AIList-REQUEST@SRI-AI.ARPA (AIList Moderator Kenneth Laws) (09/18/85)
AIList Digest Wednesday, 18 Sep 1985 Volume 3 : Issue 124
Today's Topics:
Query - AI Conference Listings,
Correction - Expert Systems in Government,
Volunteers Needed - Expert Systems in Government,
Call for Papers - Knowledge-Based Expert Systems (IEEE Software),
Seminars - Cognitive Science Calendar (MIT) &
Learning Mathematical Abstractions (UCB) &
Programming in Equational Logic (CMU) &
Explanation of Quantitative Models (CMU) &
Automating CAD Design (GMR),
Conference - Workshop on Argument Structure (Brandeis)
Course - The Logic of Robot Design (SU)
----------------------------------------------------------------------
Date: Wed, 18 Sep 85 09:47 EST
From: Atul Bajpai <bajpai%gmr.csnet@CSNET-RELAY.ARPA>
Subject: AI Conferences
Can somebody out there in the computer world provide me with a list
of major AI conferencs/symposiums/tutorials that are scheduled
through the end of 1986? Location and dates of these activities would
be useful. Thank you.
--Atul Bajpai-- CSNET: bajpai%gmr.csnet@csnet-relay
[A couple of good sources for such information are the ACM SIGART
newsletter and the monthly Communications of the ACM. IEEE Computer
carries both a calendar and calls for papers, and the AI Magazine
from AAAI has quite a few conference announcements. Yoni Malachi
(YM@SAIL) has been keeping an indexed list of conference announcements
in file confer.txt[2,2] at SAIL (Stanford AI Lab). Back issues of
AIList are a pretty good source as well. -- KIL]
------------------------------
Date: Fri 13 Sep 85 23:38:29-PDT
From: Gary Martins <GARY@SRI-CSLA.ARPA>
Subject: False advertising for "Expert Systems in Gov't"
Ailist V3 #115 includes a message dated August 28, from Marhshall Abrams
at Mitre, describing the program for a forthcoming conference on
"Expert Systems In Government".
My name appears as a panelist in a session entitled "Frontiers of KBES:
Pro & Con".
AIList readers should know that neither I nor anyone else from
Intelligent Software Inc. will be appearing in any capacity at this
festival of hype.
I have a written promise from Marshall Abrams to desist from this false
advertising. He has known for several weeks that this announcement is
false.
Misrepresentation, exaggeration, and falsehood are important elements
of contemporary AI. We would only ask that these be confined to the
technical presentations, and not spill over into administrative and
procedural announcements.
Thanks.
Gary R. Martins
------------------------------
Date: 18 Sep 85 11:12:05 EDT (Wed)
From: Marshall D. Abrams <abrams@mitre.ARPA>
Subject: Gary Martins will not be at ESIG
Due to a clerical error, the preliminary program for Expert Systems
in Government indicates that Gary Martins will participate on a panel
discussion.
I apologize for any inconvenience this error may have caused.
Sincerely,
- Marshall D. Abrams, phone: (703) 883-6938
The MITRE Corporation, 1820 Dolley Madison Blvd.
Mail Stop W458, Mc Lean, VA 22102
------------------------------
Date: 12 Sep 85 08:53:08 EDT (Thu)
From: Kamal N. Karna <m16045@mitre.ARPA>
Subject: Second EXPERT SYSTEMS IN GOVERNMENT CONFERENCE 1986
The second Expert Systems in Government Conference, sponsored by
the IEEE/CS and the MITRE Corporation, will be held in October
1986 in Washington D.C. metropolitan area.
Volunteers at all levels are solicited to participate in the
next year's program.
ESIG 1986 is anticipated to be much larger and broader in
scope. The ESIG 1986 conference shall include one day of
tutorials followed by the sessions on unclassified and classified
topics. There will be two technical co-program chairmen
to organize the technical programs.
People interested in participating in ESIG 1986 please
contact
Dr. Kamal N. Karna
Conference Chairman
Second Expert Systems in Government Conference 1986
The MITRE Corporation
1820 Dolley Madison Blvd.
McLean, Virginia 22102.
(Tel.) (703) 883-5866 (O)
(301) 921-0392 (H)
KARNA@MITRE----ARPANET
------------------------------
18-Sep-85 09:35:14-PDT,1332;000000000001
Date: Tuesday, 17 September 1985 14:52:06 EDT
From: Duvvuru.Sriram@cive.ri.cmu.edu
Subject: Call for submissions to IEEE Software
CALL FOR PAPERS
In addition to the regular features, the March 1986 IEEE Software
special issue on Knowledge-based Expert Systems (KBES) for Engineering
Applications will have a section on research being conducted on the
applications of KBES for engineering. Authors should submit a 3 page
(doubly spaced) paper, focusing on the following issues:
- goal of the project;
- architecture of the system;
- current status; and
- future plans.
