Copy Link
Add to Bookmark
Report
AIList Digest Volume 2 Issue 036
AIList Digest Thursday, 29 Mar 1984 Volume 2 : Issue 36
Today's Topics:
AI in Criminology - Request,
AI Reports - NASA Memoranda Request,
Expert Systems - Software Development Request,
Planning - JPL Planner Request,
AI Environments - Micro O/S Request,
Expert Systems - Explanatory Capability & Article & OPS5 Examples,
Seminars - Machine Learning & Incomplete Databases & Control of Reasoning
----------------------------------------------------------------------
Date: Mon, 26 Mar 84 21:00:38 pst
From: jacobs%ucbkim@Berkeley (Paul Jacobs)
Subject: information wanted on AI in criminology
I have been asked for information concerning AI applications in
criminology, particularly in locating and keeping track of criminals. I
am aware of various uses of computers in analyzing fingerprints and other
data; however, I have not heard of successful ``intelligent'' programs.
I'd appreciate any information on this matter.
Thanks,
--paul
------------------------------
Date: 26 Mar 84 12:40:01-PST (Mon)
From: hplabs!hao!seismo!rochester!jss @ Ucb-Vax
Subject: NASA tech. memorandum
Article-I.D.: rocheste.5853
I am trying to get:
NASA Technical Memorandum 85836, June 1983;
NASA Technical Memorandum 85838, Sept. 1983;
NASA Technical Memorandum 85839, Oct. 1983.
They are Volume I of their report entitled:
An overview of Artificial Intelligence and Robotics
"This report is part of the NBS/NASA series of overviews on AI and
Robotics." Any help in getting an official draft would be greatly
appreciated. (Copies aren't bad either.) A path to NASA, if one exists,
would also be appreciated. Thanks in advance.
Jon S. Stumpf
U. of Rochester
{allegra|decvax|seismo}!rochester!jss
[I have sent Jon a copy of the NTIS ordering info that I printed in
AIList V1 #81 back in October. This included the first of the reports
mentioned above; I am not sure about the others since the serial numbers
I have are for the NTIS version. The Gevarter overviews I have read
seem to be reasonably good summaries of the major projects in vision
and expert systems. -- KIL]
------------------------------
Date: 26 Mar 84 15:40:19-PST (Mon)
From: ihnp4!ihuxf!dunk @ Ucb-Vax
Subject: Expert Systems for Software Development?
Article-I.D.: ihuxf.2119
Anyone have references to papers describing the use of expert systems
in a software development environment (e.g. program synthesis,
programmer's consultant, debugging aid, etc.)? Thanks much.
Tom Duncan
AT&T Bell Laboratories
ihnp4!ihuxf!dunk
------------------------------
Date: Wed, 28 Mar 84 18:04:42 CDT
From: Mike Caplinger <mike@rice.ARPA>
Subject: JPL Planner
Can anybody give me any references to the Jet Propulsion Lab's
"autonomous space probe" project? This system is supposed to be able
to schedule different observations in a limited time frame (like a
planetary flyby) based on priorities and feedback from previous results.
Is it really AI or just some kind of optimization hack?
thanks,
Mike
------------------------------
Date: Wed, 28 Mar 84 12:39:59 pst
From: Peter T. Young <youngp%cod@Nosc>
Subject: 32/16-bit O/S Information Request
We would like to obtain descriptions of/sources for the following
operating systems:
RMX86
CP/M (Z80 & 8085)
CP/M-86
MS-DOS (Z-DOS)
UNIX
VMS
TOPS-20
that could be run on 32/16-bit or 32/32-bit CPU-based microcomputer
systems which are either already in production, or are scheduled for
production in the near future. Our aiming-point is a system that will
run a reasonably useful version of LISP or PROLOG in a real-time environ-
ment.
Could you provide us with some pointers for such information? Any help
you might provide would prove extremely useful. Thanks for considering
this request.
Peter T. Young
(Code 9411)
NOSC
San Diego, CA 92152
(619) 225-6686
<youngp@NOSC>
------------------------------
Date: 27 Mar 84 14:45:44 EST (Tue)
From: Dana S. Nau <dsn%umcp-cs.csnet@csnet-relay.arpa>
Subject: expert system algorithms
From: Toby Robison <eosp1!robison>
I think it is hopeless to demand that the algorithms instanced by
expert systems be well understood so they can be questioned. Even
when the algorithms can easily be printed, they will be hard for any
human being to comprehend, except in the most trivial systems. ...
