LAWS@SRI-AI.ARPA (03/25/85)
From: AIList Moderator Kenneth Laws <AIList-REQUEST@SRI-AI>
AIList Digest Monday, 25 Mar 1985 Volume 3 : Issue 40
Today's Topics:
Courses - AI Short Courses,
AI Tools - LISP for Amdahl470/V6 or V8 & 4th-Generation Languages,
Planning - Real-Time Multiagent Planning,
Expert Systems - Assembly Language as Expert System,
News - Opportunities in Sweden and Cambridge,
Survey - Recent Articles,
Seminars - Marker-Passing During Problem Solving (UPenn) &
Rational Interaction: Cooperation Among Intelligent Agents (UPenn) &
Vision System of Mobile Robot (UPenn) &
Counterfactual Implication (UPenn),
Conference - Rewriting Techniques and Applications
----------------------------------------------------------------------
Date: Sat, 23 Mar 85 12:35:35 est
From: Alan Oppenheim <avo@MIT-BUGS-BUNNY.ARPA>
Subject: AI short courses
I'm trying to find a TOP QUALITY 3 or 4 day short course on AI/expert systems.
One that I've recently heard about is called knowledge-based systems and AI
offered by Integrated Computer Systems. If anyone has attended this course,
or is familiar with any other similar course I would apprciate their
comments and suggestions. Please send any replies to avo@bugs-bunny.
Al Oppenheim
------------------------------
Date: Fri, 22 Mar 85 10:50 PST
From: "B.S.Radhakrishna Sharma" <bsharma%wsu.csnet@csnet-relay.arpa>
Subject: PUBLIC DOMAIN LISP FOR AMDAHL470/V6 OR V8
CAN ANYONE SEND ME POINTERS TO ANY REASONABLY GOOD LISP AVAILABLE ON
PUBLIC DOMAIN
THANX,
BSHARMA@WSU
------------------------------
Date: Friday, 22 Mar 85 10:24:20 PST
From: tekig5!sridhar%tektronix.csnet@csnet-relay.arpa
Subject: 4th generation languages
From the discussions on the above and from the March CACM, I've become
intrigued by 4GL's and am seriously thinking of incorporating them in my
present project. If anyone could give me pointers to where I can get hold of
the technical details, I'd be very grateful. Thanks.
Sridhar
(arpa) sridhar%tekig5@tektronix
(csnet) sridhar%tekig5%tektronix@csnet-relay
(usenet) { decvax,ucbvax,ogcvax,ihnp4,allegra,purdue,psu-eea, masscomp,
mit-eddie,mit-ems,uoregon,psu-cs,,uw-beaver,ucbcad ,tekred }
!tektronix!tekig5!sridhar
USMAIL: Tektronix, Inc., MS C1-952, PO Box 3500, Vancouver, WA 98661
------------------------------
Date: 22 Mar 1985 21:53-EST
From: gasser%usc-cse.csnet@csnet-relay.arpa
Subject: Real Time Multi Agent Planning
I know of many good references on multi-agent planning and decisionmaking,
but few that I know of address the issues of real time decisionmaking.
The best overall reference I have seen is a RADC report written by people
at AI&DS: "Distributed Decision Making Environments." This is easily the
best survey of Distributed AI extant (Unfortunately I don't have the
report with me at the moment and can't give you the number. It may be for
limited distribution.) Also see the work on the Vehicle Monitoring
Testbed (Lesser, Corkhill, et al. at UMASS), the Contract Net (Randy Davis
and Reid Smith), the Rand Corp Air Traffic Control and RPV projects
(Steeb, Cammarata, etc.- there is a real time element here), the SRI work
on multiagent planning (Konolige, Appelt, Genesereth, Rosenschein). Most
of these groups publish regularly in IJCAI and AAAI Proceedings since 1979
or so. Also see the reports on the annual workshops in Distributed AI in
SIGART #'s 73, 80, and 84.
