[Prev][Next][Index][Thread]
CSLI Calendar, 6 May 1998, vol. 13:32
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
______________________________________________________________________
6 May 1998 Stanford Vol. 13, No. 32
______________________________________________________________________
A weekly publication of the
Center for the Study of Language and Information (CSLI)
Stanford University, Ventura Hall, Stanford, CA 94305-4115
____________
ACTIVITIES DURING 6 MAY TO 16 MAY 1998
WEDNESDAY, 6 MAY
4:15pm EE380: Computer Systems Laboratory Colloquium
Gates B03 (NEC Auditorium)
Academic Computer Architecture Research Refocused
New Challenges, New Technology, and Old Architectures
David Patterson
UC Berkeley
Abstract below
4:15pm Seminar on Computational Learning and Adaptation (SCLA)
Cordura 100
Learning Hierarchical-Decomposition Rules for Planning
Chandra Reddy
Computer Science, Oregon State University
Abstract below
THURSDAY, 7 MAY
12 noon CSLI CogLunch
Cordura Hall, Room 100
Compositionality and the Alleged Explanatory Role of a
Compositional Semantics
Thomas Hofweber
Stanford, Philosophy
Abstract below
4:00pm Xerox PARC Forum
George Pake Auditorium, Xerox PARC
Active Services for Programmable Media Gateways:
Taming Active Networks
Steve McCanne
UC Berkeley
Abstract below
4:15pm AI-Vision-Robotics Division Colloquium
Gates 104
Model Checking Multiagent Systems
Fausto Giunchiglia
IRST, Trento Italy
Abstract below
4:15pm Frontiers of Neuroscience Seminars
M106, Medical Center
THE SECOND ANNUAL STEPHEN W. KUFFLER LECTURE
Learning to See: The Language of Vision
Dr. Dale Purves
Duke University
Host: Dr. U. J. McMahan
Abstract below
4:15pm CS548: Distributed Systems Research Seminar
Gates B01
Designing a Research Firewall: Policy, Practice &
Experience with SURF
Jonathan Stone
Stanford University
Abstract below
7:30pm Phonology Workshop
Margaret Jacks Hall, 460:146
Estimating the Probability of Historical Connections
Between Languages Using Recurrent Sound Correspondences
Brett Kessler
Department of Linguistics, Stanford University
Abstract below
FRIDAY, 8 MAY
12 noon Logic Lunch
Room 380:383N
Contextualizing Kant's Philosophy of Mathematics
Lanier Anderson
Stanford
Abstract below
12:30pm Seminar on People, Computers, and Design
Gates B03 (NEC classroom)
Persuasive Computers: Examples, Perspectives, and
Research Directions
BJ Fogg
Sun Microsystems
http://hci.stanford.edu/html/body_cs547.html
3:15pm Philosophy Department Colloquium
Bldg. 90:92Q (Philosophy)
Platonic Souls as Persons
Anthony Long
Professor of Classics, UC Berkeley
Co-Sponsored with Classics
3:15pm Computer Science Database Seminar
Gates B-12
Web-site Management with Strudel
Alon Levy
University of Washington
Abstract below
4:15pm AI-Vision-Robotics Division Colloquium
Gates 104
Exploiting Structure in Solving Planning Problems
or
Coping with the Blooming and Buzzing World that
Surrounds Us
Tom Dean
Brown University
special colloquium
Abstract below
MONDAY, 11 MAY
12 noon Berkeley Phonology Laboratory Colloquium
46 Dwinelle Hall
Substance Abuse and Dysfunctionalism: Implications for
Optimality Theory
Charles Reiss
Department of Classics, Modern Languages and Linguistics,
Concordia University, Montreal,
http://trill.berkeley.edu/Talks/Talksch.html
3:30pm Psychology Social Lab
Jordan Hall 420:100
Title to be announced
Goeffrey Fong
Waterloo
4:10pm Berkeley Linguistics Colloquium
210 Wheeler Hall (Berkeley)
The Intonational Phonology of Korean and French
Sun-Ah Jun
UCLA
http://www.linguistics.berkeley.edu/Colloquia/
4:30pm Stanford Digital Libraries Seminar
Gates B08
Marti Hearst
Berkeley
http://diglib.stanford.edu/diglib/seminars/seminars.html
WEDNESDAY, 13 MAY
12 noon Semantics Workshop
Margaret Jacks Hall 460:146
Elizabeth Traugott and Scott Schwenter
Department of Linguistics, Stanford University
Abstract below
4:15pm Seminar on Computational Learning and Adaptation (SCLA)
Cordura 100
Making Good Inferences with Missing Information and Very
Little Computation
Dan Goldstein
Stanford University.
