Copy Link
Add to Bookmark
Report
AIList Digest Volume 5 Issue 195
AIList Digest Monday, 10 Aug 1987 Volume 5 : Issue 195
Today's Topics:
Queries - Reviews of Dreyfus and Dreyfus? & Natural Language Inc. &
Multiple Copies of a Clause & Macsyma Sources,
AI Tools - Macsyma Sources & FBRL in Prolog,
Philosophy - Natural Kinds & AI and Science
----------------------------------------------------------------------
Date: 5 Aug 87 09:20 EDT
From: WAnderson.wbst@Xerox.COM
Subject: Reviews of Dreyfus & Dreyfus?
I am looking for reviews of the Dreyfus & Dreyfus book "Mind Over
Machines." Any and all references appreciated. Thanks,
Bill Anderson
<WAnderson.wbst@Xerox.COM>
[There was, of course, the preview by the brothers Dreyfus themselves
in the January 1986 Technology Review; that has been discussed at
length in AIList. Another review by Theodore Roszak appeared in
the April 3, 1986, New Scientist, pp. 46-47. Roszak doesn't add much
personal perspective, but views the book favorably: "AI's record of
barefaced public deception is unparalleled in the annals of academic
study." -- KIL]
------------------------------
Date: Wed, 5 Aug 87 09:49:05 EDT
From: Bruce Nevin <bnevin@cch.bbn.com>
Subject: Nat Lg Inc
A colleague asks me by Email about an outfit called Natural Language Inc.
He hears they claim to process `virtually unrestricted English text'
to create a relational database for query systems. He says they are
located in Berkeley. Anybody know more?
------------------------------
Date: Fri, 7 Aug 87 09:54:13 EDT
From: tim@linc.cis.upenn.edu (Tim Finin)
Subject: multiple copies of a clause in the DB
I'm studying various ways to extend Prolog's simple model of the
database (e.g. a flat, global collections of clauses) to a richer
hierarchical one with inheritance. I am trying to decide whether to
allow multiple instances of a clause in a resulting database view.
Most Prolog implementations, at least those descendant from DEC-10
Prolog, do allow the database to contain two identical clauses. Most
of the non-Prolog logic programming languages that I am familiar with
do not. I am interested in discovering what use, if any, people have
made of the ability to assert multiple copies of a clause into the
database.
I, for one, have never found a use for this in practice. In fact, it
has only effected my life by being a source of bugs. It is easy
enough to accidentally get multiple copies of a clause in the database
by consulting a file instead of reconsulting it or by defining the
same predicate in two different files. This can easily mess up your
program unless you use a rather pure logic programming style which
doen't depend on the order in which the clauses are stored in the
database.
Has anyone out there found a good use for this Prolog "feature"?
Tim
------------------------------
Date: 4 Aug 87 01:44:25 GMT
From: amdahl!meccts!cimcor!mike@ames.arpa (Michael Grenier)
Subject: Macsyma Sources
I'm looking for any PD or commercial sources or binaries of Macsyma
that will run on this Microport Unix System on the 286. Any ideas?
-Mike
ihnp4!meccts!cimcor!mike
------------------------------
Date: 5 Aug 87 17:55:52 GMT
From: jbn@glacier.stanford.edu (John B. Nagle)
Subject: Re: Macsyma Sources
A competing package is MuMath, from
Soft Warehouse
3615 Harding Av
Honolulu Hawaii 96816
808-734-5801
This is a symbolic math package written in a quaint but charming dialect
of LISP, for which an interpreter is provided. There are versions for the
Apple II and IBM PC, and recently a modern version for the PC has been
released. I've used the older version on some messy vector calculus problems
in my solid modelling work, and found it quite useful in dealing with the
grunt work of algebra and calculus. The heuristics aren't very powerful,
but the algorithms for the standard solution methods all seem to work.
Microsoft resells this package, when they remember it is in their product
line, but the developers are in Hawaii and one may as well deal directly
with them. Sometimes one of the developers answers the phone.