I disagree. One of the reasons for separating an expert system's control
structure from the knowledge base is to allow for complex behavior with a
simple control algorithm. For example, Mycin's control structure is only
about one typewritten page [1]. Jim Reggia and I at the Univ. of Maryland
are currently working on a considerably more complex expert system control
structure, but even it is not THAT hard to understand once one understands
the preliminary mathematical background [2]. We even have a proof of
correctness for the algorithm!
REFERENCES:
[1] Davis, Buchanan, and Shortliffe. Production Rules as a Representation
for a Knowledge-Based Consultation Program. ARTIFICIAL INTELLIGENCE 8
(1977), 15-45.
[2] Reggia, Nau, and Wang. A Theory of Abductive Inference in Diagnostic
Expert Systems. TR-1338, Computer Sci. Dept., Univ. of Maryland (Dec.
1983). Submitted for Publication.
------------------------------
Date: Sun 25 Mar 84 22:40:03-PST
From: Edward Feigenbaum <FEIGENBAUM@SUMEX-AIM.ARPA>
Subject: SCIENCE, 23 Mar. 1984
[Forwarded from the Stanford bboard by Laws@SRI-AI.]
The aforementioned issue of SCIENCE has a "feature story" on knowledge-
based systems in general and expert systems in particular (p1279-1282),
with various luminaries and their luminary programs mentioned. Pretty
good article.
Ed Feigenbaum
------------------------------
Date: 23 Mar 84 19:55:52-PST (Fri)
From: pur-ee!uiucdcs!parsec!ctvax!pozsvath @ Ucb-Vax
Subject: Re: Request for OPS5 examples
Article-I.D.: uiucdcs.6361
;;; /
;;; Peter Ozsvath
;;;
;;; Factorial Program in OPS-5
;;; Simulates the stack in working memory
;;;
;;; Stage 1. Fill the working memory with (fact 1) (fact 2)...(fact n)
;;;
(p fact0 (fact {<x> = 0}) --> (remove 1) (make factorial 1 1))
(p factn (fact {<n> > 0}) --> (make fact (compute <n> - 1)))
;;; Negative number entered. Quit
(p factneg (fact {<n> < 0}) -->
(write "Good-bye."))
;;; Stage 2. Sweep out the unnecessary (fact k) and (factorial k (k-1)!)
;;; and add new (factorial (k+1) k!
;;;
(p factorial (factorial <x> <y>) (fact <x>) -->
(remove 2)
(remove 1)
(make factorial (compute (<x> + 1)) (compute (<x> * <y>))))
;;; When no more (fact k) statements are left, factorial k is fount
;;;
(p factorial2 (factorial <x> <y>) -(fact <x>) -->
(write (crlf))
(write (compute <x> - 1))
(write "!! =")
(write <y>)
(write (crlf))
(make infinifact))
;;; Called once at the beginning
;;;
(p pretty_fact (start) -->
(write "Program to demonstrate the power, compactness, and robustness")
(write (crlf))
(write "of the Winning OPS 5. This program inputs numbers whose")
(write (crlf))
(write "factorials - lo and behold - it computes RECURSIVELY")
(write (crlf))
(write (crlf))
(remove 1)
(make infinifact))
;;; Circular "loop" that reads in numbers
;;;
(p infinifact (infinifact) -->
(remove 1)
(write "Enter a positive number to compute its factorial,")
(write (crlf))
(write "or a negative one to quit")
(write (crlf))
(bind <x> (accept))
(make fact <x>))
;;; Start
(start ((start)))
This program computes factorials in OPS-5. Some things seem to be
rather difficult to do in OPS-5. This same program could be
written in several lines of lisp!
------------------------------
Date: 26 Mar 84 10:06:31 EST
From: Michael Sims <MSIMS@RUTGERS.ARPA>
Subject: KELLER SPEAKING AT ML ON WED.
[Forwarded from the Rutgers bboard by Laws@SRI-AI.]