All submissions should be sent before October 15th, 1985 to:
D. Sriram/M. Rychener
Civil Engg. and Construction Labs.
Department of Civil Engineering
Carnegie-Mellon University
Pittsburgh, PA 15213
ARPAnet address: sriram@cive.ri.cmu.edu or sriram@cmu-ri-cive.arpa
(Net mail is preferable)
------------------------------
Date: Sun, 15 Sep 1985 11:42 EDT
From: Peter de Jong <DEJONG%MIT-OZ@MIT-MC.ARPA>
Subject: Cognitive Science Calendar
Reply-to: Cog-Sci-Request%MIT-OZ@MIT-MC
[Forwarded from the MIT bboard by SASW@MIT-MC.]
The Cognitive Science Calendar is a weekly listing of Cognitive
Science talks with their abstracts which are given at MIT and the
surrounding area. The listing includes the areas of Artificial Intelligence,
Cognitive Psychology, Philosophy, and Linguistics.
If you would like to join the listing or have talks listed please
reply to COG-SCI-REQUEST%MIT-OZ@MIT-MC.ARPA.
------------------------------
Date: Thu, 12 Sep 85 09:03:50 PDT
From: chertok%ucbcogsci@Berkeley (Paula Chertok)
Subject: Seminar - Learning Mathematical Abstractions (UCB)
BERKELEY COGNITIVE SCIENCE PROGRAM
Fall 1985
Cognitive Science Seminar -- IDS 237A
TIME: Tuesday, September 17, 11:00 - 12:30
PLACE: 240 Bechtel Engineering Center
(followed by)
DISCUSSION: 12:30 - 1:30 in 200 Building T-4
SPEAKER: Alan H. Schoenfeld, Education & Mathematics, UCB
TITLE: ``Obstacles To Making Sense of Mathematical
Notions, or, The Transfer Problem Rears its
Ugly Head Once Again''
It can be argued that the fundamental difficulty in mathemat-
ics learning is the transfer problem. That is, the power of
mathematics lies in the potential applicability of mathemati-
cal ideas to new situations. It doesn't matter whether the
idea is, for example, function, group, number, or triangle.
Once any particular mathematical entity is recognized as
belonging to an identified class of objects, everything known
about that class of objects applies to that entity. In such
abstaction resides much of the power and utility of mathemat-
ics. This paper explores some theoretical and some pragmatic
obstacles to students' abstraction of mathematical notions.
We look at two domains, whole number arithmetic and plane
geometry. Some parallels between the two domains are drawn,
to indicate that the processes of abstraction are similar in
both. In the case of number, we examine a theoretical para-
dox: the use of good ``hands on'' manipulatives to help stu-
dents make sense of base 10 addition and subtraction may make
it harder to understand the nature of ``number.'' In the case
of geometry, we discuss an empirical obstacle. When things
are compartmentalized in the curriculum, connections that we
would hope are ``natural'' turn out to be very hard to make.
------------------------------
Date: 17 September 1985 1308-EDT
From: Theona Stefanis@CMU-CS-A
Subject: Seminar - Programming in Equational Logic (CMU)
Name: Michael J. O'Donnell - The University of Chicago
Date: Monday, 23 September
Time: 3:00
Place: WeH 5409
Title: Equational Logic as a Programming Language
In logic programming languages programs are logical assertions with no
explicit procedural information, and execution consists of the efficient
derivation of certain logical consequences of a program.
Prolog, and relational database query languages are both logic
programming languages based on the predicate calculus. The advantages
of logic programming are clarity of programs, simplicity of semantics,
and the potential for parallel execution without timing dependence.
In this talk, I describe a programming language based on the logic of
equations. A prototype implementation exists, and has been used in a
number of experiments. The talk will focus on examples illustrating the
advantages of equational programming, and the differences between
equational programming and Prolog programming.
Mike O'Donnell
------------------------------
Date: 17 Sep 85 16:16:10 EDT
From: Jeanne.Bennardo@CMU-RI-ISL1
Subject: Seminar - Explanation of Quantitative Models (CMU)
Intelligent Systems Lab Seminar
Topic: Presentation of Ergo Project
Speaker: Dr. Donald Kosy
Place: DH3313
Date: Friday, September 20
Time: 11:00am - 12:00noon
The Ergo project is an investigation into computer methods for explaining
results produced by quantitative planning models. It is an outgrowth of
previous research in which we developed a procedure that explains differences
in results by comparing their derivations. For example, given a suitable
quantitative model and a series of results from it such as:
1982 1983 1984
Profit $683.00 $676.40 $951.60
this procedure can answer "why" questions about the results such as:
Why did profit go down in 1983?
