Copy Link
Add to Bookmark
Report
AIList Digest Volume 4 Issue 141
AIList Digest Thursday, 5 Jun 1986 Volume 4 : Issue 141
Today's Topics:
Seminars - Synchronizing Plans among Intelligent Agents (SRI) &
Model-Based Reasoning with Causal Ordering (CMU) &
Tree Adjoining Grammars (UPenn) &
Connectionist Expert Systems (GTE) &
Knowledge-Based Design & Qualitative Process Theory (SU) &
FP Rewrite Rules & Parallel Unification (IBM-SJ),
Seminar Series - CSMIL (UMich),
Conference - Call for Papers for IJCAI-87
----------------------------------------------------------------------
Date: Wed 28 May 86 14:31:47-PDT
From: Amy Lansky <LANSKY@SRI-WARBUCKS.ARPA>
Subject: Seminar - Synchronizing Plans among Intelligent Agents (SRI)
SYNCHRONIZING PLANS AMONG INTELLIGENT AGENTS
VIA COMMUNICATION
Charlie Koo (KOO@SUSHI)
Stanford University
11:00 AM, MONDAY, June 2
SRI International, Building E, Room EJ228 (new conference room)
In a society where a group of agents cooperate to achieve certain
goals, the group members perform their tasks based on certain plans.
Some tasks may interact with tasks done by other agents. One way to
coordinate the tasks is to let a master planner generate a plan and
distribute tasks to individual agents accordingly. However, there are
two difficulties. Firstly, the master planner needs to know all the
expertise that each agent has. The amount of knowledge sharply
increases with the number of specialties. Secondly, the
master-planning process will be computationally more expensive than if
each agent plans for itself, since the planning space for the former
is much larger. Therefore, distributed planning is motivated.
The objective of this on-going research is to formalize a model for
synchronizing and monitoring plans independently made by nonhostile
intelligent agents via communication. The proposed model also will
provide means to monitor the progress of plan execution, to prevent
delays, and to modify plans with less effort when delays happen.
In this talk, a commitment-based communication model which allows
agents to track their commitments during execution of plans will be
proposed. It includes a language, a set of communication operators
and a set of commitment tracking operators. The process of
synchronizing plans based on this communication model will also be
described.
Relevant work: Contract Net, nonlinear planners, distributed planners.
------------------------------
Date: 28 May 1986 1217-EDT
From: Yumi Iwasaki <IWASAKI@C.CS.CMU.EDU>
Subject: Seminar - Model-Based Reasoning with Causal Ordering (CMU)
I will be presenting my thesis proposal as follows:
Date : Tuesday, June 3, 1986
Time : 2 pm
Place : WeH 5409
Title : Model-Based Reasoning of Device Behavior with Causal Ordering
Causality plays an important role in human understanding of the world. While a
number of artificial intelligence systems have been built that reason with
causal knowledge, few have addressed the issue of not only representing and
using causal knowledge but also of discovering causal relations in the domain
based on an operational definition of causality. We propose to study
discovering, representing, and using causal knowledge based on the definition
of causal relations given by the theory of causal ordering. The proposed
scheme for causal reasoning has several levels of representation of knowledge,
namely the network representation of processes, equation model of components,
and causal ordering structure. The scheme links the knowledge at the level of
intuitive understanding of processes to the diagnostic level via an
intermediate, more formal model represented as a system of equations. In this
research, we will study application of the concept of causal ordering to a task
of reasoning about physical device behavior by implementing a causal reasoning
program, ACORD, in the domain of a coal power plant. We also expect to
contribute to better understanding of advantages and disadvantages of
model-based and evidential reasoning.
------------------------------
Date: Mon, 2 Jun 86 21:53 EDT
From: Tim Finin <Tim%upenn.csnet@CSNET-RELAY.ARPA>
Subject: Seminar - Tree Adjoining Grammars (UPenn)
A STUDY OF TREE ADJOINING GRAMMARS
Vijayshanker
Ph.D. dissertation proposal
1:30pm June 9, 1986; Room 337, Towne Building
The goal of this research is to study a grammatical formalism called Tree
Adjoining Grammars (TAG's). The original motivation for TAG's was linguistic
and subsequent work established their linguistic relevance. Our study
consists of two parts. The first part deals with formal properties of TAG's:
for example, closure properties ; automaton characterizing classes of string
languages and tree languages generated by TAG's. In the second part of our
study, we outline how a syntax driven scheme for providing compositional
semantics of natural languages can be given with the Tree Adjoining Grammars.
