[Prev][Next][Index]
CSLI Calendar, 9 February 1995, vol.10:15
-
To: friends
-
Subject: CSLI Calendar, 9 February 1995, vol.10:15
-
From: Tom Burke <burke>
-
Date: Wed, 8 Feb 1995 13:03:19 -0800
C S L I C A L E N D A R O F P U B L I C E V E N T S
______________________________________________________________________________
9 February 1995 Stanford Vol. 10, No. 15
______________________________________________________________________________
A weekly publication of the Center for the Study of Language and
Information (CSLI), Stanford University, Ventura Hall, Stanford, CA 94305-4115
____________
CSLI ACTIVITIES DURING 8 -- 17 FEBRUARY 1995
WEDNESDAY, 8 FEBRUARY
4:00 - Seminar on Computational Learning and Adaptation
Cordura Hall, Room 100
Learning Bayesian Networks Using Feature Selection
Gregory M. Provan, Institute for Decision Systems Research
Abstract below
8:00pm - 1995 Wesson Lectures in Problems of Democracy, I
Building 370, Room 370
Agreements Without Theories
Cass Sunstein, Univ Chicago Law School
THURSDAY, 9 FEBRUARY
12:00 - STASS Seminar (NOTE SPECIAL TIME)
Cordura Hall, Room 100
Situations, Worlds, Concepts, and Mathematical Objects:
Some Recent Experimental Results from the Metaphysics
Research Lab
Edward N. Zalta, CSLI
Abstract below
2:15 - CSLI Seminar
Series on Visual Programming
Cordura Hall, Room 100
ToonTalk -- An Animated Programming Environment for Children
Ken Kahn, Animated Programs
Abstract below
3:15 - Philosophy Colloquium
Building 90, Room 91-A
Choice and Deliberation in Aristotle's Ethics
Heda Segvic, UC Santa Barbara
4:15 - SSP Forum
Building 60, Room 61-F
Symbolic Modeling in Engineering
John Kunz, Stanford Civil Engineering
Abstract below
7:30 - Phonology Workshop
Building 460, Seminar Room 146
Split Prominence, Stress Balance, and No Prominence
Ove Lorentz, U Tromso and UC Santa Cruz
Abstract below
8:00pm - 1995 Wesson Lectures in Problems of Democracy, II
Building 370, Room 370
Reasonable Politics
Cass Sunstein, Univ Chicago Law School
FRIDAY, 10 FEBRUARY
12:00 - Logic Lunch
Building 380, Room 383-N
Modal Quantificational Logic with Nonrigid Constants,
and the Problem of Cut-Elimination
Charles Parsons, Harvard Philosophy, visiting CASBS
Abstract below
12:30 - HCI Seminar
Terman Auditorium
Sign Language Interfaces
Nancy Frishberg, Apple Computer
Abstract below
3:15 - 1995 Wesson Lectures in Problems of Democracy
Building 90, Room 91-A
Discussion Seminar
Cass Sunstein, Univ Chicago Law School
(John Ferejohn, Stanford Political Science, Discussant)
3:30 - Linguistics Colloquium
Cordura Hall, Room 100
Negation and "Mode of Judgment"
Bill Ladusaw, UC Santa Cruz
Abstract below
THURSDAY, 16 FEBRUARY
12:00 - STASS Seminar
Cordura Hall, Room 100
An Organic Programming System for Cooperative Architecture
Hideyuki Nakashima, Electrotechnical Laboratory, Japan
Abstract below
2:15 - CSLI Seminar
Series on Visual Programming
Cordura Hall, Room 100
KidSim: End-user Programming of Simulations
Allen Cypher and David C. Smith, Apple Computer
Abstract below
4:15 - SSP Forum
Building 60, Room 61-F
Paradoxes of Rationality
Jean-Pierre Dupuy, Stanford French and Political Science,
and Ecole Polytechnique Philosophy
Abstract below
FRIDAY, 17 FEBRUARY
12:30 - HCI Seminar
Terman Auditorium
The Internet at the Turn of the Millenium
Pavel Curtis, Xerox PARC <pavel@parc.xerox.com>
Abstract below
____________
The CSLI Calendar appears on Thursday of each week throughout the academic
year. Announcements, abstracts, and other information to appear in the
Calendar on a given Thursday should be submitted by e-mail to
incalendar@csli.stanford.edu by 5:00 p.m. on the previous Tuesday.