John Nagle
------------------------------
Date: Sat, 8 Aug 87 23:21:33 EDT
From: tim@linc.cis.upenn.edu (Tim Finin)
Subject: FBRL in Prolog
Date: Wed, 5 Aug 87 12:18 EDT
From: The Mad Debugger <emerson@uvm-gen.UUCP>
Subject: FBRL in Prolog
Does anyone know of any FBRL's written in Prolog or that support logical
inference? I know HSRL (from Carnegie-Mellon) and KRYPTON (from
XEROX PARC) have a logical basis to them, but both are written in LISP.
I am currently writing a FBRL interpreter embedded in C-Prolog, and would
like to ''compare notes'' with other such systems, if they're out there.
I would also appreciate any thoughts on the implementation of frame theory
in Prolog.
Thanks in advance,
Tom E.
You may want to start by looking at Pat Hay's paper "The Logic of
Frames" from the mid to late seventies. He gives a logical account
for the semantics underlying the basic ideas in FBRL's. The paper is
reprinted in Brachman and Levesque's book "Readings in Knowledge
Representation" published by Morgan Kaufmann (1985).
I can point you toward three things involving FBRLs in Prolog that you
may want to look at:
[1] WIth a group from RCA, I built a frame-based representation
language in prolog called PINE. We used it to build an expert system
for diagnosing faults in ATE equipment. It is described in:
FOREST - An Expert System for Automatic Test Equipment; Tim Finin, Pamela
Kleinosky and John McAdams; Proceedings of the First Conference on
Artificial Intelligence Applications; (IEEE), Denver, Colorado, 1984.
A somewhat longer version is available as
technical report MS-CIS-84-09, Dept. of Computer and Information Science,
University of Pennsylvania, Philadelphia PA 19104
[2] Arity Corp offers an expert systems building toolkit (written by
David Drager) which is based on a FBRL. It's written in Arity Prolog,
of course. It is really quite powerful. I'd characterize it as a
cross between EMYCIN and KEE.
[3] The AI research group at UNISYS's Paoli Research Lab has been
using a FBRL implemented in Prolog to build many of their expert
systems for quite some time. There system is called KNET and is
similar to KL-ONE. An early reference is:
KNET - A logic Based Associative Network Framework for Expert
Systems"; Freeman, M., L. Hirschman and D. McKay; SDC, A Burroughs
Company; technical memo LBS 12; Sept. 1983.
I believe that there are several descriptions of it in the open
literature, but I'm not sure where they can be found.
Tim
------------------------------
Date: 30 Jul 87 00:34:27 GMT
From: uunet!mnetor!utzoo!dciem!nrcaer!cognos!roberts@seismo.css.gov
(Robert Stanley)
Subject: Re: natural kinds
In article <1526@botter.cs.vu.nl> hansw@cs.vu.nl (Hans Weigand) writes:
> (3) anthropic/functional kinds, existing by virtue of readiness_to_hand
> Examples: chair, cup, house, knife, game
>.... Thus we may recognize an Eskimo iglo, and an African pile-dwelling both
>as "houses". I think it is not so much the form (iconicity) that matters,
>but rather that we feel that, when we would live in Greenland
>(resp. the jungle), we would naturally appreciate or use these things
>as houses too (to protect us against cold, dangers)....
This raises some very interesting points, most particularly the fact that
anthropic kinds cannot generally have simple definitions. A very young child
gets away with calling a crude drawing or sand castle a 'house', but an
architect or construction engineer sees a house in much more specific terms.
In fact, we are entering the realms of the working vocabulary, and what is the
lowest common denominator which allows for completely successful transfer
between two disparate working sets.
Perhaps a strong example will serve. Kenya became an independent nation in
1964, and was faced with the problem of codifying laws, and deciding on
official languages. The two numerically superior tribal groupings were the Luo
and the Kikuyu, each with their own language, but colonial administration had
been exclusively English (at least in writing), and the standard interlingua of
the whole East African coast was Swahili (an Arabic-based patois). To further
complicate the issue, the very powerful, nomadic tribe of the Masai (with their
own language) had do be taken into account.