Several people have tried to integrate time into planning - notably:
Steve Vere at JPL (DEVISER) - see IEEE PAMI 5/83 pg. 246 (One of the best
actual planning systems.)
James Allen at Rochester: "Towards a General Theory of Action and Time"
AI Journal 23 (1984) pg. 123-154.
William Long, Reasoning about state from causation and time in a medical
domain, Proc AAAI 83
Drew McDermott, A temporal logic for reasoning about processes and plans,
Cog Sci, 6 1982, pg 101-157
Kahn, K. and G.A. Gorry: Mechanizing temporal knowledge, AI Journal,
9, 1977, pg 87-108
Gary Hendrix, Modelling Simultaneous actions and continuous processes
AI Journal, 8, 1977, pg 47-68.
Mark Fox's thesis and follow-on work out of CMU Robotics Institute.
We have started a distributed problem solving project here at USC Computer
Science as well, and are investigating temporal planning, joint planning
among agents, evolutionary-opportunistic reasoning, the roles of agent
models, and how agents jointly handle anomalies. Our particular focus
applies knowledge from organization theory and the sociology of work to
multi-agent problem solving.
-- Les Gasser
Asst. Professor
Dept. of Computer Science, SAL-200
USC
Los Angeles, CA. 90089-0782
ARPANET: GASSER%USC-CSE@CSNET-RELAY
------------------------------
Date: Sat, 23 Mar 85 16:14:58 pst
From: Vish Dixit <dixit%usc-cse.csnet@csnet-relay.arpa>
Subject: Assembly Language as Expert System
I agree with Curtis Goodhart's comments on Expert systems.
It raises the following questions:
1. is an Expert System simply a programming methodology?
2. must it be written in a particular language for it to
be an ES?
3. Must it embody large/imprecise body of knowledge and
employ heuristic/adhoc rules?
4. finaly, what is the most important charasteristic of an ES -
programming style, language, or the domain?
Also, there seems to be a craze for writing ES for every problem.
(You know that when even the campus recruiters start asking about them.)
If only programming methodology (Forward production System)
is the criterion Assembly language programs would qualify to be called
Expert Systems. (?)
Here it goes:
An assembly language program is written as a sequence of instructions
of the form <operation> <operand> <operand> ...
The <operands> are usualy the internal registers and some memory.
The internal registers and the memory could be considered as the
database (short term memory). Each instruction is simply a Rule of
the expert system. The Rules are <condition> <action> pairs.
In the assembly program, the conditions and some
actions are implicit to save the space.
Thus the instructions
ORG 2000H
MOV A,B
....
ORG 3000H
MOV A,B
...
JNZ 2000H
actually stand for the rules: (PC is program counter)
<PC=2000H OR PC=3000H> then <MOV A,B> <PC++>
<PC=??? AND NZ> then <PC := 2000H>
The order of these rules is immaterial.
The microprocessor control simply looks at whatever rule has its condition
satisfied, fetches it and executes it until it is explicitly put into
Halt mode or it none of rules can be applied (PC runs over). The
actions, conditions, etc are very restricted, nevertheless they are there.
signed -:)
USENET: ...!{sdcrdcf,randvax}!uscvax!dixit
CSNET: dixit@usc-cse.csnet
ARPA: dixit%usc-cse@csnet-relay.arpa
MABEL: (213) 747-3684
USMAIL: EE-systems, SAL-337, Univ. of Southern Calif.,
Los Angeles, CA 90089-0781
------------------------------
Date: 21 Mar 85 1808 PST
From: Bengt Jonsson <BXJ@SU-AI.ARPA>
Subject: Possibilities to visit Sweden
[Forwarded from the Stanford bboard by Laws@SRI-AI.]
A new research institute for Computer Sciences is now being formed in
connection with the Royal Institute of Technology, Stockholm, Sweden.
They are interested in furthering contacts with other research groups.