Committee: J. H. Gallier
A. K. Joshi (Supervisor)
A. Kroch
R. Larson (MIT)
W. Rounds (U of Michigan, Ann Arbor)
B. L. Webber (Chairperson)
------------------------------
Date: Wed, 4 Jun 86 11:18:18 edt
From: Rich Sutton <rich%gte-labs.csnet@CSNET-RELAY.ARPA>
Subject: Seminar - Connectionist Expert Systems (GTE)
"Connectionist Expert Systems in a Noisy World"
by Stephen I. Gallant
This talk will describe a model for connectionist expert systems
(MACIE) and show how it is well suited to noisy and redundant
environments.
Connectionist expert systems are diagnostic expert systems
based upon a connectionist model with several interesting features:
-- They can be generated from training examples (and/or rules)
-- They perform forward chaining to make conclusions and
backward chaining to elicit additional information
-- They give IF-THEN rules to justify their inferences, even
though their knowledge base contains no such rules
-- They are arguably less prone to brittle behavior than
traditional expert systems.
In the talk it will be shown how an expert system for a noisy
and redundant problem can be constructed from: (1) a noise-free
model of an underlying process (perhaps a traditional expert
system) and (2) a model for the noise involved. System generation
is entirely automated.
Where: GTE Labs, Waltham, MA
When: June 11th, 9:30 am
Contact: Rich Sutton, rich@gte-labs.csnet, 617-466-4133 (or 466-4207)
Net address of speaker: sig@northeastern
Also: That afternoon we will have an informal research meeting of
connectionists from GTE, UMass, and Northeastern
Welcome: Visitors are welcome!
------------------------------
Date: Wed 4 Jun 86 10:58:14-PDT
From: Christine Pasley <pasley@SRI-KL>
Subject: Seminars - Knowledge-Based Design & Qualitative Process Theory (SU)
CS529 - AI In Design & Manufacturing
Instructor: Dr. J. M. Tenenbaum
Speaker: Sanjay Mittal
From: Xerox Palo Alto Research Center
Title: Pride: A Knowledge-Based Framework for Design
Guest Speaker: Kenneth Forbus
From: Qualitative Reasoning Group
University of Illinois
Title: Qualitative Process Theory: Selected Topics
Date: Wednesday, June 4, 1986
Time: 4:00 - 5:30
Place: Terman 556
Sanjay Mittal's abstract:
This talk will describe the Pride project at Xerox. The first part
of the talk will be about an expert system for the design of paper
transports inside copiers. A prototype version of the system has been in
field test for a year. It has been successfully used on real copier
projects inside Xerox - both for designing and for checking designs
produced by engineers. From an applications point of view we have been
motivated by the following observations: knowledge is often distributed
among different experts; the process of generating designs is
unnecessarily separated from their analysis, leading to long design
cycles; and design is an evolutionary process, i.e., a process of
exploration.
The second part of the talk will describe the framework in Pride for
representing design knowledge and using it to support the design
process. In this framework, the process of designing an artifact is
viewed as knowledge guided search in a multi-dimensional space of
possible designs. The dimensions of such a space are the design
parameters of the artifact. In this view, knowledge is used not only to
search the space but also to define the space. Domain knowledge is
organized in terms of design plans, which are organized around goals.
Conceptually, goals decompose a problem into sub-problems and are the
units for structuring knowledge. Design goals have design methods
associated with them, which specify alternate ways to make decisions
about the design parameters of the goal. The third major element of a
plan are constraints on the design parameters. The framework provides a
problem solver for executing these plans. The problem solver extends
dependency-directed backtracking with an advice mechanism and a context
mechanism for simultaneously maintaining multiple designs.
Kenneth Forbus' abstract:
Much of our commonsense knowledge of the physical world appears to be
organized around a notion of physical processes. Qualitative Process
theory provides a formal language for describing such processes,
including a qualitative representation of differential equations and
the conditions under which they apply. This talk will briefly review
Qualitative Process theory and discuss two topics of current research:
Interpreting measurements taken across time, and a new implementation,
based on an assumption-based truth maintenance system, that provides
roughly two orders of magnitude performance improvement.