English and Swahili both were adopted as official languages, and a determined
effort made to create a formal body of law in both. In the Swahili version is
a formal definition of house which runs to some 96 pages of text! Why?
Because the term house has a whole slew of legal meanings in English common
law, on which Kenya's laws are based, which are totally alien to many of the
Kenyan tribes, especially the nomadic Masai. Therefore, each and every such
legal referent has to be precisely defined.
I leave as an exercise to the reader.......
I am not sure that house or any other cultural artifact can be called a natural
object unless its cultural matrix is expressly defined as part of the object's
name. Or that all objects in a given grouping are stated to exist within an
explicitly defined cultural context. I am absolutely sure that when I say
house and an Eskimo says igloo we are not talking about the same thing at all.
In fact the only common denominator appears to be shelter from the elements in
the winter months, albeit those are different for the two of us.
--
Robert Stanley Compuserve: 76174,3024 Cognos Incorporated
uucp: decvax!utzoo!dciem!nrcaer!cognos!roberts 3755 Riverside Drive
or ...nrcaer!uottawa!robs Ottawa, Ontario
Voice: (613) 738-1440 - Tuesdays only (don't ask) CANADA K1G 3N3
------------------------------
Date: 3 Aug 87 20:29:44 GMT
From: smadar@jarre.rutgers.edu (Cabelli)
Subject: Structure, Function and Intention in Natural Kinds
Ken Laws writes:
>Semantic classification thus requires at least three viewpoints:
>structure, intended function, and perceived or implemented function.
There has been alot of research recently in machine learning on
formulating concepts with these viewpoints in mind. I am amazed at
the omittion of any relevant AI work in this discussion on natural
kinds! For example, no mention was made of Winston's work on learning
structural descriptions from functional definitions (AAAI-83), (I was
surprised Minsky omitted that work).
My work on "Formulating Concepts According to Purpose" (AAAI-87)
presents a prototype system which formulates definitions of a "cup" based
on the purpose for which an agent intends to use it (one specialized
notion of intention). If the agent intends to use a cup to drink hot
liquids from, one definition is automatically generated. If on the
other hand, the cup has an ornamental purpose, a different definition
can be formed.
The key idea of the technique is to simulate the plan of actions the
agent will go through in drinking hot liquids from a cup, (say POUR,
GRASP, PICKUP, DRINK). Then, computing the (weakest) preconditions of
this plan derives a functional description (must contain hot liquids,
must be graspable by agent with hot liquid, must be liftable, and so
on). A technique like Winston's is then used to compute the
structural description from the functional one.
Smadar Kedar-Cabelli
Rutgers University
------------------------------
Date: 02 Aug 87 2159 PDT
From: John McCarthy <JMC@SAIL.STANFORD.EDU>
Subject: AI and science
Like mathematics, philosophy and engineering, AI differs
from the (other) sciences. Whether it fits someone's definition
of a science or not, it has need of scientific methods including
controlled experimentation.
First of all, it seems to me that AI is properly part of
computer science. It concerns procedures for achieving goals
under certain conditions of information and possibility for action.
We can even consider it analogous to linear programming. Indeed if
achieving one's goals always consisted finding the values of
a collection of real variables that would minimize a linear
function of these variables subject to a collection of linear
inequalities, then AI would coincide with linear programming.
However, the relation between goals, available actions,
the information initially available and that can later be acquired
is sometimes more complex than in any of the branches of computer
sciences the main character of whose scientific treatment consists
of mathematical theorems. We don't have a mathematical formalization
of the general problem faced in AI let alone general mathematical
methods for their solution. Indeed what we know of human intelligence
doesn't suggest that a conventional mathematical formalization of
the problems intelligence is used to solve even exists. For this
reason AI is to a substantial degree an experimental science.
The fact that a general mathematical formalization of the problems
intelligence solves is unlikely doesn't make mathematics useless in AI.
Many aspects of intelligence are formalizable, and languages of
mathematical logic are useful for expressing facts about the common
sense world, and logical reasoning, especially as extended by non-monotonic
reasoning is useful for drawing conclusions.