Past issues of the CSLI Calendar, a quarterly schedule of upcoming CSLI
events, and other information about CSLI are available on the World Wide Web:
<http://csli-www.stanford.edu/>. The Calendar, with available abstracts, is
also posted each week to the csli.bboard newsgroup.
____________
SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION
on Wednesday, 8 February
4:00 p.m., Cordura Hall, Room 100
Learning Bayesian Networks Using Feature Selection
Gregory M. Provan
Institute for Decision Systems Research
<provan@camis.stanford.edu>
We introduce a novel enhancement for learning Bayesian networks with a bias
for small, high-accuracy networks, using greedy search (i.e., K2-like)
algorithms. The new approach selects a subset of features which maximize
predictive accuracy prior to the network learning phase. We examine
explicitly the effects of two aspects of the algorithm: (a) feature selection
and (b) node ordering. Our approach generates networks which are
computationally simpler to evaluate and which (for most databases tested so
far) display predictive accuracy comparable to the larger networks generated
using a K2-like approach that incorporates all features.
This talk describes work done jointly with Moninder Singh.
The goal of this seminar is to increase communication among local researchers
with interests in computational approaches to learning and adaptation. If you
would like to be added to (or removed from) the mailing list, or if you are
interested in giving a talk in the seminar, please send email to
<langley@cs.stanford.edu>.
____________
1995 WESSON LECTURES IN PROBLEMS OF DEMOCRACY
on Wednesday, 8 February
8:00 p.m., Building 370, Room 370
Agreements Without Theories
Cass Sunstein
University of Chicago Law School
Cass Sunstein is Karl H. Llewellyn Professor of Jurisprudence and Co-Director
of the Center on Constitutionalism in Eastern Europe at the University of
Chicago Law School. This is the first 1995 Wesson Lecture, to be followed by
a second lecture Thursday evening, and a discussion seminar on Friday
afternoon in the Philosophy Department.
____________
STASS SEMINAR
on Thursday, 9 February
12:00 noon, Cordura Hall, Room 100
Situations, Worlds, Concepts, and Mathematical Objects:
Some Recent Experimental Results from the
Metaphysics Research Lab
Edward N. Zalta
CSLI
<zalta@csli.stanford.edu>
After a brief review of the machinery by which we can isolate and characterize
situations and worlds, we provide confirmation of more recent predictions of
the theory of abstract objects. In particular, we isolate and characterize
Leibnizian concepts and various kinds of mathematical objects (including
"natural cardinals" and "natural sets"). We then show that these objects
behave the way we would expect them to (for example, we derive Leibniz's
"calculus" of concepts). In the seminar, we explain how our machinery detects
these abstract objects and suggest that theoretical principles embodied by the
machinery offer a unified theory of logical space.
For more information about the Metaphysics Research Lab, the URL is
<http://mally.stanford.edu/>. Zalta's home page is
<http://mally.stanford.edu/zalta.html>.
____________
CSLI SEMINAR
SERIES ON VISUAL PROGRAMMING
on Thursday, 9 February
2:15 p.m., Cordura Hall, Room 100
ToonTalk -- An Animated Programming
Environment for Children
Ken Kahn
Animated Programs
<kahn@csli.stanford.edu>
Seymour Papert once described the design of the Logo programming language as
taking the best ideas in computer science about programming language design
and "child engineering" them. Twenty-five years after Logo's birth, there has
been tremendous progress in programming language research and in
computer-human interfaces. Programming languages exist now that are very
expressive and mathematically very elegant and yet are difficult to learn and
master. We believe the time is now ripe to attempt to repeat the success of
the designers of Logo by child engineering one of these modern languages.
When Logo was first built, a critical aspect was taking the computational
constructs of the Lisp programming language and designing a child friendly
syntax for them. Lisp's "CAR" was replaced by "FIRST", "DEFUN" by "TO",
parentheses were eliminated, and so on. Today there are totally visual
languages in which programs exist as pictures and not as text. We believe
this is a step in the right direction, but even better than visual programs
are animated programs. Animation is much better suited for dealing with the
dynamics of computer programs than static icons or diagrams. While there has
been substantial progress in graphical user interfaces in the last twenty-five
years, we chose to look not primarily at the desktop metaphor for ideas but
instead at video games. Video games are typically more direct, more concrete,
and easier to learn than other software. And more fun too.
