Copy Link
Add to Bookmark
Report
AIList Digest Volume 5 Issue 279
AIList Digest Monday, 7 Dec 1987 Volume 5 : Issue 279
Today's Topics:
Announcement - Prolog Source Library,
Seminar - Composing and Decomposing Universal Plans (SRI),
Conference - AI Workshop in Singapore 1989
----------------------------------------------------------------------
Date: 3-DEC-1987 22:48:59 GMT
From: POPX@VAX.OXFORD.AC.UK
Subject: Prolog Source Library
From: Jocelyn Paine,
St. Peter's College,
New Inn Hall Street,
Oxford OX1 2DL.
Janet Address: POPX @ OX.VAX
PROLOG SOURCE LIBRARY
I teach AI to undergraduates, as a one-term practical course in the
Experimental Psychology degree. For the course, I use Poplog Prolog, on
a VAX under VMS. During the course, I talk about topics like scripts,
mathematical creativity, planning, natural language analysis, and expert
systems; I exemplify them by mentioning well-known programs like GPS,
Sam, and AM.
I would like my students to be able to run these programs, and to
investigate their mechanism and limitations. For students to incorporate
into their own programs, I'd also like to provide a library of Prolog
tools such as chart parsers, inference engines, search routines, and
planners. Unfortunately, published descriptions of the famous programs
give much less information than is necessary to re-implement them. As
for tools like planners and inference engines: the literature is often
more helpful, but I still have to do a lot of work which must have been
done before, even if it's merely typing in code from excellent textbooks
like "The Art of Prolog".
I'm sure other Prolog programmers have this problem too.
I have therefore set up a LIBRARY OF PROLOG SOURCE CODE, which I will
distribute over the British academic network (Janet) and nets like
Bitnet connected to Janet, to anybody who wants it. I will take
contributions from anyone who wants to provide them, subject to a few
conditions mentioned below. I proposed this in AIList Bulletin V5 267:
here are the details of how the library works. If you want to contribute
entries, or to request them, please read on...
How to send contributions.
Please send Prolog source for the library, to user POPX at Janet
address OX.VAX (the Vax-Cluster at Oxford University Computing
Service). If a file occupies more than about 1 megabyte, please send a
short message about it first, but don't send the large file itself
until I reply with a message requesting it. This will avoid the
problems we sometimes have where large files are rejected because
there isn't enough space for them.
I accept source on the understanding that it will be distributed to
anyone who asks for it. I intend that the contents of the library be
treated in the same way as (for example) proofs published in the
mathematical literature, and algorithms published in computer science
textbooks - as publicly available ideas which anyone can experiment
with, criticise, and improve.
I will try to put an entry into the library within one working week of
its arrival.
Catalogue of entries.
I will keep a catalogue of contributions available to anyone who asks
for it.
The catalogue will contain for each entry: the name and geographical
address of the entry's contributor (to prevent contributors receiving
unwanted electronic mail, I won't include their electronic mail
addresses unless I'm asked to do so); a description of the entry's
purpose; and an approximate size in kilobytes (to help those whose
mail systems can't receive large files easily).
I will also include my evaluations of its ease of use, of its
portability and standardness (by the standards of Edinburgh Prolog);
and my evaluation of any documentation included.
Quality of entries.
Any contribution may be useful to someone out there, so I'll start by
accepting anything. I'm not just looking for elegant code, or logical
respectability. However, it would be nice if entries were to be
adequately documented, to come with examples of their use, and to run
under Edinburgh Prolog as described in "Programming in Prolog" by
Clocksin and Mellish. If you can therefore, I'd like you to follow the
suggestions below.
The main predicate or predicates in each entry should be specified
so that someone who knows nothing about how they work can call them.
This means specifying: the type and mode of each argument, including
details of what must be instantiated on call, and what will have
become instantiated on return; under what conditions the predicate
fails, and whether it's resatisfiable; any side-effects, including
transput and clauses asserted or retracted; whether any initial
conditions are required, including assertions, operator
declarations, and ancilliary predicates. In some cases, other
information, like the syntax of a language compiled by the
predicate, may be useful.