Abstract below
4:15pm EE380: Computer Systems Laboratory Colloquium
Gates B03 (NEC Auditorium)
The List Foundation: An Example Virtual Community
Craig Neumark and Nancy Melone
List Foundation
Abstract below
THURSDAY, 14 MAY
12 noon CSLI CogLunch
Cordura Hall, Room 100
On Actions, Causes, and Counterfactuals
Judea Pearl
UCLA, Computer Science
Abstract below
12 noon CTL: Award-Winning Teachers on Teaching
CERAS 204
Classroom Burn-Out: Experiencing It, Dealing With It, and
Learning From It
Prof. Christina Maslach
Psychology, University of California, Berkeley
http://www-ctl.stanford.edu/lectures/awtmain.html
4:00pm Xerox PARC Forum
George Pake Auditorium, Xerox PARC
Strategies for Seed Funding to a Successful Start-up
Steve Jurvetson
Managing Director, Venture Capital
Draper Fisher & Jurvetson
http://www.parc.xerox.com/ops/projects/forum
4:15pm AI-Vision-Robotics Division Colloquium
Gates 104
To be announced
Judea Pearl
UCLA, Computer Science
http://www-formal.stanford.edu/aicolloq/
4:15pm Frontiers of Neuroscience Seminars
Munzer Auditorium
Repulsive neuronal growth cone guidance
Dr. Alex Kolodkin
The Johns Hopkins University
Host: Dr. Barbara Barres
http://www.stanford.edu/dept/nbio/Spring98.html
7:30pm Linguistics Department Colloquium
Margaret Jacks Hall 460:146
Three Day Colloquium series: Day 1
Paul Smolensky
Johns Hopkins University
http://www-linguistics.stanford.edu/colloq/colloq.html
4:15pm CS548: Distributed Systems Research Seminar
Gates B01
To be announced
http://www.stanford.edu/class/cs548/
FRIDAY, 15 MAY
12:30pm Seminar on People, Computers, and Design
Gates B03 (NEC classroom)
Experience from the Design of Visio
Peter Mullen
Visio
http://hci.stanford.edu/html/body_cs547.html
3:15pm Cognitive Seminar
Jordan Hall 420:100
David Fetherstonhaugh
3:15pm Philosophy Department Colloquium
Bldg. 90:92Q (Philosophy)
Feminist Internationalism: the Role of Religion
Martha Nussbaum
Prof of Phil and Law, University of Chicago
Co-Sponsored with the Philosophy Reading Group
http://www-philosophy.stanford.edu/general/colloqs.html
3:15pm Computer Science Database Seminar
Gates B-12
Disk Array Architectures
Walt Burkhard
Computer Science and Engineering UC San Diego
http://www-db.stanford.edu/dbseminar/dbseminar.html
3:30pm Linguistics Department Colloquium
Margaret Jacks Hall 460:146
Three Day Colloquium series: Day 2
Paul Smolensky
Johns Hopkins University
http://www-linguistics.stanford.edu/colloq/colloq.html
SATURDAY, 16 MAY
10:00am Linguistics Department Colloquium
Margaret Jacks Hall 460:146
Three Day Colloquium series: Day 3
Paul Smolensky
Johns Hopkins University
http://www-linguistics.stanford.edu/colloq/colloq.html
____________
EE380: COMPUTER SYSTEMS COLLOQUIUM
on Wednesday, 6 May 1998, 4:15pm to 5:30pm
NEC Auditorium (B03), Gates Computer Science Building
http://www.stanford.edu/class/ee380/
Academic Computer Architecture Research Refocussed
New Challenges, New Technology, and Old Architectures
David A. Patterson
University of California, Berkeley
http://iram.cs.berkeley.edu/
For two decades architecture research has been focussed on desktop or
server machines. As a result of that attention, today's
microprocessors are 1000 times faster than original Berkeley RISC and
Stanford MIPS chips. Given the looming consolidation of desktop
microprocessor architectures, it may be time to declare victory and
look for new research challenges.
One candidate is personal mobile computing, where portable devices are
used for visual computing and personal communications tasks. Such a
device supports in an integrated fashion all the functionalities
provided by a portable computer, a cellular phone, a digital camera
and a video game today. In addition, we believe speech I/O will be the
cornerstone of such devices.
