Copy Link
Add to Bookmark
Report
AIList Digest Volume 2 Issue 026
AIList Digest Wednesday, 7 Mar 1984 Volume 2 : Issue 26
Today's Topics:
Seminars - Extended Prolog Theorem Prover &
A Model of LISP Computation &
YOKO Random Haiku Generator &
Emulation of Human Learning &
Circuit Design by Knowledge-Directed Search &
Knowledge Structures for Automatic Programming &
Mathematical Ontology &
Problem Solving in Organizations &
Inequalities for Probablistic Knowledge
Conference - STeP-84 Call for Papers
----------------------------------------------------------------------
Date: 29 Feb 84 13:54:56 PST (Wednesday)
From: Kluger.PA@PARC-MAXC.ARPA
Reply-to: Kluger.PA@PARC-MAXC.ARPA
Subject: HP Computer Colloquium 3/8/84
[Forwarded from the SRI-AI bboard by Laws@SRI-AI.]
Mark E. Stickel
SRI International
A Prolog Technology Theorem Prover
An extension of Prolog, based on the model elimination theorem-proving
procedure, would permit production of a Prolog technology theorem prover
(PTTP). This would be a complete theorem prover for the full first-order
predicate calculus, not just Horn clauses, and provide capabilities for
full handling of logical negation and indefinite answers. It would be
capable of performing inference operations at a rate approaching that of
Prolog itself--substantially faster than conventional theorem-proving
systems.
PTTP differs from Prolog in its use of unification with the "occurs
check" for soundness, the complete model elimination input inference
procedure, and a complete staged depth-first search strategy. The use of
an input inference procedure and depth-first search minimize the
differences between this theorem-proving method and Prolog and permit the
use of highly efficient Prolog implementation techniques.
Thursday, March 8, 1984 4:00 pm
Hewlett Packard
Stanford Division
5M Conference room
1501 Page Mill Rd
Palo Alto
*** Be sure to arrive at the building's lobby on time, so that you may
be escorted to the meeting room.
------------------------------
Date: Wed 29 Feb 84 13:07:26-PST
From: MESEGUER@SRI-AI.ARPA
Subject: A Model of LISP Computation
[Forwarded from the CLSI bboard by Laws@SRI-AI.]
REWRITE RULE SEMINAR AT SRI-CSL
Wednesday March 7, 3:00 pm
A Model of Computation
Theory and application to LISP-like systems
Carolyn Talcott
Stanford University
The goal of this work is to provide a rich context in which a
variety of aspects of computation can be treated and where new
ideas about computing can be tested and developed. An important
motivation and guide has been the desire to understand the construction
and use of LISP like computation systems.
The first step was to define a model of computation and develop the
theory to provide basic tools for further work. The main components are
- basic model and notion of evaluation
- equivalence relations and extensionality
- an abstract machine as a subtheory
- formalization of the metatheory
Key features of this theory are:
- It is a construction of particular theories uniformly
from given data structures (data domain and operations).
- Focus is on control aspects of computation
- A variety of objects
Forms -- for describing control aspects of computation
Pfns -- abstraction of form in an environment
-- elements of the computation domain
-- computational analogue of partial functions
Carts -- for collecting arguments and values
Envs -- intepretation of symbols appearing in forms
cTrees -- objects describing particular computations
Applications of this theory include
- proving properties of pfns
- implementation of computation systems
- representing and mechanizing aspects of reasoning
In this talk I will describe RUM - the applicative
fragment (flavor). RUM is the most mathematically
developed aspect of the work and is the foundation
for the other aspects which include implementation
of a computation system called SEUS.
------------------------------
Date: 1 Mar 1984 10:00:33-EST
From: walter at mit-htvax
Subject: GRADUATE STUDENT LUNCH
[Forwarded from the MIT-MC bboard by SASW@MIT-MC.]
Computer Aided Conceptual Art (CACA)
Eternally Evolving Seminar Series
presents
YOKO: A Random Haiku Generator
Interns gobble oblist hash | We will be discussing YOKO and the
Cluster at operations | related issues of computer modeling
Hidden rep: convert! | of artists, modeling computer artists,
| computer artists' models, computer
Chip resolve to bits | models of artists' models of computers,
Bus cycle inference engine | artist's cognitive models of computers,
Exposing grey codes | computers' cognitive models of artists
| and models, models' models of models,
Take-grant tinker bucks | artists' models of computer artists,
Pass oblist message package | modelling of computer artists' cognitive
Federal express | models and artist's models of cognition.