MACHINE LEARNING BROWN BAG SEMINAR
Speaker: Richard Keller
Date: Wednesday, March 28, 1984 - 12:00-1:30
Location: Hill Center, Room 254
Placing Learning in Context:
SOURCES OF CONTEXTUAL KNOWLEDGE FOR CONCEPT LEARNING
(Alternatively titled: The Mysterious Origins of LEX's Learning Goal)
In this talk, I will describe a new source of knowledge for concept
learning: knowledge of the learning context. Most previous research
in machine learning has failed to recognize contextual knowledge as a
distinct and useful form of learning knowledge. Contextual knowledge
includes, among other things, knowledge of the purpose for learning
and knowledge of the performance task to be improved by learning. The
addition of this meta-knowledge, which describes the learning process,
provides a broader perspective on learning than has been available to
most previous learning systems.
In general, learning systems that omit contextual knowledge have an
insufficient vantagepoint from which to supervise learning activity.
Both AM [Lenat-79] and LEX [Mitchell-83], for instance, were limited
by an inability to adapt to changes in their respective learning
environments, even when the changes were a result of their own
learning behavior. This limitation is not particularly surprising;
neither of these systems contained an explicit representation of the
task they were performing (specifically, mathematical discovery and
integral calculus problem solving, respectively). Nor did these
systems contain any knowledge about the relationship between learning
and the task performance. Before it is reasonable to expect a
learning system to adapt to changes in the task environment, it is
necessary to represent task knowledge and to incorporate this
knowledge into learning procedures. My research, therefore, focuses
on the representation and use of contextual knowledge -- including
task knowledge -- as guidance for concept learning.
In this talk, I will describe a learning framework that incorporates
the use of contextual knowledge. In addition, I will introduce
various alternative methods of representing contextual knowledge, and
sketch the design of some learning algorithms that utilize contextual
knowledge. Examples will be drawn, in large part, from my work on
incorporating contextual knowledge within the LEX learning system.
------------------------------
Date: 27 Mar 84 14:22:39 EST
From: DSMITH@RUTGERS.ARPA
Subject: Rutger's Computer Science Colloquium
[Forwarded from the Rutgers bboard by Laws@SRI-AI.]
Department of Computer Science
COLLOQUIUM
SPEAKER: Dr. Witold Lipski, Jr.
Polish Academy of Sciences
TITLE: LOGICAL PROBLEMS RELATED TO INCOMPLETE INFORMATION IN DATABASES
A general methodology for modeling incomplete information in databases is
described, and then illustrated in the case of three concrete models of a
database. We emphasize the distinction between two different interpretations
of a query language -- the external interpretation, which refers queries
directly to the real world modeled by the database; and the internal
interpretation, which refers queries to the information about this world
available in the database. Our methodology stresses the need for a precise
definition of the semantics of the query language by means of a non-procedural
specification, and for a correct procedural implementation of this
specification. Various logical -- and, at times, combinatorial -- problems
connected with information incompleteness are discussed. Related work is
surveyed and an extensive bibliography is included.
DATE: Friday, March 30, 1984
TIME: 2:50 p.m.
PLACE: Room 705 - Hill Center
Coffee at 2:30
------------------------------
Date: 28 Mar 1984 09:48 EST (Wed)
From: Crisse Ciro <CRISSE%MIT-OZ@MIT-MC.ARPA>
Subject: Genesereth Talks on Control of Reasoning
[Forwarded from the MIT bboard by SASW@MIT-MC.]
Procedural Hints in the Control of Reasoning
Michael R. Genesereth
Computer Science Department
Stanford University
DATE: Thursday, March 29
TIME: 4:00 PM
PLACE: NE43 8th Floor Playroom
[This talk is also being given at IBM San Jose on
Friday, April 6, 10:00. -- KIL]
One of the key problems in automated reasoning is control of
combinatorics. Whether one works forward from given premises or
backward from desired conclusions, it is usually necessary to consider
many inference paths before one succeeds in deriving useful results.
In the absence of advance knowledge as to which path or paths are
likely to succeed, search is the only alternative.
In some situations, however, advance knowledge is available in the
form of procedural hints like those found in math texts. Such hints
differ from facts about the subject of reasoning in that they are
prescriptive rather than descriptive; they say what a reasoner OUGHT
to do rather than what is TRUE.
This talk describes a language for expressing hints to control the
process of reasoning and provides an appropriate semantic account in
the form of an interpreter that behaves in accordance with the hints.
The work is relevant to understanding the phenomenon of introspection
and is of practical value in the construction of expert systems.
HOST: Prof. Randy Davis
------------------------------
End of AIList Digest
********************
