Copy Link
Add to Bookmark
Report
AIList Digest Volume 5 Issue 174
AIList Digest Sunday, 12 Jul 1987 Volume 5 : Issue 174
Today's Topics:
Seminars - Object-Oriented Databases (IBM) &
A Model for Distributed Planning (SU) &
Pengi: A Theory of Activity (UCB) &
Learning Conjunctive Concepts (SU),
Conferences - Concurrent Logic Programming, Open Systems Programming &
NEXUS meeting at AAAI &
Fifth International Machine Learning Conference &
Second International Conference on AI in Engineering
----------------------------------------------------------------------
Date: Wed, 01 Jul 87 13:54:22 PDT
From: IBM Almaden Research Center Calendar <CALENDAR@ibm.com>
Subject: Seminars - Object-Oriented Databases (IBM)
IBM Almaden Research Center
650 Harry Road
San Jose, CA 95120-6099
Excerpts from
RESEARCH CALENDAR
July 6 - 10, 1987
EFFICIENT SUPPORT FOR DERIVED OBJECTS IN RELATIONAL DATABASE SYSTEMS
E. Hanson, University of California at Berkeley
Comp. Sci. Sem. Tues., July 7 10:00 A.M. Room: B3-247
Recently, an incremental algorithm known as Algebraic View Maintenance
(AVM) was proposed for maintaining materialized copies of views.
Another incremental view maintenance algorithm called Rete View
Maintenance (RVM) is presented in this talk. RVM is based on the Rete
network, a type of discrimination network used to support efficient
forward chaining rule interpreters in expert systems shells. RVM is
known as a statically optimized view maintenance algorithm because the
execution plan for maintaining views is compiled in advance into the
Rete network. In contrast, AVM is dynamically optimized since an
execution plan for maintaining a view is found after each base
relation update that affects the view. A statically optimized
variation of AVM is also presented. Using algorithms for view
maintenance as a starting point, a collection of methods is proposed
to allow other kinds of derived objects to be maintained. These
include aggregates, database procedures, and views and procedures
containing aggregates.
Host: S. Finkelstein
EPIDEMIC ALGORITHMS FOR REPLICATED DATABASE MAINTENANCE
A. Demers, Xerox Palo Alto Research Center
Bay Area Syst. Symp. Fri., July 10 11:15 A.M. Room: Front Aud.
When a database is replicated at many sites, maintaining mutual
consistency among the sites in the face of updates is a significant
problem. This paper describes several randomized algorithms for
distributing updates and driving the replicas toward consistency. The
algorithms are very simple and require few guarantees from the
underlying communication system, yet they ensure that the effect of
every update is eventually reflected in all replicas. The cost and
performance of the algorithms are tuned by choosing appropriate
distributions in the randomization step. The algorithms are closely
analogous to epidemics, and the epidemiology literature aids in
understanding their behavior. One of the algorithms has been
implemented in the Clearinghouse servers of the Xerox Corporate
Internet, solving long-standing problems of high traffic and database
inconsistency.
Host: L.-F. Cabrera
OBJECT-ORIENTED INTERFACES TO RELATIONAL DATABASES
R. Cattell, SUN Microsystems
Bay Area Syst. Symp. Fri., July 10 1:30 P.M. Room: Front Aud.
Users of engineering workstations have requirements that traditional
database systems often do not address. A number of research projects
have recently examined addressing engineering requirements with
additional semantics that an "object-oriented" database system can
provide over a relational database system. My talk will focus on two
topics that have received relatively little attention: (1)
object-oriented end-user interfaces to databases exploiting the
capabilities of an engineering workstation, and (2) the *performance*
that these tools and engineering applications require from a database
system, without which additional semantics are useless. Examples will
be provided from some of our own database and user interface work
combining features of object-oriented and relational database models.
Users may graphically view and edit a database schema, view database
objects that span multiple relations, browse through databases by
pointing with a mouse, and display specialized objects such as
documents and images stored in a database. To quantify performance,
we have proposed a set of benchmarks that measure the simple
object-oriented operations that we believe engineering applications
most typically execute. I will discuss the results of performing
these benchmarks on several relational database systems, and the
implications for database system architecture for engineering
applications.
Host: L.-F. Cabrera
For further information on individual talks, please contact the host
listed above.