Hosts: Claudia Smith and Crisse Ciro
REFRESHMENTS WILL BE SERVED
------------------------------
Date: 1 Mar 84 09:26:46 EST
From: PETTY@RUTGERS.ARPA
Subject: VanLehn Colloquium on Learning
[Forwarded from the Rutgers bboard by Laws@SRI-AI.]
SPEAKER: Dr. Kurt VanLehn
Xerox Corp.
Palo Alto Research Center
TITLE: "FELICITY CONDITIONS FOR HUMAN SKILL ACQUISITION"
A theory of how people learn certain procedural skills will be
presented. It is based on the idea that the teaching and learning
that goes on in a classroom is like an ordinary conversation. The
speaker (teacher) compresses a non-liner knowledge structure (the
target procedure) into a linear sequence of utterances (lessons). The
listener (student) constructs a knowledge structure (the learned
procedure) from the utterance sequence (lesson sequence). In recent
years, linguists have discovered that speakers unknowingly obey
certain constraints on the sequential form of their utterances.
Apparently, these tacit conventions, called felicity conditions or
conversational postulates, help listeners construct an appropriate
knowledge structure from the utterance sequence. The analogy between
conversations and classrooms suggests that there might be felicity
conditions on lesson sequences that help students learn procedures.
This research has shown that there are. For the particular kind of
skill acquisition studied here, three felicity conditions were
discovered. They are the central hypotheses in the learning theory.
The theory has been embedded in a model, a large AI program. The
model's performance has been compared to data from several thousand
students learning ordinary mathematical procedures: subtracting
multidigit numbers, adding fractions and solving simple algebraic
equations. A key criterion for the theory is that the set of
procedures that the model "learns" should exactly match the set of
procedures that students actually acquire including their "buggy"
procedures. However, much more is needed for psychological validation
of this theory, or any complex AI-based theory, than merely testing
its predictions. Part of the research has involved finding ways to
argue for the validity of the theory.
DATE: Tuesday, March 6, 1984
TIME: 11:30 a.m.
PLACE: Room 323 - Hill Center
------------------------------
Date: 1 Mar 84 09:27:06 EST
From: PETTY@RUTGERS.ARPA
Subject: Tong Colloquium on Knowledge-Directed Search
[Forwarded from the Rutgers bboard by Laws@SRI-AI.]
SPEAKER: Christopher Tong
TITLE: "CIRCUIT DESIGN AS KNOWLEDGE-DIRECTED SEARCH"
The process of circuit design is usefully viewed as search through
a large space of circuit descriptions. The search is knowledge-diverse
and knowledge-
intensive: circuits are described at many levels of abstraction (e.g.
architecture, logic, layout); designers use many kinds of knowledge and
styles of reasoning to pursue and constrain the search.
This talk presents a preliminary categorization of knowledge about
the design process and its control. We simplify the search by using a
single processor-oriented language to cover the function to structure
spectrum of circuit abstractions. We permit the circuit design and the
design problem (i.e. the associated goals) to co-evolve; nodes in the
design space contain explicit representations for goals as well as
circuits. The design space is generated by executing tasks, which
construct and refine circuit descriptions and goals (aided by libraries
of components of goals). The search is guided locally by goals and
tradeoffs; globally it is resource-limited (in design time and quality),
conflict-
driven, and knowledge-intensive (drawing on a library of strategies).
Finally, we describe an interactive knowledge-based computer
program called DONTE (Design ONTology Experiment) that is based on the
above framework. DONTE transforms architectural descriptions of a
digital system into circuit-level descriptions.
DATE: Thursday, March 8, 1984
TIME: 2:50 p.m.
PLACE: Room 705 - Hill Center
* Coffee Served at 2:30 p.m. *
------------------------------
Date: 1 Mar 84 09:27:23 EST
From: PETTY@RUTGERS.ARPA
Subject: Ferrante Colloquium on Automatic Programming
[Forwarded from the Rutgers bboard by Laws@SRI-AI.]