We have constructed a general-purpose concurrent programming system, ToonTalk
(TM), in which the source code is animated and the programming environment is
a video game. Every abstract computational aspect is mapped into a concrete
metaphor. For example, a computation is a city, an active object or agent is
a house, birds carry messages between houses, a method or clause is a robot
trained by the user and so on. The programmer controls a "programmer persona"
in this video world to construct, run, debug and modify programs. We believe
that ToonTalk is especially well suited for giving children the opportunity to
build real programs in a manner that is easy to learn and fun to do.
A live demo of ToonTalk will be given.
This talk is the first in a series of CSLI seminars in the next few weeks on
visual programming. Other speakers on the schedule include Allen Cypher and
David Smith (Apple Computer), Kurt Schmucker (Apple Computer), and Barbara
Hayes-Roth (Stanford Knowledge Systems Laboratory).
____________
PHILOSOPHY COLLOQUIUM
on Thursday, 9 February
3:15 p.m., Building 90, Room 91-A
Choice and Deliberation in Aristotle's Ethics
Heda Segvic
UC Santa Barbara
<hedas@humanitas.ucsb.edu>
____________
SYMBOLIC SYSTEMS FORUM
on Thursday, 9 February
4:15 p.m., Building 60, Room 61-F
Symbolic Modeling in Engineering
John Kunz
Stanford Civil Engineering
<kunz@cive.stanford.edu>
In engineering and science, models are descriptive (of the entities of a
system, their attributes and their relationships), and they are predictive (of
the values of dependent system attributes following change in independent
attribute values). Non-numeric (symbolic) models are now useful to formalize,
describe and analyze complex engineering processes, devices and systems.
Symbolic modeling systems use Artificial Intelligence techniques to represent
principled engineering knowledge.
Symbolic models of products and processes have now been implemented in support
of a number of related engineering applications. These model-based systems
are the heart of a new generation of integrated systems that enable a single
specialist or small team to develop a complete design concept. Suites of
integrated modeling and analysis tools enable "desktop engineering." Desktop
engineering will have both technical and social similarities with the current
practice of desktop publishing.
With case studies and theoretical background, this talk will discuss both the
current practice of symbolic modeling in engineering and its emerging practice
in desktop engineering.
JOHN KUNZ is Senior Research Associate in the Center for Integrated Facility
Engineering, an industrially-sponsored center of the Departments of Civil
Engineering and Computer Science. He first taught symbolic modeling at
Stanford in 1989. He has extensive experience in developing non-numeric
symbolic models in engineering, manufacturing, and medicine.
____________
PHONOLOGY WORKSHOP
on Thursday, 9 February
7:30 p.m., Building 460, Seminar Room 146
Split Prominence, Stress Balance, and No Prominence
Ove Lorentz
U Tromso and UC Santa Cruz
<lorentz@ling.ucsc.edu>
Typically, stressed syllables are characterized by a convergence of prominence
features. In this talk, I will look at cases where such prominence features
are either separated or not present.
One kind of split prominence occurs when pitch prominence, alias sentence
accent, does not associate with a syllable bearing main stress, but either
immediately follows it, goes to a secondary stress, or goes to an edge. This
is found in Mainland Scandinavian when the syllable bearing the main stress
already has a lexical accent assigned to it.
Most dialects of Mainland Scandinavian require that a syllable bearing stress
is bimoraic, but in some dialects a foot may consist of two light syllables,
which are said to have Stress Balance or equilibrium, since it is difficult to
hear which of the two syllables bears the stress. A split prominence analysis
for these will be compared to the bisyllabic head analysis of Riad (1992).
A language which appears to have split prominence on the sentential level, is
Indonesian Malay, which has been the focus of several recent papers.
Researchers and native speakers alike seem to disagree on which of the two
syllables of the final foot bears the stress. The separation of tone from
rhythmic head will be proposed to account for this disagreement, and for the
case where Cohn & McCarthy (1994) propose an iamb in the middle of trochees in
Indonesian.