------------------------------
Date: Wed, 04 Jun 86 17:46:49 PDT
From: Almaden Research Center Calendar <calendar@IBM.com>
Subject: Seminars - FP Rewrite Rules & Parallel Unification (IBM-SJ)
IBM Almaden Research Center
650 Harry Road
San Jose, CA 95120-6099
RESEARCH CALENDAR
June 9 - 13, 1986
GOOD REWRITE STRATEGIES FOR FP
E. Wimmers, IBM Almaden Research Center
Computer Science Seminar Wednesday, June 11 2:30 P.M. Room: B2-307
In order to implement a language based on rewrite rules, it does not
suffice to know that there are enough rules in the language; we also
need to have a good strategy for determining the order in which to
apply them. But what is good? Corresponding to each notion of having
enough rules, there is a corresponding notion of a good rewrite
strategy. We examine and characterize these notions of goodness, and
give examples of a number of natural good strategies. Although we
have confined ourselves to FP here, we believe that our techniques
(some of which are nontrivial extensions of techniques first used in
the context of lambda-calculus) will apply well beyond the realm of FP
rewriting systems.
Host: J. Backus
...
ON THE PARALLEL COMPLEXITY OF UNIFICATION OF TERMS AND RELATED PROBLEMS
C. Dwork, IBM Almaden Research Center
Comp. Sci. Colloquium Thursday, June 12 3:00 P.M. Room: Rear Audit.
Unification of terms is a well known problem with applications to a
variety of symbolic computation problems. Two terms s and t,
involving function symbols and variables, are unifiable if there is a
substitution for the variables under which s and t become
syntactically identical. For example, f(x,x) and f(g(y),g(g(c))) are
unified by substituting g(c) for y and g(g(c)) for x. As parallel
architectures become technologically feasible, researchers in logic
programming have sought parallel unification algorithms running at
speeds subpolynomial in the length of the input. Unfortunately, the
existence of such an algorithm has been shown to be "popularly
unlikely," in that it would violate commonly held beliefs about the
structure of the class P of problems solvable in polynomial time. Two
special cases of unification are term matching and equivalence
testing, in which one or both of the terms contain no variables,
respectively. In contrast to the case for general unification, term
matching and testing for equivalence can both be solved
deterministically in time O((log n)**2) for inputs of size n, using
M(n**2) processors, where M(k) is the number of sequential operations
needed to multiply k-by-k matrices (roughly k**2.5). The processor
bound can be improved to M(k) if randomization is allowed. This is
joint work with Paris Kanellakis and Larry Stockmeyer.
Host: R. Strong
...
------------------------------
Date: Sun, 1 Jun 86 20:20:38 EDT
From: Gary_M._Olson%UMich-MTS.Mailnet@MIT-MULTICS.ARPA
Subject: Seminar Series - CSMIL (UMich)
The Cognitive Science and Machine Intelligence Laboratory
(CSMIL) at the University of Michigan has been conducting a
major lecture series this spring, consisting of the
following speakers:
March 31 -- John Anderson, Carnegie-Mellon
April 21 -- Shimon Ullman, M.I.T.
May 5 -- Allen Newell, Carnegie-Mellon
May 12 -- Bobby Inman, M.C.C.
May 19 -- Roger Schank, Yale
June 24 -- Randy Davis, M.I.T.
Anyone interested in further information should contact:
Gary Olson, Director
Cognitive Science and Machine Intelligence Laboratory
University of Michigan
904 Monroe Street
Ann Arbor, Michigan 48109
313-747-4948
net address: Gary_Olson%UMich-MTS.Mailnet@MIT-Multics.Arpa
------------------------------
Date: Wed, 4 Jun 86 23:09:06 edt
From: walker@mouton.bellcore.com (Don Walker)
Subject: Conference - Call for Papers for IJCAI-87
CALL FOR PAPERS: IJCAI-87
Tenth International Joint Conference on Artificial Intelligence
August 23-28, 1987
Milan, Italy
The IJCAI conferences are the main forums for the presentation of artificial
intelligence research to an international audience. The goal of IJCAI-87 is to
promote scientific interchange, within and between all subfields of AI, among
researchers from all over the world. The conference is sponsored by the
International Joint Conferences on Artificial Intelligence, Inc. (IJCAII).