This new challenge brings new demands for architects. This application
cares much more about real-time performance than the performance
target of today's out-of-order microprocessors (average case
performance or SPEC performance). These programs typically operate on
vectors of 8-bit or 16-bit samples of audio and visual data and 32-bit
floating point data, not the 64-bit data of today's machines. In
addition to high performance for multimedia and DSP functions,
requirements include energy efficiency and area efficient, scalable
designs.
As a starting point, we propose reviving vector architectures. Vector
architectures match the narrower widths and real-time demands of
multimedia. They also scale well with increasing number of transistors
and wire-delay challenges of future integrated circuits. Unlike
conventional DSPs, they have a foundation of compiler research which
allows them to be programmed in high-level languages. And unlike the
MMX-style instruction set extensions, vector architectures have an
elegant and fast interface to memory and scale well with vector
length.
A vector machine benefits from a low-latency, high bandwidth memory.
Intelligent RAM, or IRAM, merges processing and memory into a single
chip to lower memory latency, increase memory bandwidth, improve
energy efficiency, and reduce size. Hence IRAM appears to be an
excellent technology for mobile computing. Surprisingly, the
integration of the processor/cache/memory of IRAM with with high-speed
serial I/O lines may also lead to very good I/O performance.
I conclude by describing the design of VIRAM-1, a microprocessor
designed by graduate students that may well have more transistors than
the contemporary Intel microprocessor. The goal is that in 2-3 years
VIRAM-1 will consume less than 2 watts of power, contain 16-32 MBytes
of memory, have about 1 GByte/sec of I/O, and crunches at the rate of
1-2 GFLOPS (64-bit floating point) and 4-16 GOPS (16-bit fixed point).
It may also challenge DSP performance even though programmed in
high-level programming languages.
The repercussions of success extend beyond the architecture research
community. Today's semiconductor industry is sharply divided into
processor and memory camps. If IRAM proves successful, unification may
come to the semiconductor industry. In such a future, its unclear
which company will ship the most processors.
Biography: David A. Patterson (University of California at Berkeley)
has taught computer architecture since joining the faculty in 1977,
and is holder of the E.H. and M.E. Pardee Chair of Computer
Science. He is a member of the National Academy of Engineering, is a
Fellow of the Computer Society of the Institute of Electrical and
Electronic Engineers (IEEE), and is also a Fellow of the the
Association for Computing Machinery (ACM). His teaching has been
honored by the ACM with the Karl V. Karlstrom Outstanding Educator
Award, by IEEE with the Undergraduate Teaching Award, by the
University of California with the Distinguished Teaching Award, and by
his department with the Diane S. McEntyre Award for Excellence in
Teaching. He received the inaugural Outstanding Alumnus Award of the
UCLA Computer Science Department as part of its 25th Anniversary and
in 1995 he received the IEEE Technical Achievement Award.
He is past chair of the CS Division in the EECS department at
Berkeley, the ACM Special Interest Group in Computer Architecture, and
the Computing Research Association. He has consulted for many
companies, including Digital, Hewlett Packard, Intel, and Sun
Microsystems, and is also co-author of five books (including two with
John Hennessy.)
At Berkeley, he led the design and implementation of RISC I, likely
the first VLSI Reduced Instruction Set Computer. This research became
the foundation of the SPARC architecture, currently used by Fujitsu,
Sun Microsystems, and Xerox. (In 1996 Microprocessor Report and COMDEX
named SPARC as one of the significant microprocessors as part of the
celebration of the 25th anniversary of the microprocessor.) He was
also a leader of the Redundant Arrays of Inexpensive Disks (RAID)
project, which led to high performance storage systems from many
companies. These projects led to three distinguished dissertation
awards from the ACM. His current research interests are in large-scale
computing using networks of workstations (NOW) and in building novel
microprocessors using Intelligent DRAM (IRAM).
____________
SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION (SCLA)
on Wednesday, 6 May 1998, 4:15pm to 5:30pm
Cordura 100
http://www-csli.stanford.edu/cll/scla.html
Learning Hierarchical-Decomposition Rules for Planning
Chandra Reddy
Department of Computer Science
Oregon State University
Corvallis, OR
Most industrial-strength planners employ hierarchical planning by
recursively decomposing tasks into subtasks. In spite of this success,
there have hardly been any learning systems applicable to hierarchical
planning. There are two opportunities for learning useful knowledge
for hierarchical planning: (1) learning task decompositions
themselves; and (2) learning control knowledge for selecting/rejecting
among various task decompositions. My work focuses on the former---in
particular, on learning goal-decomposition rules (d-rules) that
recursively decompose goals into subgoals under particular conditions.