A set of example calls would be useful, showing the inputs given,
and the outputs expected. Use your discretion: if you contribute an
expert system shell for example, I'd like a sample rulebase, and a
description of how to call the shell from Prolog, and some
indication of what questions I can ask the shell, but I don't
require that the shell's dialogue be reproduced down to every last
carriage return and indentation.
For programmers who want to look inside an entry, adequate comments
should be given in the source code, together perhaps with a more
general description of how the entry works, including any relevant
theory.
In the documentation, references to the literature should be given,
if this is helpful.
Entries should be runnable using only the predicates and operators
described in "Programming in Prolog" (if they are not, I may not be
able to test them!). I don't object to add-on modules being included
which are only runnable under certain implementations - for example,
an add-on with which a planner can display its thoughts in windows
on a high-resolution terminal - but they will be less generally
useful.
As mentioned earlier, I will evaluate entries for documentation and
standardness, putting my results into the catalogue. If I can, I
will also test them, and record how easy I found them to use, by
following the instructions given.
I emphasise that I will accept all entries; the comments above suggest
how to improve the quality of entries, if you have the time.
Requesting entries.
I can't afford to copy lots of discs, tapes, papers, etc, so I can
only deal with requests to send files along the network. Also, I can't
afford to send along networks that I have to pay to use from Janet.
You may request the catalogue, or a particular entry in it. I will
also try to satisfy requests like "please send all the natural
language parsers which you have" - whether I can cope with these will
depend on the size of the library.
I will try to answer each request within seven working days. If you
get no reply within fourteen working days, then please send a message
by paper mail to my address. Give full details of where your
electronic mail messge was sent from, the time, etc. If a message
fails to arrive, this may help the Computing Service staff discover
why.
Although I know Lisp, I haven't used it enough to do much with it,
though I'm willing just to receive and pass on Lisp code, and to try
running it under VAX Lisp or Poplog version 12 Lisp.
------------------------------
Date: Thu, 3 Dec 87 16:03:52 PST
From: Amy Lansky <lansky@venice.ai.sri.com>
Subject: Seminar - Composing and Decomposing Universal Plans (SRI)
COMPOSING AND DECOMPOSING UNIVERSAL PLANS
Marcel Schoppers
Advanced Decision Systems (MARCEL@ADS.ARPA)
11:00 AM, MONDAY, December 7
SRI International, Building E, Room EJ228
``Universal plans'' are representations for robot behavior; they are
unique in being both highly reactive and automatically synthesized. As
a consequence of this plan representation, subplans have conditional
effects, and hence there are conditional goal conflicts. When block
promotion (= subplan concatenation) cannot remove an interaction, I
resort not to individual promotion (= subplan interleaving) but to
confinement (falsifying preconditions of the interaction). With
individual promotion out of the way, planning is a fundamentally
different problem: plan structure directly reflects goal structure,
plans can be conveniently composed from subplans, and each goal
conflict needs to be resolved only once during the lifetime of the
problem domain. Conflict analysis is computationally expensive,
however, and interactions may be more easily observed at execution
time than predicted at planning time.
All conflict elimination decisions can be cached as annotated
operators. Hence it is possible to throw away a universal plan, later
reconstructing it from its component operators without doing any
planning. Indeed, an algorithm resembling backchaining mindlessly
reassembles just enough of a universal plan to select an action that
is helpful in the current world state. Since the selected action is
both a situated response and part of a plan, recent rhetoric about
situated action as *opposed* to planning is defeated.
VISITORS: Please arrive 5 minutes early so that you can be escorted up
from the E-building receptionist's desk. Thanks!
------------------------------
Date: Thu, 03 Dec 87 10:46:56 SST
From: Joel Loo <ISSLPL%NUSVM.BITNET@wiscvm.wisc.edu>
Subject: Conference - AI Workshop in Singapore 1989
Thanks for those who expressed interest in the call for papers
posted by me recently. Due to the overwhelming queries, it might
be beneficial to post a detailed one here for your convenience.
2nd International IFIP/IFAC/IFORS
Workshop on
ARTIFICIAL INTELLIGENCE
IN ECONOMICS AND MANAGEMENT
SINGAPORE
January 11-13, 1989
Organized by the Institute of Systems Science
National University of Singapore
+-----------------+
! CALL FOR PAPERS !