In response to the growing interest in engineering issues within the AI
community, IJCAI-87's Technical Program will have two distinct tracks: science
and engineering. The science papers, presented Sunday through Wednesday
(August 23-26), will stress the computational principles underlying cognition
and perception in man and machine. The engineering papers, presented Tuesday
through Friday (August 25-28), will highlight pragmatic issues that arise in
applying these computational principles. Tutorials will be presented on Sunday
and Monday in parallel with the first two days of the science paper
presentations. Meetings or workshops focussed on specific research issues
might most appropriately be held on Thursday or Friday.
TOPICS OF INTEREST
Authors are invited to submit papers to either the science or engineering
tracks within one of the following topic areas:
- Architectures and Languages (including logic programming, user
interface technology)
- Reasoning (including theorem proving, planning, explaining)
- Knowledge Acquisition and Learning (including knowledge-base
maintenance)
- Knowledge Representation (including task domain analysis)
- Cognitive Modeling
- Natural Language Understanding
- Perception and Signal Understanding (including speech, vision, data
interpretation)
- Robotics
REQUIREMENTS FOR SUBMISSION:
Authors are requested to prepare full papers, no more than 7 proceedings' pages
(approximately 5600 words), or short papers, no more than 3 proceedings' pages
(approximately 2400 words). The full-paper classification is intended for
well-developed ideas, with significant demonstration of validity, while the
short-paper classification is intended for descriptions of research in
progress. Authors must ensure that their papers describe original
contributions to or novel applications of AI, regardless of length
classification, and that the research is properly compared and contrasted with
relevant literature.
DETAILS OF SUBMISSION:
Authors should submit six (6) copies of their papers (hard copy only -- we
cannot accept on-line files) to the Program Chair no later than Monday, January
5, 1987. The following information must be included on the title page:
- Author's name, address, telephone number and computer mail address
(if applicable)
- Paper type (full-paper or short-paper), topic area, track (science or
engineering), and a few keywords for further classification within
the topic area
- An abstract of 100-200 words
The timetable is as follows:
- Submission deadline: 5 January 1987 (papers received after January 5
will be returned unopened)
- Notification of acceptance or rejection: 17 March 1987
- Camera ready copy due: 10 April 1987
The language of the conference is English; all papers submitted should be
written in English.
REVIEW CRITERIA:
Each paper will be reviewed by at least two experts. Acceptance will be based
on the overall merit and significance of the reported research, as well as on
the quality of the presentation. A paper may be reviewed by experts
responsible for an area or track other than the one to which it was submitted
if, in the opinion of a program committee member, it can thereby be more fairly
reviewed.
Papers submitted to the science track should make an original and significant
contribution to knowledge in the field of artificial intelligence.
Papers submitted to the engineering track should focus on pragmatic issues that
arise in reducing AI principles and techniques to practice. Such papers could
identify the critical features of some successful application system's approach
to reasoning or knowledge acquisition or natural language understanding. Of
particular interest are papers that demonstrate insightful analysis of a task
domain motivating the selection of a computational and representational
approach.
CONTACT POINTS:
Submissions and inquiries about the program should be sent to the Program
Chair:
John McDermott
Department of Computer Science
Carnegie-Mellon University
Pittsburgh, PA 15213
USA
1-412-268-2599
McDermott@cmu-cs-a.arpa
Inquiries about registration, tutorials, exhibits, and other local arrangements
should be sent to the Local Arrangements Chair:
Marco Somalvico
Dipartimento di Elettronica
Politecnico di Milano
Piazza Leonardo Da Vinci N.32
I-20133 Milano
ITALY
39-2-236-7241
somalvic!prlb2@seismo
Other inquiries should be directed to the General Chair:
Alan Bundy
Department of Artificial Intelligence
University of Edinburgh
80 South Bridge
Edinburgh EH1 1HN
UK
44-31-225-7774 ext 242
Bundy%edinburgh.ac.uk@ucl-cs.arpa
------------------------------
End of AIList Digest
********************