------------------------------
Date: Fri 3 Jul 87 14:17:35-PDT
From: Charlie Koo <KOO@Sushi.Stanford.EDU>
Subject: Seminar - A Model for Distributed Planning (SU)
A Model for Distributed Performance --
Synchronizing Plans among Intelligent Agents
via Communication
Charles C. Koo
July 8, Wednesday
9:00am - 10:00am
Room 352
Margaret Jacks Hall
In a society where a group of agents cooperate to achieve certain goals,
agents perform their tasks based on certain plans. Some tasks may interact
with tasks done by other agents. One way to coordinate those tasks is to let
a master planner generate a plan to begin with, and distribute tasks to
individual agents accordingly. However, there are two difficulties
with this approach, given that agents are resource-limited. First, the
master planner needs to know all the expertise that each agent has. The
amount of knowledge sharply increases with the number of specialties.
Second, the centralized planning process takes longer turn-around time than
if each agent plans for itself. This causes a lot of computing resources
not being utilized. Thus, distributed planning is desirable.
In this presentation, I will describe a model for synchronizing and
monitoring plans autonomously made by intelligent agents via communication.
The model suggests an planning algorithm that allows agents to plan in
parallel and then synchronize their plans via a commitment-based
communication vehicle. Represenation as well as reasoning issues in the
distributed environment will be addressed.
Communication plays an integral role in planning for synchronization
purposes. The communication vehicle includes a minimal set of protocols
that enables the synchronization, a set of communication operators and a
set of commitment tracking operators. The tracking operators provide means
to monitor the progress of plan execution, to prevent delays, and to modify
plans with less effort when delays happen. A deadlock detection scheme will
also be described.
------------------------------
Date: Mon, 6 Jul 87 08:48:17 PDT
From: teresa@ernie.berkeley.edu (teresa diaz)
Subject: Seminar - Pengi: A Theory of Activity (UCB)
Special Seminar
Phil Agre
Artificial Intelligence Laboratory
Massachusettes Institute of Technology
Pengi: An Implementation
of a
Theory of Activity
2:00 - 4:00 p.m.
Friday, July 10, 1987
1011 Evans Hall
AI has typically sought to understand the organized
nature of human activity in terms of the making and execu-
tion of plans. There can be no doubt that people use plans.
But before and beneath any plan-use is a continual process
of moment-to-moment improvisation. An improvising agent
might use a plan as one of its resources, just as it might
use a map, the materials on a kitchen counter, or a string
tied round its finger. David Chapman and I have been study-
ing the organization of the most common sort of activity,
the everyday, ordinary, routine, familiar, practiced,
unproblematic activity typified by activities like making
breakfast, driving to work, and stuffing envelopes. Our
theory describes the central role of improvisation and the
inherent orderliness, coherence, and laws of change of
improvised activity. The organization of everyday routine
activity makes strong suggestions about the organization of
the human cognitive architecture. In particular, one can
get along remarkably well with a peripheral system much as
described by Marr and Ullman and a central system made of
combinational logic. Chapman has built a system with such
an architecture. Called Pengi, it plays a commercial video
game called Pengo, in which a player controls a penguin to
defend itself against ornery and unpredictable bees. The
game requires both moderately complex tactics and constant
attention to opportunities and contingencies. I will out-
line our theory of activity, describe the Pengi program, and
indicate the directions of ongoing further research.
______________________________________________________________________
This information is also kept in usr/public/seminars.
------------------------------
From: Peter Karp <KARP@SUMEX-AIM.STANFORD.EDU>
Subject: Seminar - Learning Conjunctive Concepts (SU)
[Forwarded from the AFLB list.]
David Haussler from UC Santa Cruz will be giving a talk at the GRAIL
learning seminar this Thursday 7/9 at the Welch Road Conference room at
1:15. This is Room A1110 in Building A at 701 Welch Road, across from
the Stanford Barn.
Learning Conjunctive Concepts in Structural Domains
David Haussler
Department of Computer Science,
University of California, Santa Cruz, CA 95064
We study the problem of inductively learning conjunctive concepts from
examples on structural domains like the blocks world. This class of
concepts is formally defined and it is shown that even when each example
(positive or negative) is a two-object scene it is NP-complete to
determine if there is any consistent concept in this class. We
demonstrate how this result affects the feasibility of Mitchell's
version space approach and how it shows that it is unlikely that this
class of concepts is polynomially learnable from random examples in the
sense of Valiant. On the other hand, we show that for any fixed number
of objects per scene this class is polynomially learnable from random
examples if
(1) we allow a larger hypothesis space, or
(2) we answer cetrain types of queries in addition to providing
random examples.
------------------------------
Date: Mon, 6 Jul 87 14:43:14 PDT
From: Ken Kahn <Kahn.pa@Xerox.COM>
Reply-to: Kahn.pa@Xerox.COM
Subject: Conference - Concurrent Logic Programming, Open Systems
Programming
We are pleased to announce that Xerox PARC with support from AAAI will
host a workshop on concurrent logic programming, meta-programming, and
open systems programming on September 8 and 9 (the first business days
after the Fourth IEEE Symposium on Logic Programming in San Francisco).