Indonesian also has the phenomenon of "no prominence", namely the vowel schwa,
which does not carry stress in Cohn & McCarthy's dialect, even though it is
targeted for stress by rhythm rules. I will look at some dialects where schwa
may carry stress, and argue for a more differentiated weight hierarchy for
Indonesian.
If time permits, I will also look at the well-known case of reduced vowels in
Eastern Mari (Cheremis), discussed, e.g., in Kenstowicz 1994, which also do
not carry stress except if there are no full vowels in the word. Other facts
of Mari throw some light on the nature of these vowels and on what is required
of a stress host.
The main conclusion of the discussion will be that a stress rhythm may be
built over a varied assortment of prominence properties. When such properties
are not present, an already established rhythm may be imposed on unstressable
syllables.
____________
1995 WESSON LECTURES IN PROBLEMS OF DEMOCRACY
on Thursday, 9 February
8:00 p.m., Building 370, Room 370
Reasonable Politics
Cass Sunstein
University of Chicago Law School
Cass Sunstein is Karl H. Llewellyn Professor of Jurisprudence and Co-Director
of the Center on Constitutionalism in Eastern Europe at the University of
Chicago Law School. This is the second 1995 Wesson Lecture. A discussion
seminar on Friday afternoon in the Philosophy Department is also scheduled as
part of Wesson program.
____________
LOGIC LUNCH
on Friday, 10 February
12:00 noon, Building 380, Room 383-N
Modal Quantificational Logic with Nonrigid Constants,
and the Problem of Cut-Elimination
Charles Parsons
Harvard Philosophy and CASBS Stanford
The addition of modalities to first-order quantificational logic necessitates
choices that are largely independent of the choice of the modal propositional
logic. Customarily, in possible-worlds semantics objects are assigned to
variables rigidly (i.e., independently of the possible world). One can still
have individual constants that are not rigid designators. Then the
characteristic problems discussed under the heading of "referential opacity"
can be formulated. In spite of this, although logics of this kind have
existed since the 1960's they are not as well known as they might be.
We will discuss the issues involved in formulating these logics, attending to
the difference between ordinary ("full") quantificational logic and free
logic, which allows terms that fail to designate existing objects, following
the approach to the free-logical case briefly outlined in the author's
_Mathematics in Philosophy_ (Cornell, 1983), pp. 331-333.
With the propositional logics K, T and S4, a cut-free formulation is
relatively straightforward in the absence of identity (see chapter 2 of
Schuette, _Vollstaendige Systeme modaler und intuitionistischer Logik_
(Springer, 1968)). We will report on the state of our work on the question
whether the obstacles posed by identity can be overcome, and greater
generality with respect to propositional logic obtained, using the methods of
Mints's talk earlier this year.
____________
SEMINAR ON HUMAN-COMPUTER INTERACTION
on Friday, 10 February
12:30 p.m., Terman Auditorium
Sign Language Interfaces
Nancy Frishberg
Apple Computer
<nancyf@seiden.com>
with a guest appearance by RALPH,
shown by David L. Jaffe, Palo Alto VA Hospital
<jaffe@roses.stanford.edu>
Surveying current work in computer-based methods for teaching, learning,
storing, retrieving and manipulating sign language forms offers a new window
on questions of human-computer interaction, natural language processing, and
gesture generation and recognition. The sign languages referred to are fully
formed natural languages linked to cultural values and social behaviors in
deaf communities.
Collaborations between technology specialists and sign language specialists
could result in truly useful and usable applications. Understanding the
expected audiences and settings of use as well as choices in technology are
crucial to interface design. Are we designing for hearing literate adults
learning a second language at home or deaf children with limited experience in
any language using equipment at school? Do we need random access to lots of
full screen video or will a smaller video display mixed with live camera work
better? How does the content developer's interface differ from the ultimate
user's interface? One positive development is number of sign language
interface projects which include deaf people as part of the design and
implementation teams. Equally exciting is the fact that projects are no
longer only experimental, but are reaching maturity and coming into the
marketplace.
Projects in notation and transcription systems, first and second language
acquisition, dictionary making, non-text representations and displays from the
U.S. and nearly a dozen other countries will be described, demonstrated or
sampled on videotape. In addition, we will consider some experiments in
automatic recognition of sign languages and electronic transmission of
signing. And, we'll enjoy a live demo of RALPH, a computer-controlled
electromechanical hand that produces fingerspelling for people who are
deaf-blind.