I map the problem of learning d-rules to the problem of learning Horn
programs. Then, I describe theoretical results on learning special
cases of Horn programs. Next, I present two implementations for
learning d-rules loosely based on the theoretical results---one by
generalizing the example plans executed by the user, and the other by
solving user-given exercises with search. Finally, I demonstrate the
performance of these implementations on two planning domains.
____________
CSLI COGLUNCH
on Thursday, 7 May 1998, 12 noon
Cordura Hall, Room 100
http://www-csli.stanford.edu/csli/Coglunch/
Compositionality and the Alleged Explanatory Role
of a Compositional Semantics
Thomas Hofweber
Stanford
http://www.stanford.edu/~hofweber/
In the debate about compositionality there are two separate,
but not independent, issues under discussion, namely:
(1) Does every natural language have a compositional semantics?
(2) Does a compositional semantics play an explanatory role?
The standard answer seems to be to say "yes" to (2) and therefore
"yes" to (1). It has also been argued that the right answer is "no" to
(1) and thus "no" to (2). In this talk I will mainly look at (2). I
will argue that the answer to (2) is, most likely, "no". The main part
of the talk will be occupied with different models of what role a
semantic theory might have in an account of language understanding
(and production). Building on work by Stephen Schiffer, I will outline
a model of language understanding in which a semantic theory plays no
role, and say something about why this model is more plausible than
its alternatives. The way of looking at language understanding to be
proposed will not only have implications for the psychology of
language understanding, but also for the epistemology of language
understanding, in particular the issue of whether or not knowledge we
come to have through what other people tell us can have a priori
warrant.
Overall, I guess the best bet is to say "no" to (2), but "yes" to (1).
____________
XEROX PARC FORUM
on Thursday, 7 May 1998, 4:00pm - 5:00pm
George Pake Auditorium, Xerox
http://www.parc.xerox.com/ops/projects/forum/
Active Services for Programmable Media Gateways:
Taming Active Networks
Steven McCanne, U.C. Berkeley
Joint work with Elan Amir, U.C. Berkeley
http://www.cs.berkeley.edu/~mccanne/
Several recent proposals for an ``active networks'' architecture
advocate the placement of user-defined computation within the network
as a key mechanism to enable a wide range of new applications and
protocols, including reliable multicast transports, mechanisms to foil
denial of service attacks, intra-network real-time signal transcoding,
and so forth. This laudable goal, however, creates a number of very
difficult research problems, and although a number of pioneering
research efforts in active networks have solved some of the
preliminary small-scale problems, a large number of wide open problems
remain. In this talk, I will an alternative approach to active
networks that addresses a restricted and more tractable subset of the
active-networks design space. Our approach, which we (and others)
call ``active services'', advocates the placement of user-defined
computation within the network as with active networks, but unlike
active networks preserves all of the routing and forwarding semantics
of current Internet architecture by restricting the computation
environment to the application layer. Because active services do not
require changes to the Internet architecture, they can be deployed
incrementally in today's Internet.
I will develop and motivate our active services architecture by
describing its evolution from Amir's work on media gateways to its
present form as a novel substrate for injecting user programmable
functionality like media transcoding into the network. To motivate
this work, I will first outline the ITU model for multiparty
communication (i.e., H.323) and point out some of its constraints and
limitations. I will then show how these limitations can be overcome
with active services and present the core building blocks and
abstractions of our active service framework. I will also describe
the impressive active service prototype called "AS1" that Amir has
built on the Berkeley campus utilizing the Network of Workstations
(NOW) compute cluster, and report on its use to mitigate bandwidth
disparity among a number of real users attending a remote seminar from
last Fall (see http://bmrc.berkeley.edu/courseware/cscw/fall97/).
Finally, I will describe ongoing work in our research group that is
generalizing the active service framework for multimedia archive and
reliable-multicast proxy networks.