Areas of interest include parallel machine architectures, logic
programming, design of integrated circuits and computer networks.
There is also interest if someone likes to visit Stockholm for X
months of research. Persons interested in this can send a message to
BXJ@SAIL
------------------------------
Date: Fri 22 Mar 85 10:33:28-PST
From: Evan Cohn <COHN@SU-SCORE.ARPA>
Subject: Free Trip to Cambridge (England)
[Forwarded from the Stanford bboard by Laws@SRI-AI.]
Each year the University of Cambridge Computer Laboratory pays for a graduate
student from North America to visit the Laboratory for a few weeks over the
summer. By doing this we hope to learn about some of the latest research in
the USA .. and to make a friend. All we expect is that the selected person
will interact with relevant students and faculty members at Cambridge. We will
pay travel to and from Cambridge and provide accomodation in Wolfson College
plus reasonable living expenses (including hire of a bicycle). This year we
can only provide accomodation between 8 July and 14 September.
The selection of our summer visitor is competitive. We base our decision
mainly on the research achievements and potential of the candidate, but also
on the relevance of his interests to those of the people here.
If you would like to apply for the visiting studentship please send a one
page desciption of your current research to:
Mike Gordon
University of Cambridge Computer Laboratory
Corn Exchange Street
Cambridge
CB2 3QG
England
Please also arrange for a letter of reference on your behalf to be sent to
this address.
------------------------------
Date: Sat 23 Mar 85 14:27:10-CST
From: CL.SHANKAR@UTEXAS-20.ARPA
Subject: Plug for FREE TRIP TO CAMBRIDGE
[Forwarded from the UTexas-20 bboard by Laws@SRI-AI.]
I was one of the two people who made the trip out there last summer,
and I must say it is a great deal. Cambridge is a wonderful little
university town and the CS department is really friendly. I didn't
find out everything about the dept. but their main strengths are:
Networks, Dist. systems, Comp. Architecture, Prog. language theory
and implementation, Natural language understanding, graphics,
VLSI design automation, Program and Hardware verification, etc.
The place is infested with americans, and one of these, Larry Paulson,
will be visiting here in early April. It's really a fantastic
experience, and I'll be glad to provide more details to anyone
who is interested.
Shankar
------------------------------
Date: 23 Mar 1985 09:51-EST
From: leff%smu.csnet@csnet-relay.arpa
Subject: Recent Articles
Electronics Week, February 4 1985
Page 28-30 "Investment Shifts in AI"
Describes Corporate and venture capital funding for AI
Page 51-62
Discusses many AI efforts and systems in different areas (including many
that are not the usual examples given of AI successes). Also discusses
AI tools.
____________________________________________________________________________
High Technology, March 1985
Page 6: short article on IBM efforts in speech recognition
Page 80: shot article on Hitachi's robot hand. Based on shape-memory
alloy techniques. This hand is 1/10 the previous weight and is
described as being almost as good as the human hand.
Page 16-25: articles on expert system shells, a one page business
outlook for expert system shells, and a special article on Intellicorp.
------------------------------
Date: Sun, 24 Mar 85 19:34 EST
From: Tim Finin <Tim%upenn.csnet@csnet-relay.arpa>
Subject: seminar announcements (PENN)
Here are descriptions of some upcomming seminars at the University of
Pennsylvania. All will be given in the Moore school building (33rd and
Walnut, Philadelphia).
MARKER-PASSING DURING PROBLEM SOLVING; Jim Hendler (Brown)
3pm Tuesday March 26; 216 Moore School
A standard problem in Artificial Intelligence systems that do planning or
problem solving is called the "late-information, early-decision paradox." This
occurs when the planner makes a choice as to which action to consider, prior
to encountering information that could either identify an optimal solution or
that would present a contradiction. As the decision is made in the absence of
this information it is often the wrong one, leading to much needless
processing.