NANCY FRISHBERG has written and spoken extensively on sign language,
interpreting, language teaching and evaluation for the past twenty-five years.
Frishberg holds a Ph.D. in linguistics from UCSD, is a certified sign language
interpreter and instructor. Her book, "Interpreting: An Introduction", is the
primary text for sign language interpreter instruction in the U.S.
A former academic, she has been involved in technology work for the past
decade. Her work on multimedia interfaces and usability from IBM's Watson
Research Center has been presented at CHI and won an IICS award. She is
currently with Apple's Newton Licensing Programs. She also serves on the
board of the Association for Software Design.
RALPH will be shown by DAVID L. JAFFE of the VA Medical Center, Palo Alto.
____________
1995 WESSON LECTURES IN PROBLEMS OF DEMOCRACY
on Friday, 10 February
3:15 p.m., Building 90, Room 91-A
Discussion Seminar
Cass Sunstein
University of Chicago Law School
with John Ferejohn, Discussant
Stanford Political Science
Cass Sunstein is Karl H. Llewellyn Professor of Jurisprudence and Co-Director
of the Center on Constitutionalism in Eastern Europe at the University of
Chicago Law School.
____________
LINGUISTICS COLLOQUIUM
on Friday, 10 February
3:30 p.m., Cordura Hall, Room 100
Negation and "Mode of Judgment"
Bill Ladusaw
UC Santa Cruz
<ladusaw@ling.ucsc.edu>
In this paper I advocate the adoption of what could be termed a "semantically
exocentric" interpretation of natural language (descriptive) negation. This
entails adopting what Horn terms a "symmetrist" position with respect to
negation and affirmation as modes of predication (or "judgment"). I justify
the adoption of the well-known typological generalization that the morphology
of clausal negation does not routinely iterate and by its utility in
accounting for why there is no "semantic pressure" against negative concord
interpretations of clauses with multiple occurrences of negative morphology.
The postulation of a class of modes of judgment, which map the basis for a
proposition into a propositional meaning focuses our attention on the question
of what the possible bases for propositions are. In the spirit of a
"structured meaning" approach, I propose that the distinction between unary
bases consisting of Davidsonian event descriptions and binary bases consisting
of an individual and a (possibly derived) property corresponds to the
traditional distinction between thetic and categorical "judgments." This
views the etiology of the contrast (and related "Milsark" effects) as a matter
of semantics rather than as only an "information packaging" issue.
____________
STASS SEMINAR
on Thursday, 16 February
12:00 noon, Cordura Hall, Room 100
An Organic Programming System for Cooperative Architecture
Hideyuki Nakashima
Cooperative Architecture Project Team
Electrotechnical Laboratory, Japan
<nakashim@etl.go.jp>
At ETL, we are developing a new software methodology for building large,
complicated systems out of simple units in the New Models for Software
Architecture project to which CSLI also joins. The emphasis is on the
architecture (called cooperative architecture) which is used to combine the
units, rather than on the intelligence of individual units.
We named the new programming methodology "organic programming" after the
flexibility of organic systems such as plants and animals. The design takes
the concept of situated reasoning as its base, and thus an extension of
Prosit. Organic programming is also a natural extension of object oriented
programming.
I will also describe the application of an organic programming language Gaea
to the description of multi-agent system. One of the advantages resides in
that we can program the system in a subsumptive manner.
____________
CSLI SEMINAR
SERIES ON VISUAL PROGRAMMING
on Thursday, 16 February
2:15 p.m., Cordura Hall, Room 100
KidSim: End-user Programming of Simulations
Allen Cypher and David C. Smith
Apple Computer
<cypher@apple.com,david_c._smith@powertalk.apple.com>
KidSim is an environment that allows children to create their own simulations.
They create their own characters, and create rules that specify how the
characters are to behave and interact. KidSim is programmed by demonstration,
so that users do not need to learn a conventional programming language or
scripting language. Informal user studies have shown that children are able
to create simulations in KidSim with a minimum of instruction, and that KidSim
stimulates their imagination.
Allen Cypher is a Research Scientist at Apple Computer, Inc. His main
interest is in end-user programming -- giving all computer users capabilities
that have traditionally belonged to programmers. He is a co-creator of
KidSim. Allen also created the Eager system, which observes a users' actions
and creates programs to automate repetitive activities. He edited the book
"Watch What I Do: Programming by Demonstration", published in 1993 by MIT
Press. He received a Ph.D. from Yale University in 1980.