SPEAKER: Jeanne Ferrante
IBM Thomas J. Watson Research Center
Yortown Heights, NY
TITLE: "PROGRAMS = CONTROL + DATA
A new program representation called the program dependence graph or
PDG is presented which makes explicit both the data values on which
an operation depends (through data dependence edges) and the control
value on which the execution of the operation depends (through control
dependence edges). The data dependence relationships determine the
necessary sequencing between operations with the same control
conditions, exposing, exposing potential parallelism. In this talk we
show how the PDG can be used to solve a traditional stumbling block in
automatic program improvement. A new incremental solution to the
problem of updating data flow following changes in control flow such
as branch deletion is presented.
The PDG is the basis of current work at IBM Yorktown Heights for
compiling programs in sequential languages like FORTRAN to exploit
parallel architectures.
DATE: Friday, March 9, 1984
TIME: 2:50 p.m.
PLACE: Room 705 - Hill Center
* Coffee Served at 2:30 p.m. *
------------------------------
Date: 5 Mar 84 17:45 PST
From: Guibert.pa@PARC-MAXC.ARPA
Subject: Talk by David McAllester: Mon. Mar. 12 at 11:00 at PARC
[Forwarded from the CSLI bboard by Laws@SRI-AI.]
Title: "MATHEMATICAL ONTOLOGY"
Speaker: David McAllester (M.I.T.)
When: Monday March 12th at 11:00am
Where: Xerox PARC Twin Conference Room, Room 1500
AI techniques are often divided into "weak" and "strong" methods. A
strong method exploits the structure of some domain while a weak method
is more general and therefore has less structure to exploit. But it may
be possible to exploit UNIVERSAL structure and thus to find STRONG
GENERAL METHODS. Mathematical ontology is the study of the general
nature of mathematical objects. The goal is to uncover UNIVERSAL
RELATIONS, UNIVERSAL FUNCTIONS, and UNIVERSAL LEMMAS which can be
exploited in general inference techniques. For example there seems to
be a natural notion of isomorphism and a standard notion of essential
property which are universal (they can be meaningfully applied to ALL
mathematical objects). These universal relations are completely ignored
in current first order formulations of mathematics. A particular theory
of mathematical ontology will be discussed in which many natural
universal relations can be precisely defined. Some particular strong
general inference techniques will also be discussed.
------------------------------
Date: 5 Mar 1984 22:41 EST (Mon)
From: "Daniel S. Weld" <WELD%MIT-OZ@MIT-MC.ARPA>
Subject: AI Revolving Seminar
[Forwarded from the MIT-XX bboard by Laws@SRI-AI.]
Wednesday, March 7 4:00pm 8th floor playroom
Knowledge and Problem Solving Processes
in Organizations
Gerald Barber
Human organizations have frequently been used as models for AI systems
resulting in such theories as the scientific community metaphor, the
society of mind and contract nets among others. However these human
organizational models have been limited by the fact that do no take
into account the epistemological processes involved in organizational
problem solving. Understanding human organizations from an
epistemological perspective is becoming increasingly important as a
source of insight into intelligent activities and for computer-based
technology as it becomes more intricately involved in organizational
activities.
In my talk I will present the results of an organizational study which
attempted to identify problem solving and knowledge processing
activities in the organization. I will also outline the possibilities
for development of both human organizational models and artificial
intelligence systems in light of this organizational study. More
specifically, I will also discuss the shortcoming of organizational
theories and application of the results of this work to highly
parallel computer systems such as the APIARY.
------------------------------
Date: Tue 6 Mar 84 09:05:05-PST
From: Juanita Mullen <MULLEN@SUMEX-AIM.ARPA>
Subject: SIGLUNCH ANNOUNCEMENT -- Friday, March 9, 1984
[Forwarded from the SIGLUNCH distribution by Laws@SRI-AI.]
Friday, March 9, 1984
LOCATION: Braun Lecture Hall (smaller), ground floor of Seeley Mudd
Chemistry Building (approx. 30 yards west of Gazebo)
12:05
SPEAKER: Ben Grosof
Stanford University, HPP
TOPIC: AN INEQUALITY PARADIGM FOR PROBABILISTIC KNOWLEDGE
Issues in Reasoning with Probabilistic Statements
BACKGROUND: Reasoning with probabilistic knowledge and evidence is
a key aspect of many AI systems. MYCIN and PROSPECTOR were pioneer
efforts but were limited and unsatisfactory in several ways. Recent
methods address many problems. The Maximum Entropy principle
(sometimes called Least Information) provides a new approach to
probabilities. The Dempster-Shafer theory of evidence provides a new
approach to confirmation and disconfirmation.