+-----------------+
The Second International Conference on AI in Economics and
Management will be held in Singapore during the 2nd week of
January 1989. The workshop will address issues relevant to the
use of AI Technology in Economic and Management communities.
Topics for the workshop will cover both technology and
applications.
Professor Herbert Simon, Nobel Laureate will be the Keynote
Speaker.
This workshop will address research and applications of
artificial intelligence techniques and tools, in the areas of:
finance, accounting, marketing, banking, insurance, economics,
human-resource management, assets adminstration, decision
support systems, public and private services, office automation,
law, and manufacturing planning.
The techniques to be presented should be explicitly relevant to
the above application areas, and include: knowledge
representation, search and inference, knowledge acquisition,
intelligent interfaces, knowledge base validation, natural
language analysis, planning procedures, and task support systems.
The tools to be presented should also be specific in design or in
use to the application areas discussed at the workshop, and may
cover: application specific expert systems, front-ends to
decision support systems, interfaces to database systems,
interfaces between symbolic and procedural processors, object
oriented environment.
The workshop will have contributed papers and case sessions.
There will be separate tutorials on the use of AI technology on
January 9 & 10.
** Paper Submission Procedure **
Authors should submit 700 word extended abstracts, typed with
double-spacing, in 2 copies before July 1, 1988 to:
Mrs Vicky TOH
Institute of Systems Science
National University of Singapore
Kent Ridge
Singapore 0511
Each abstract should include full address of all authors, and
references in numerical order. Authors of accepted submission
will be notified by September 1, 1988. Papers not received in
full by this date will not be included for presentation. All
papers must be in English.
** Software Submission Procedure **
Authors not willing to submit a paper, but ready to demonstrate
an artifical intelligence software program are encouraged to do
so. The submission procedure is the same as for papers. The host
computers, operating systems, utilities and all interfaces must
be specified exactly, as well as the architecture and principles
underlying the program. Authors will have to be responsible for
all logistics, including supply of computers etc.
All authors of accepted papers or of accepted software demos, are
expected to present their work in person. Failure to do so will
result in the corrsponding paper not appearing in the workshop
proceedings.
** Exhibit **
Companies interested in exhibiting publications equipment or
software falling within the scope of the workshop, should contact
the organizing committee.
---------------------------------
Important Dates
Tutorials : 9 & 10 Jan 1989
Workshop : 11-13 Jan 1989
For submission
of extended
abstract : 1 Jul 1988
Notification of
Acceptance : 1 Sep 1988
Camera Ready
Papers Due : 1 Nov 1988
---------------------------------
Language : Throughout the workshop, English will be the
official language. Translation facilities will NOT
be available.
Proceedings : Proceedings will be published after the workshop,
with Y.H. Pao, L.F.Pau, J.Motiwalla and H.H.Teh as
editors. Copyrights for accepted papers are thus
transferred to the publishers.
Registration: US$200 for Tutorials
Fees US$200 for Workshop
US$300 for the complete Workshop & Tutorials.
(fees cover freshments, lunches and conference
documentation)
Hotels : The price range for 5-star hotels in Singapore is
US$50-US$75
Travel : Arrangements will be made for special excursion air
fares.
(Request for information should be directed as well to Mrs Vicky
Toh at the above address (Telex: ISSNUS RS 39988, Fax: 7782571,
BITNET: ISSVCT@NUSVM))
*** Conference Committee ***
Chairman : Juzar MOTIWALLA, Institute of Systems Science,
National University of Singapore
Program Committee Chairmen:
Yoh-Han PAO, Case Institute of Technology, US
L.F. PAU, Technical University, Denmark
Hoon-Heng TEH, Institute of Systems Science, Singapore
Organizing Committee Chairman:
Desai NARASIMHALU, Institute of Systems Science, Singapore
*** International Program Committee *** (tentative)
Jan Alkins, AION, US
Jason Catlett, Univ. of Sydney, AUS
C.H. Hu, Academy of Sciences, PRC
Jae Kyu Lee, KAIST, KOREA
Peng Si Ow, CMU, US
Suzanne Pinson, Univ. of Paris, FRANCE
Edison Tse, Stanford Univ., US
Andrew Whinston, Purdue Univ., US
------------------------------
End of AIList Digest
********************