Participation is by invitation only. The purpose of the workshop is to
promote informal scientific interchanges between members of various
laboratories doing research centered around concurrent logic programming
languages such Guarded Horn Clauses and KL1 at ICOT, Parlog at Imperial
College, FCP at Weizmann Institute of Science, and Vulcan at Xerox PARC.
Other topics of interest include meta-programming to support programming
abstractions and issues related to programming large open distributed
systems. The format of the workshop will consist of informal
presentations and discussions of work in progress. Presentations given
at the Fourth SLP are not to be repeated. There will be several panel
discussions on topics such as the different proposals for dataflow
synchronization in these languages, the role of meta-programming in
supporting abstractions, and why it is that there are several indepenent
implementation efforts for different dialects of concurrent logic
programming languages (or are they committed choice programming
langauges or open systems programming languages?).
Live demonstrations of software is encouraged. Available computers
include Xerox computers running Xerox Common Lisp, Vaxes under Unix
4.2BSD, Sun 3's, IBM PC's, and Macintoshes (SE and II).
We will not be covering participants' transportation or living expenses.
Lunches will be provided. We are expecting between 20 and 40
participants. If you are interested in coming, or know someone who
might be, please send a letter or electronic message indicating what you
would like to talk about or demo to:
Kenneth Kahn
Xerox PARC
3333 Coyote Hill Road
Palo Alto, CA 94304
(415) 494-4390
or
ArpaNet: Kahn.pa@Xerox.Com
Here's the preliminary list of invitees:
Ehud Shapiro, Weizmann Institute
Shmuel Klinger, Weizmann Institute
Vijay Saraswat, CMU
Leon Sterling, Case Western Reserve
Keith Clark, Imperial College
Steve Gregory, Imperial College
Andrew Davison, Imperial College
M. Huntbach, Imperial College
Mitsuhiro Kishimoto, Fujitsu
Y. Takayama, ICOT
A. Okumura, ICOT
Y. Kimura, ICOT
H. Seki, ICOT
T. Chikayama, ICOT
Kazonuri Ueda, ICOT
K. Furukawa, ICOT
Fernando Pereira, SRI
Tony Kusalik, Univ. of Saskatchewan
Leon Alkalaj, UCLA
Richard O'Keefe, Quintus Compter Systems
Bill Kornfeld
Lee Naish, Melbourne University
G. Levi, University of Pisa
Walter Wilson, IBM
M. Maher, IBM
Carl Hewitt, MIT
Will Clinger, Tektronics
Mark Miller, Xerox PARC
Danny Bobrow, Xerox PARC
Curtis Abbott, Xerox PARC
Ken Kahn, Xerox PARC
Eric Tribble, Xerox PARC
------------------------------
Date: Wed, 8 Jul 87 21:34:05 CDT
From: Dan Cerys <cerys@XX.LCS.MIT.EDU>
Subject: Conference - NEXUS meeting at AAAI
I don't recall seeing the following on these lists, but this meeting is
probably interesting to those interested in the TI Explorer. Please
post if it will appear before July 15.
The National Explorer Users' Society will meet during AAAI-87 in
Conference Room A of the Conference Center House at Seattle Center in
Seattle, Washington on Wednesday, the fifteenth of July from three
o'clock until six o'clock.
3:00 Welcome, Introductions, and Organization
Rich Acuff, Stanford University
3:20 Explorer II
Chuck Corley, Strategic Systems Engineering, TI
3:30 New Customer Support Offerings
Phil Campbell, Technical Support Center, TI
3:35 Release 3.0 Summary
Joyce Statz, User Interface Branch, TI
3:45 TGC, System Training
Jim Mynatt, AI Technical Consultant, TI
3:55 Networking, Namespace, and Generic Network Interface
Roger Frech, Networking Branch, TI
4:05 New Compiler Features
Merrill Cornish, member, Group Technical Staff, TI
4:15 Future Directions
Henry Carr, Explorer Software Development, TI
4:25 Educational Marketing Survey
John Alden, Educational Marketing, TI
4:30 Bi-Directional Question and Answer
5:10 Break into groups to talk about
Explorer II with Chuck Corley
LX/Multiprocessing with Kari Karhi
Networking with Roger Frech
TI Prolog with Dan Cerys
NEXUS, the National Explorer Users' Society, met last year as
the Explorer Users' Group at AAAI-86. The purpose of the group is to
share technical information about the Explorer. There are no dues or
membership fees. Membership is open to all Explorer users. To join,
send your name, address, phone number, and net address to either of the
following addresses:
Rich Acuff
Stanford University
251 Medical School Office Building
Stanford CA 94305
acuff@sumex-aim.stanford.edu
Glenda S. McKinney M/S 2201
Texas Instruments
P. O. Box 2909
Austin TX 78769
mckinney%dsg%ti-csl@csnet-relay
Conference Room A is on the second floor of the Conference
Center House, in the northeast corner. The Conference Center House is
across the plaza from the Coliseum.