Profit went down in 83 primarily because overhead cost went up and
profit = gross margin - overhead cost. Although gross margin went up
by 19%, overhead cost went up by 41% and the latter outweighed the
former. Would you like me to continue?
The explanation can be continued all the way down to the lowest level of the
model, i.e. down to its input variables.
The intent of the current work is to extend the explanatory power of this
procedure and to evaluate its potential for practical use. In particular
we are interested in extensions to cover:
*Models that involve simultaneous equations
*Models that involve conditional expressions
*Explanations of differential magnitude, e.g. Why did overhead cost
go up so much?
*Inferring the referents of comparison, e.g. Why is production cost
high in 83?
*Explaining lack of change, e.g. Why did material cost remain
constant?
*Explaining differences between "what-if" cases and a base case, e.g.
Why is profit in case 1 higher than for the base case?
------------------------------
Date: Thu, 12 Sep 85 14:24 EST
From: "S. Holland" <holland%gmr.csnet@csnet-relay.arpa>
Subject: Seminar - Automating CAD Design (GMR)
Presentation is to be held at General Motors Research Laboratories in
Warren, Michigan.
Function, Form and Fabrication:
Considerations in Automating Design
James R. Rinderle
Assistant Professor
Carnegie-Mellon University
Monday, September 16, 1985
Any design modification requires knowledge of whether changed features
affect the functionality of a product, impact the producibility or are
simply arbitrary specifications. Good designers simultaneously consider
the form of the product, the functional requirements and the means for
fabrication. Even the most modest attempts at automating design are doomed
to fail unless these considerations are included at some level.
A CAD database should consist of multiple representations of a product in
terms of function, form and fabrication analogous to the traditional
specification list, drawings and process plan. The relationships among
these representations is the basis for design. Design modification may be
facilitated by developing a syntax for describing function, form, and
fabrication and by developing strategies for manipulating the descriptions.
Dr. James R. Rinderle is an Assistant Professor of Mechanical Engineering
at Carnegie-Mellon University. He was awarded the Ph.D. degree by the
Mechanical Engineering Department of the Massachusetts Institute of
Technology in 1982. He also received the Master of Science and Bachelor of
Science degrees from MIT. He is a member of Sigma Xi, Pi Tau Sigma, the
ASME, Robotics International and CASA of SME. He was named a Presidential
Young Investigator in 1985.
-Steve Holland
------------------------------
Date: Sat, 14 Sep 1985 03:00 EDT
From: INGRIA%MIT-OZ@MIT-MC.ARPA
Subject: Workshop on Argument Structure - Brandeis
The
Linguistics
and Cognitive Science Programs
Brandeis University
Present
A Workshop on
Argument
Structure
Friday, September 20th
9AM to 5PM
Sachar International Center
Presentations will be one (1) hour long with time for questions.
Speakers:
9:30 Ken Hale
``A View from the Middle''
11:00 Ray Jackendoff
``The Role of Thematic Relations in Linguistic Theory''
1:30 Edwin Williams
``Argument Structure and Heads''
2:45 Armin Mester and Jane Grimshaw
``Light Verbs in Japanese'
4:00 Howard Lasnik
``Subjects of NPs?''
5:00 - 7:00
Reception - Sachar International Center
For more information, call (617)-647-2986.
------------------------------
Date: Wed 18 Sep 85 08:22:53-PDT
From: Leslie Kaelbling <KAELBLING@SRI-AI.ARPA>
Subject: New course: The Logic of Robot Design (SU)
[Forwarded from the Stanford bboard by Laws@SRI-AI.]
THE LOGIC OF ROBOT DESIGN
This course will explore theoretical issues in the design of software
for intelligent agents. Its aim is to provide conceptual tools for
coping with complexity in robot design, covering processes from the
sensorimotor level through reasoning, planning, and linguistic
communication, emphasizing the role of formal methods in analysis and
synthesis of robot software.
The following topics will be covered:
- applications of epistemic and temporal logic to robotics
- automata-theoretic models of knowledge
- inference and planning
- logic-based tools for programming intelligent robots.
Some familiarity with basic logic and computer programming will be
assumed. Coursework will consist of problem sets and one programming
assignment.
Instructor : Stan Rosenschein (stan@sri-ai; 859-4167)
Time : TTh 11-12:15
Place : 460-252
Course number : CS428
Units : 3
TA : Leslie Kaelbling (kaelbling@sri-ai, pack@su-sushi,
k.kaelbling@su-lots-b; 859-2578)
------------------------------
End of AIList Digest
********************