In this talk I describe how the technique known as "marker-passing" can be
used by a problem-solver. Marker-passing, which has been shown in the past to
be useful for such cognitive tasks as story comprehension and word sense
disambiguation, is a parallel, non-deductive, "spreading activation"
algorithm. By combining this technique with a planning system the paradox
described above can often be circumvented. The marker-passer can also be used
by the problem-solver during "meta-rule" invocation and for finding certain
inherent problems in plans. An implementation of such a system is discussed
as are the design "desiderata" for a marker-passer.
RATIONAL INTERACTION: COOPERATION AMONG INTELLIGENT AGENTS
Jeffrey S. Rosenschein (Stanford)
10:30 - 12:00 Thursday, April 11; 216 Moore School
The development of intelligent agents presents opportunities to exploit
intelligent cooperation. Before this can occur, however, a framework must be
built for reasoning about interactions. This talk describes such a framework,
and explores strategies of interaction among intelligent agents.
The formalism that has been developed removes some serious restrictions that
underlie previous research in distributed artificial intelligence,
particularly the assumption that the interacting agents have identical or non-
conflicting goals. The formalism allows each agent to make various
assumptions about both the goals and the rationality of other agents. In
addition, it allows the modeling of restrictions on communication and the
modeling of binding promises among agents.
VISION SYSTEM OF MOBILE ROBOT, Saburo Tsuji (Osaka University)
3pm Monday, April 1st, 216 Moore School
This paper describes model-guided monitoring of a building environment
by a mobile robot. The prior knowledge on the environment is used as a
priori world model and constraints for image analysis. The world model
is arranged in a hierarchy with three levels so as to provide coarse to
fine structures of environment; 1D route, 3D work space and 2D patterns
of specific obects. In the preliminary experiments, a mobile platform
with a TV camera is driven around passages of a building via a given
route and reports changes there to a human operator. It stops every few
meters, takes pictures and finds correspondences between line features
detected in the image and those in the image model generated from the
work space model. Mismatched lines are further examined to detect
changes in the scenes. The drawbacks in the system design are
discussed, and a brief overview of hardware and software systems of a
new mobile robot is described.
COUNTERFACTUAL IMPLICATION; Matt Ginsberg (Stanford)
2pm Friday, May 3rd, 216 Moore School
Counterfactuals are a form of commonsense non-monotonic inference that has
been of long-term interest to philosophers. In this talk, I discuss the
problem of deriving counterfactual statements from a predicate calculus
database, and present a formal description of this derivation that allows the
encoding of some context-dependent information in the choice of a sublanguage
of the logical language in which we are working. The construction is formally
identical to the "possible worlds" interpretation due to David Lewis.
A concrete example is given which uses counterfactual implication for the
purpose of diagnosing digital hardware, and the talk concludes with a
discussion of possible applications of counterfactuals elsewhere in AI.
------------------------------
Date: 15 Mar 1985 1049-PST
From: JOUANNAUD@SRI-CSL.ARPA
Subject: Conference - Rewriting Techniques and Applications
*****************************************
* First International Conference *
* on *
* Rewriting Techniques and Applications *
*****************************************
To be held
in
Dijon, Burgundy, FRANCE
from May, 20 to May, 22, 1985.
Sponsored
by
NSF, CNRS, ADI
Bull
University of Dijon
City Council of Dijon.
Program Committee
*****************
Jan Bergstra, Amsterdam
Joseph Goguen, SRI-International
John Guttag, MIT
Jean-Pierre Jouannaud, Nancy (co-chairman)
Pierre Lescanne, CRIN
David Musser, General Electric Labs (co-chairman)
Peter Padawitz, Passau
David Plaisted, Chapel Hill
Ravi Sethi, Bell Labs
David Turner, Kent
Warning: Since Dijon is a very attractive city, and welcomes many
peope late spring and summer, your Hotel reservation Form must arrive
no later than April 10th. No reservation can be garanteed after that
date. [...]
------------------------------
End of AIList Digest
********************