Biography: Steven McCanne received the B.S. degree with high honors in
Electrical Engineering and Computer Science and the Ph.D. degree in
Computer Science from the University of California, Berkeley, in 1990
and 1996 respectively. From 1988 to 1996, he was on the staff at the
Lawrence Berkeley National Laboratory where he co-developed the
network and application architecture and software that underlies the
Internet Multicast Backbone or ``MBone'' tools. In 1996, he joined the
faculty of the Department of Electrical Engineering and Computer
Science at the University of California, Berkeley, where he is
currently an Assistant Professor. His research interests include
multimedia networking, multicast communication protocols, signal
compression, remote collaboration technologies, and network simulation
architectures. His recent dissertation work, which was awarded the ACM
doctoral dissertation award for 1996-97, investigates methods for
scalable video representation and transmission over heterogeneous
multicast networks. Dr. McCanne received the Best Student Paper award
at both Winter USENIX '93 and ACM Multimedia '95, the AT&T Graduate
Scholarship, and the 1995 R&D 100 Award for the development of the
MBone tools. He is a member of the IEEE, the ACM, the AAAS, and the
Internet Engineering Task Force.
____________
AI-VISION-ROBOTICS DIVISION COLLOQUIUM
on Thursday, 7 May 1998, 4:15pm until 5:30pm
Gates 104
http://www-formal.stanford.edu/aicolloq/
Model Checking Multiagent Systems
Fausto Giunchiglia
Istituto per la Ricerca Scientifica e Tecnologica (IRST),
Istituto Trentino di Cultura (ITC),
Trento, Italy.
and
Dipartimento di Informatica e Studi Aziendali (DISA)
Universita` di Trento
Trento, Italy.
Model checking is a very successful technique which has been applied
in the design and verification of finite state concurrent reactive
processes. In this talk I will show how this technique can be lifted
to be applicable to multi-agent systems. Our approach allows us to
reuse the technology and tools developed in model checking, to design
and verify multi-agent systems in a modular and incremental way, and
also to have a very efficient model checking algorithm.
Biography: Fausto Giunchiglia is an Associate Professor at the
University of Trento, Italy. At IRST he is the coordinator of the
Reasoning Area, a group of people and projects focused on Knowledge
Representation and Reasoning, Mechanized Reasoning and Formal Methods
and Reasoning for Decision Support. His work spans the topics of
Interactive theorem proving, the GETFOL system,
Integration/interaction of mechanized reasoning systems(OMRS),
Meta-theoretic reasoning, Abstraction, Formal Methods, Contextual
reasoning, Commonsense Reasoning and the foundations/semantics of
planning. Besides being on the board of numerous conferences and
journals he is has co-won the ECAI best paper award (1984), he is the
winner of the CNR prize for outstanding research (1986), and he is the
winner of the the "Artificial Intelligence" prize of the Italian
Association for Artificial Intelligence (1992).
____________
FRONTIERS OF NEUROSCIENCE
THE SECOND STEPHEN W. KUFFLER LECTURE
Thursday, 7 May 1998, 4:15pm (time unsure)
M106, Stanford Medical Center
http://www.stanford.edu/dept/nbio/Spring98.html
Learning to See: The Language of Vision
Dale Purves, M.D.
George Barth Geller Professor for Research in Neurobiology
and
Chairman, Department of Neurobiology, Duke University
http://www.neuro.duke.edu/faculty/Purves/Purves.htm
The title of Dr. Purves' lecture will be "Learning to see: the
language of vision". As a result of Stephen Kuffler's seminal
observations on the organization of retinal ganglion cell receptive
fields in the early 1950s, much visual research has been predicated on
the idea that perception is built up from the receptive field
properties of cells at different stages in the visual pathway. Based
on the phenomenology of visual perception, Dr. Purves will suggest a
different analysis, the gist of which is that vision, much like
language, depends on learned associations.
Dr. Purves received his undergraduate degree at Yale and medical
training at Harvard, where as a student he first came to know Steve
Kuffler. After postdoctoral work in Kuffler's department, and
subsequently at University College London, he joined the faculty at
Washington University School of Medicine. He remained there until 1990
when he moved to his present position at Duke. Dr. Purves has worked
on the effects of activity on the membrane properties of muscle
fibers, synaptic specificity in autonomic ganglia, trophic
interactions in the formation and maintenance of neural connections,
the role of neuronal geometry in determining convergence and
divergence in the nervous system, the structure of individual synaptic
connections monitored over long periods, the development of the
olfactory system, and the role of neural activity in cortical
development and the allocation of cortical space. In the last few
years he has turned his attention to visual perception.
Dr. Purves was editor-in-chief of The Journal of Neuroscience from
1988 to 1993. In addition to numerous primary research articles, he
has authored five books on neurosciences. He was elected to the
National Academy of Sciences in 1989, and the Institute of Medicine in
1996.
