd@hplb.uucp necessarily those of | |0272-799910 ..!mcvax!ukc!hplb!ijd my employer. | +---------------------+--------------------------+------------------------+ ------------------------------ End of AIList Digest ********************