------------------------------
Date: Fri, 10 Jul 87 11:01:21 EDT
From: laird@caen.engin.umich.edu (John Laird)
Subject: Conference - Fifth International Machine Learning Conference
CALL FOR PAPERS
FIFTH INTERNATIONAL CONFERENCE ON MACHINE LEARNING
Ann Arbor, Michigan
June 12-15, 1988
The Fifth International Conference on Machine Learning will be held at the
University of Michigan, Ann Arbor, during June 12-15, 1988. The goal of the
conference is to bring together researchers from all areas of machine learning.
The conference will have open attendance and registration fees.
REVIEW CRITERIA
In order to ensure high quality papers, each submission will be reviewed
by two members of the program committee and judged on clarity, significance,
and originality. The best papers will be published in the proceedings, and
their authors will be invited to give a talk on their work or describe it at
a poster session. All submissions should contain new work, new results, or
major extensions to prior work. Summaries and overviews are discouraged.
The ideal paper will present a clear description of the learning task being
addressed and the proposed solution to that problem. If the paper describes
a running system, it should explain that system's representation of inputs
and outputs, its performance component, and its learning methods. It should
include a detailed example, as well as relate the work to earlier research.
Most important, all papers should include some evaluation of the work in the
form of substantive results. Papers are not required to take this form, but
authors are strongly encouraged to follow this format.
SUBMISSION OF PAPERS
Each paper must have a cover sheet with the title, authors' names, primary
author's address and telephone number, and an abstract of about 200 words. The
cover page should also give three keywords that specify the problem area,
general approach, and evaluation criteria. Some examples of each are:
PROBLEM AREA GENERAL APPROACH EVALUATION CRITERIA
Concept learning Genetic algorithms Empirical evaluation
Learning and planning Empirical methods Theoretical analysis
Language learning Explanation-based Psychological validity
Learning and design Connectionist
Machine discovery Analogical reasoning
The body of the paper must not exceed 13 double-spaced pages in twelve point
font, including figures but excluding references. Authors should send four
copies of their papers to:
Machine Learning Conference
Cognitive Science and Machine Intelligence Laboratory
The University of Michigan
904 Monroe Street
Ann Arbor, MI 48109-1234
Internet: ml88@csmil.umich.edu
The deadline for submission of papers is January 15, 1988. Authors will be
notified of acceptance by March 1, 1988. Final camera-ready copies of the
papers will be due April 1, 1988.
Organizing Committee
J. E. Laird (chairman) University of Michigan
J. H. Holland, S. L. Lytinen, G. M. Olson University of Michigan
J. G. Carbonell, T. M. Mitchell Carnegie-Mellon University
P. Langley University of California, Irvine
R. S. Michalski University of Illinois
------------------------------
Date: Fri, 10 Jul 87 23:40:17 EDT
From: sriram@ATHENA.MIT.EDU
Subject: Conference - Second International Conference on AI in
Engineering
SECOND INTERNATIONAL CONFERENCE ON
AI IN ENGINEERING
The program agenda for the above conference, which is to be held in
Boston from August 3-8, 1987, can be obtained from Ms Sandra Elliott,
Computational Mechanics Inst., 25 Bridge Street, Billerica, MA 01821,
USA (Tel. No: (617)667 5841). I have a copy of the agenda online. If
you are interested in getting a copy, send me mail.
Some program highlights:
Keynote speaker: Dr. Randy Davis, MIT, USA
Banquet speaker: Dr. Mark Stefik, Xerox,USA
Invited speakers:
Dr. John Gero, Univ. of Sydney, Australia
Dr. Jean-Claude Latombe, ITMI, France
(Currently atStanford Univ.)
Dr. B. Chandrasekaran, OSU, USA
Panels on:
AI in Mechancial Engineering: The Commerical Reality
AI in Electrical Engineering: The Commerical Reality
AI in Engineering Design: The Research Issues
AI in Engineering: The Engineer's Perspective
Over 80 papers dealing with various applications of knowledge-based
systems, robotics, and natural language processing will be presented.
Sriram@athena.mit.edu
------------------------------
End of AIList Digest
********************