____________
CS548: DISTRIBUTED SYSTEMS RESEARCH SEMINAR
on Thursday, 7 May 1998, 4:15pm
Gates B01
http://www.stanford.edu/class/cs548/
Designing a Research Firewall:
Policy, Practice and Experience with SURF
Jonathan Stone
Stanford University
Corporate network firewalls are well-understood and are becoming
commonplace. These firewalls establish a security perimeter that aims
to block (or heavily restrict) both incoming and outgoing network
communication. We argue that these firewalls are neither effective
nor appropriate for academic or corporate research environments
needing to maintain information security while still supporting the
free exchange of ideas.
In this paper, we present the Stanford University Research Firewall
(SURF), a network firewall design that is suitable for a research
environment. While still protecting information and computing
resources behind the firewall, this firewall is less restrictive of
outward information flow than the traditional model; can be easily
deployed; and can give internal users the illusion of unrestricted
e-mail, anonymous FTP, and WWW connectivity to the greater Internet.
Our experience demonstrates that an adequate firewall for a research
environment can be constructed for minimal cost using off-the-shelf
software and hardware components.
____________
STANFORD PHONOLOGY WORKSHOP
on Thursday, 7 May 1998, 7:30pm
Margaret Jacks Hall 460:146
http://www-linguistics.stanford.edu/Linguistics/pinterest/
Estimating the Probability of Historical Connections
Between Languages Using Recurrent Sound Correspondences
Brett Kessler
Department of Linguistics
Stanford University
http://ling-pc31.stanford.edu/~kessler/
Historical linguistics is in the unenviable position of having no
generally accepted methodology for statistically estimating whether
the connections it documents between languages actually represent
historical realities or are purely coincidental. There have been good
elucidations of the mathematical folly underlying most attempts to
demonstrate language relatedness on the basis of surface lexical
similarities (Rosenfelder, 1998), and there have been reasonably valid
attempts to put similarity-based techniques on a firm mathematical
footing (Oswalt, 1991). But few have approached the problem in a way
meaningful to historical linguists, whose comparative method teaches
that one should seek recurrent sound correspondences, not
similarities. The major exception is Ringe (1992), who proposes a full
methodology based on correspondences. However, as with most initial
releases, there seem to be several bugs in his system. In this talk I
discuss the problems and propose several fixes. In particular, I
propose that the statistical methodology be based on the use of
permutation tests. This computational approach allows us to get around
problems of paucity of data. Perhaps more importantly, it lets us
apply metrics that linguists feel are most valid (such as the number
of recurrent sound correspondences), instead of leaving us stuck with
approximations (such as association statistics like Chi Squared).
In addition to sharing interim results, I hope to get feedback about
my own bugs, and to discuss suggestions for further research. Among
such possibilities are additional metrics, including some based on
explicit word alignments and multilateral comparisons. If time
permits, I would also like to briefly discuss variations that would
address the question of quantifying how close the connection is
between historically connected languages.
____________
LOGIC LUNCH
on Friday, 8 May 1998, 12 noon
Math Corner 380:383N
http://www-logic.stanford.edu/seminars.html
Contextualizing Kant's Philosophy of Mathematics
Lanier Anderson
Stanford
The most famous claim of Kant's philosophy of mathematics is that
mathematical cognition is not analytic, but synthetic, and that this
fact is shown by the role of "construction" in mathematical argument.
Much of the recent discussion of Kant's view can be organized around
two positions on the role of this construction: one side (Hintikka,
Friedman) attributes a *logical* role to construction, and the other
side (Parsons, et al.) thinks it has a *phenomenological* role. On
the logical reading, mathematical constructions like the diagrams of
Euclidean geometry play an ineliminable role in arguments for
particular theorems, whereas on the phenomenological reading,
constructions provide a kind of direct evidence in support of the
axioms of elementary mathematics. I will discuss some recent advances
in our understanding of Kant that support a version of the logical
reading, but I will also argue that prominent recent versions of the
logical readings tend to overemphasize issues that seem prominent only
from a post 19th c. point of view, and underplay key aspects of the
role of diagrammatic constructions in the sort of arguments Euclid
himself gives. These latter considerations are closer to what Kant
himself had in mind in attributing such importance to construction.
If there is time, I will point out some of the wider epistemological
consequences of this notion of construction in Kant's theory of
cognition.