THE TALK: We begin by relating probabilistic statements to logic. We
then review the motivations and shortcomings of the MYCIN and
PROSPECTOR approaches. Maximum Entropy and Dempster-Shafer are
presented, and recent work using them is surveyed. (This is your big
chance to get up to date!) We generalize both to a paradigm of
inequality constraints on probabilities. This paradigm unifies the
heretofore divergent representations of probability and evidential
confirmation in a formally satisfactory way. Least commitment is
natural. The interval representation for probabilities includes in
effect a meta-level which allows explicit treatment of ignorance and
partial information, confidence and precision, and (in)dependence
assumptions. Using bounds facilitates reasoning ABOUT probabilities
and evidence. We extend the Dempster-Shafer theory significantly and
make an argument for its potential, both representationally and
computationally. Finally we list some open problems in reasoning with
probabilities.
------------------------------
Date: Fri, 2 Mar 84 11:18 EST
From: Leslie Heeter <heeter%SCRC-VIXEN@MIT-MC.ARPA>
Subject: STeP-84
[Forwarded from the SRI-AI bboard by Laws@SRI-AI.]
In addition to the call for papers below, Eero Hyvonen
has asked me to announce that they are looking for a lecturer
for the tutorial programme. The tutorial speaker should preferably
have experience in building industrial expert systems. For a few
hours' lecture, they are prepared to pay for the trip, the stay, and
some extra.
Exhibitors and papers are naturally welcome, too.
C A L L F O R P A P E R S
STeP-84
Finnish Artificial Intelligence Symposium
(Tekoalytutkimuksen paivat)
Otaniemi, Espoo, Finland
August 20-22, 1984
Finnish Artificial Intelligence Symposium (STeP-84) will be held at
Otaniemi campus of Helsinki University of Technology. The purpose of
the symposium is to promote AI research and application in Finland.
Papers (30 min) and short communications (15 min) are invited on
(but not restricted to) the following subfields of AI:
Automaattinen ohjelmointi (Automatic Programming)
Kognitiivinen mallittaminen (Cognitive Modelling)
Asintuntijajarjestelmat (Expert Systems)
Viidennen polven tietokoneet (Fifth Generation Computers)
Teolliset sovellutukset (Industrial Applications)
Tietamyksen esittaminen (Knowledge Representation)
Oppiminen (Learning)
Lisp-jarjestelmat (Lisp Systems)
Logikkaohjelmointi (Logic Programming)
Luonnollinen kieli (Natural Language)
Hahmontunnistus (Pattern Recognition)
Suunnittelu ja etsinta (Planning and Search)
Filosofiset kysymykset (Philosophical Issues)
Robotiikka (Robotics)
Lauseen todistaminen (Theorem Proving)
Konenako (Vision)
The first day of the symposium is reserved for the Tutorial programme
on key areas of AI presented by foreign and Finnish experts. There will
be an Industrial Exhibition during the symposium. Submission deadline
for one page abstracts of papers and short communications is April 15th.
Camera ready copy of the full text is due by July 31st. The address of
the symposium is:
STeP-84
c/o Assoc. Prof. Markku Syrjanen
Helsinki University of Technology
Laboratory of Information Processing Science
Otakaari 1 A
02150 Espoo 15 Telex: +358-0-4512076
Finland Phone: 125161 HTKK SF
Local Arrangements:
Eero Hyvonen, Jouko Seppanen, and Markku Syrjanen
helsinki University of Technology
Program Committee:
Kari Eloranta Erkki Lehtinen
University of Tampere University of Jyvaskyla
Seppo Haltsonen Seppo Linnainmaa
Helsinki University of Tech. University of Helsinki
Rauno Heinonen Klaus Oesch
State Technical Research Centre Nokia Corp.
Harri Jappinen Martti Penttonen
Sitra Foundation University of Turku
Matti Karjalainen Matti Pietikainen
Helsinki University of Tech. University of Oulu
Kimmo Koskenniemi Matti Uusitalo
University of Helsinki Finnish CAD/CAM Association
Kari Koskinen
Finnish Robotics Association
Organised under the auspices of Finnish Computer Science Society.
Conference languages will be Finnish, Swedish, and English.
------------------------------
End of AIList Digest
********************