In my view a large part of AI research should consist of the
identification and study of intellectual mechanisms, e.g. pattern
matching and learning. The problems whose computer solution exhibits
these mechanisms should be chosen for reasons of scientific perspicuousness
analogously to the fact that genetics uses fruit flies and bacteria.
A. S. Kronrod once said that chess is the {\it Drosophila} of artificial
intelligence. He might have been right, but the organizations that
support research have taken the view that problems should be chosen
for their practical importance. Sometimes it is as if the geneticists
were required to do their work with elephants on the grounds that
elephants are useful and fruit flies are not. Anyway chess has been
left to the sportsmen, most of whom only write programs, not scientific
papers and compete about who can get time on the largest computers or
get someone to support the construction of specialized chess computers.
Donald Norman's complaints about the way AI research is
conducted have some validity, but the problem of isolating
intellectual mechanisms and making experiments worth repeating is
yet to be solved, so it isn't just a question of deciding to
be virtuous.
Finally, I'll remark that AI is not the same as cognitive
psychology although the two studies are allied. AI concentrates
more on the necessary relations between means and ends, while
cognitive psychology concentrates on how humans and animals
achieve their goals. Any success in either endeavor helps the other.
Methodology in AI is worth studying, but acceptance of its results
should be moderated by memory of the behaviorist catastrophe in
psychology. Doctrines arising from methodological studies crippled the
science for half a century. Indeed psychology was only rescued by ideas
arising from the invention of the computer --- and at least partly ideas
originating in AI.
------------------------------
Date: 5 Aug 87 12:23:37 GMT
From: Gilbert Cockton <mcvax!hci.hw.ac.uk!gilbert@seismo.CSS.GOV>
Reply-to: Gilbert Cockton <mcvax!hci.hw.ac.uk!gilbert@seismo.CSS.GOV>
Subject: Re: AI, science, and pseudo-science
In article <8707270710.AA05885@ucbvax.Berkeley.EDU>
mckee@CORWIN.CCS.NORTHEASTERN.EDU writes:
a lot, but his go at describing types of not-quite-sciences is
interesting. For me, AI should be one of the
>
>* Interdisciplinary Sciences: Materials Science, Neuroscience
> (characterized by their subject matter not yielding coherently
> to any single experimental technique or theoretical paradigm.)
>
My criticism of AI is that most of the workers I meet are pretty
ignorant of the CRITICAL TRADITIONS of ESTABLISHED disciplines which
can say much about AI's supposed object of study. When AI folk do stop
hacking (LISP, algebra or logic - it makes no difference, logic finger
and algebra wrist are just as bad as the well known 'computer-bum'),
they may do so only to raid a few concepts and 'facts' from some
discipline, and then go and abuse them out of sight of the folk who
originally developed them and understand their context and deductive
limitations. What some of them do to English is even worse :-)
>(However, I can't resist throwing in my excuse: programming is fun;
> science is hard, often boring, work. Science is far more rewarding, though.)
I think the nail's been hit squarely on the head, but to programming we
should add amateur philosophy and idealist logic/algebra as other fun
pasttimes pursued instead of hard, critical, rigorous argument. I think
the major turn-off of AI work can be summed up as a complete lack of
candid scholarship. The same is unfortunately true for much
applications-driven research in computing. Without reining in AI (or
computer applications research) under proper disciplines, I can't
really see any prospect for workers developing their critical faculties
up to the highest standards of established disciplines.
NB - yes there are uncritical, unimaginative automata and disreputable
charlatans in all disciplines. But these sorts are not the type who
make a DISCIPLINE. AI seems to have few folk who do want it to be a
discipline.
--
Gilbert Cockton, Scottish HCI Centre, Ben Line Building, Edinburgh, EH1 1TN
JANET: gilbert@uk.ac.hw.aimmi ARPA: gilbert%aimmi.hw.ac.uk@cs.ucl.ac.uk
UUCP: ..!{backbone}!aimmi.hw.ac.uk!gilbert
------------------------------
End of AIList Digest
********************