____________
SEMINAR ON PEOPLE, COMPUTER, AND DESIGN
on Friday, 8 May 1998, 12:30-2:00pm
Gates B03 (NEC Classroom)
http://www-pcd.stanford.edu/seminar/
(SITN Channel E2)
Persuasive Computers:
Examples, Perspectives, and Research Directions
BJ Fogg
Sun Microsystems, Stanford University
http://www.captology.org/
The study of computers as persuasive technologies ("captology") has
developed over the last few years at Stanford. In exploring this
domain, I've found that computers can change attitudes and behaviors
by functioning in three ways: as persuasive tools, media, and social
actors (a framework I call the "functional triad"). Various examples
of persuasive computing already exist, and many more--both good and
bad--will likely emerge as computing technology becomes increasingly
pervasive. To be sure, the Internet opens up new possibilities for
persuasive technologies, not just by providing information
persuasively but also by creating persuasive experiences. To achieve
the potentials of persuasive computing and to avoid the pitfalls, I
outline key directions for research; this includes an ongoing inquiry
into the ethics of persuasive technologies.
(Note: Some students from the captology research group will
participate in this talk by presenting their conceptual designs for
persuasive technologies.)
Biography: BJ Fogg works at Sun Microsystems, where he directs the
"Sun Summits on the Future," a series of workshops that explores the
implications of a networked work for basic human activities and
needs. In addition to his projects at Sun, last fall BJ taught a
Stanford HCI course (CS377) on persuasive computing. As a follow-up on
the course, he now directs a research group in the analysis, design,
and theory of persuasive technologies.
While completing his Ph.D. at Stanford, BJ worked at Interval Research
and at HP Labs. At both companies, he researched and designed new
types of interactive technologies.
____________
COMPUTER SCIENCE DATABASE SEMINAR
on Friday, 8 May 1998, 3:15pm - 4:30pm
Gates B12
http://www-db.stanford.edu/dbseminar/dbseminar.html
Web-site Management with Strudel
Alon Levy
University of Washington
The World-Wide Web is a prime vehicle for disseminating information.
Consequently, Web sites are growing in size, have increasingly complex
structure, and often serve information derived from multiple data
sources. Managing the content and structure of such Web sites is a
novel data management problem.
I will describe the Strudel system, which is the first system to apply
concepts from database management systems to the problem of Web site
management. In Strudel we separate three distinct tasks in building
Web sites that are usually interdependent in current Web site
management tools: (1) managing the data underlying the site, (2) the
management of the structure of the site (i.e., specifying the data
contained within each page and the links between pages), and (3)
designing the graphical presentation of pages.
The key idea in Strudel is that the structure and the content of the
Web site are specified declaratively in a high-level query language,
StruQL. As a result, it is possible to easily restructure and modify
Web sites, and to create multiple versions of a Web site from the same
underlying data. The underlying declarative representation of the Web
site also provides a platform for specifying and enforcing integrity
constraints on sites, and for designing data warehouses for their
support.
After describing Strudel and some experiences we had in using the
system, I will argue the following: (1) Web-site management is an
important application of semistructured data (2) StruQL, as a language
for querying semistructured data has several advantages over related
languages, and, most importantly, (3) Web-site management is an
important field for database research.
The Strudel system was developed jointly with Mary Fernandez (AT&T
Labs), Daniela Florescu (INRIA, France), Jaewoo Kang (Severa Inc.) and
Dan Suciu (AT&T Labs).
Biography: Alon Levy joined the faculty of the Computer Science and
Engineering Department of the University of Washington in January,
1998. Previously, he was a principal member of technical staff at
AT&T (previously, Bell) Laboratories. He received his Ph.D in Computer
Science from Stanford University in 1993. Alon's interests are in
Database Systems, Artificial Intelligence and the interactions between
the two fields. In particular, he is interested in query optimization,
data integration, management of semistructured data, Web site
management systems and knowledge representation. His most recent
projects include the Information Manifold data integration system and
the Strudel Web site management system.
____________
AI-VISION-ROBOTICS DIVISION COLLOQUIUM
special additional colloquium
on Friday, 8 May 1998, 4:15pm until 5:30pm
Gates 104
http://www-formal.stanford.edu/aicolloq/
Exploiting Structure in Solving Planning Problems
or
Coping with the Blooming and Buzzing World that Surrounds Us
Tom Dean
Computer Sciences Department
Brown University
http://www.cs.brown.edu/people/tld/
In describing the world for planning purposes, the relevant features
often multiply, seeming to indicate the need for complicated
descriptions of the underlying dynamics. For many decision problems,
however, we can structure our knowledge so as to realize compact,
factored representations of the dynamics [McCarthy and Hayes, 1969]
that take advantage of independence relationships involving the
problem features. Factored representations have long been a mainstay
of AI approaches to planning [Fikes and Nilsson, 1971]. Unfortunately,
the existence of a compact representation does not imply a simple
decision problem. In the first part of this talk, we consider methods
based on model reduction and related to the work of Hartmanis and
Stearns [1966] that attempt to simplify decision making by exploiting
structure implicit in factored representations.