While on the Xerox "Star" project, David Canfield Smith invented icons and the
desktop metaphor for computer interfaces, today used by 100 million people.
For the past several years at Apple, Dave has worked on educational software,
particularly on finding a way for children to program computers. KidSim(tm)
is the culmination of that work. Dave's research interests include
human-computer interaction, educational software, programming language design,
programming environments, end-user programming, and getting rid of the
"priests" of computing. The unifying theme behind his work for the past
twenty years has been the attempt to make computers more accessible to
ordinary people.
____________
SYMBOLIC SYSTEMS FORUM
on Thursday, 16 February
4:15 p.m., Building 60, Room 61-F
Paradoxes of Rationality
Jean-Pierre Dupuy
Stanford French and Political Science
and Ecole Polytechnique Philosophy
<jpdupuy@leland.stanford.edu>
Rational choice theory and game theory are now undergoing a foundational
crisis. This is openly acknowledged by some of their most eminent
representatives. The class of paradoxes we will discuss constitutes one of
the most serious symptoms of this. These are known as paradoxes of backwards
induction. For most decision problems with a finite horizon, working
backwards in time, step by step, allows in principle to reach a complete
solution. Since the invention of dynamic programming, this method has become
the rational way for dealing with this sort of problem. Yet in the last
decade its very foundations have come to appear less solid than once thought.
My claim is that there is no other foundation for the pretended obviousness of
backwards induction than a metaphysical principle according to which, as far
as rationality is concerned, only the future matters. I have built a general
framework to contest the universality of this principle and I will show that
it is sufficient to resolve the difficulties posed by the method in question.
More precisely, I will establish that the paradoxes of backwards induction are
Newcomb problems.
It is widely thought that Newcomb problems are fantasies of philosophers or
theologians. The various forms that the paradoxes of backwards induction take
are quite another story: the possibility of reciprocal exchange; the stability
of agreements, promises and contracts; the effectiveness of threats and
deterrence are only a few examples, and it must be admitted there would be no
viable human society if people had not succeeded in ensuring the stability of
these raw materials of social relations.
Here I will limit myself to two cases: the credibility of promises and the
effectiveness of deterrence. Most authors consider the latter to be the
inverse symmetrical image of the former. The solution I propose reveals that
the problem of deterrence is appreciably more complex than that of promising.
JEAN-PIERRE DUPUY is professor of social and political philosophy at Ecole
Polytechnique, Paris, and is director of research at the CNRS (Philosophy).
He is founding director of CREA (Centre de Recherche en Epistemologie
Appliquee), the philosophical research group of the Ecole Polytechnique. He
is also full professor part of the year in the Departments of French and by
courtesy Political Science at Stanford.
____________
SEMINAR ON HUMAN-COMPUTER INTERACTION
on Friday, 17 February
12:30 p.m., Terman Auditorium
The Internet at the Turn of the Millenium
Pavel Curtis
Xerox PARC
<pavel@parc.xerox.com>
We keep hearing about the net and the millions of new people using it every
day, so why don't I ever run into any of them there? Current visions of the
"National Information Infrastructure" make it sound like a big, vacant
reference library. Instead, it should be more like a town, with many people
interacting with each other as they go about their activities.
I'll talk about our research into the design and construction of "network
places," including a discussion of Jupiter, our current system based around
multimedia extensions to a MUD, and the Fabric, our vision for the next
system, intended to scale to the size of the full Internet.
DR. PAVEL CURTIS has been a member of the research community at the Xerox Palo
Alto Research Center since 1983, during which time he has worked on aspects of
the Smalltalk-80, Interlisp-D/Xerox Lisp, and Cedar programming environments
and on other projects mostly related to the design and implementation of
programming languages, including leadership of the SchemeXerox project
exploring large-scale software development in the Scheme programming language.
He is the founder and chief administrator of LambdaMOO, one of the most
popular recreational social virtual realities on the Internet. His current
work centers on the Social Virtual Reality project, investigating the
implementation, applications, and implications of systems that allow multiple
simultaneous users to communicate and interact in pseudo-physical
surroundings.
____________