If factored representations constitute the beans and chili peppers of
AI planning, then what's the big burrito, the whole enchilada, and do
you really need beans to put together a satisfying meal? In many
planning and control problems, accurate models are hard to come by. It
may not be necessary to construct a factored representation enroute to
a satisfactory solution to a planning problem. In the second part of
the talk, we re-examine our motives for studying planning and propose
an alternative perspective that combines traditional planning with
learning. The result is an AI version of adaptive control with an
emphasis on factored representations and a distinctly Bayesian flavor.
____________
STANFORD SEMANTICS WORKSHOP
on Wednesday, 13 May 1998, 2:15pm
Margaret Jacks Hall 460:146
http://www-linguistics.stanford.edu/Linguistics/semgroup/
Invoking Scalarity: The Development of IN FACT
Elizabeth Traugott and Scott Schwenter
Department of Linguistics, Stanford University
We identify and analyze the discourse contexts in which the adverbial
_in fact_ was recruited in the history of English to pragmatic marker
status. As a pragmatic marker it invokes scalarity in two domains:
that of epistemic sentence adverb (in fact2, IPAdv) , and that of
rhetorical discourse marker (in fact 3, DM). The particular analysis
of _in fact_ leads to the formulation of general hypotheses about how
scalar meanings emerge from non-scalar sources, and also about the
diachronic relationships among scalar polysemies of the same lexeme.
____________
SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION (SCLA)
on Wednesday, 13 May 1998, 4:15pm to 5:30pm
Cordura 100
http://www-csli.stanford.edu/cll/scla.html
Making Good Inferences with Missing Information
and Very Little Computation
Daniel Gray Goldstein
Engineering-Economic Systems & Operations Research
Stanford University
In contrast to those who seek more complicated ways to combine all
available information to make optimal inferences, I have investigated
what happens when one uses simple heuristics which neither use nor
combine all available information and also violate various axioms of
reasonableness, such as transitivity or the Archimedian axiom. These
heuristics, derived from human decision-making strategies, give
surprising results, and suggest some new directions one might take in
machine learning.
____________
EE380: COMPUTER SYSTEMS COLLOQUIUM
on Wednesday, 13 May 1998, 4:15pm to 5:30pm
NEC Auditorium (B03), Gates Computer Science Building
http://www.stanford.edu/class/ee380/
The List Foundation
An Example Virtual Community
Nancy Melone, Craig Newmark
The List Foundation
http://www.listfoundation.org/
The List Foundation is a Bay Area virtual community that grew from
craigs-list. It provides useful resources to the local community,
including jobs, events, apartments, for sale items, and so on.
Communication is via a family of email lists and a web site.
Over the course of its evolution, the List Foundation has built a
community where people connect with each other, and has somehow built
a feeling of community mediated primarily by electronic means. This is
recently been complemented by starting a series of "block parties."
The plans of the List Foundation include expansion to other cities and
job mentoring and training.
Biographies: Nancy was bred and born in Berkeley, and raised in the
walnut orchards of Alamo. She returned to her roots and attended Cal,
graduating with both a bachelor's and a master's degree. She also
pursued Ph.D. studies at Cal. She has held a variety of positions in
industry, all dealing with emerging technology.
She has a passion (actually she has a lot of passion) for alpine
skiing, France, and cooking - not necessarily in that order.
Craig is a hardcore Java and Web programmer who grew up wearing a
plastic pocket protector and thick black glasses, taped together, the
full nerd cliche.
He started craigslist, the predecessor to listfoundation.org, in early
'95 as a means of better connecting to people by letting them know
about cool or useful events happening around San Francisco. It rapidly
grew and built a large community of people mostly in the new media
industry.
____________
CSLI COGLUNCH
on Thursday, 14 May 1998, 12 noon
Cordura Hall, Room 100
http://www-csli.stanford.edu/csli/Coglunch/
On Actions, Causes, and Counterfactuals
Judea Pearl
UCLA
http://bayes.cs.ucla.edu/jp_home.html
Technical Reports R-218, R-248, R-236T, R-240, R-250
The talk will describe a mathematical theory of causation based on
modifiable structural equations -- a generalization of causal models
used in the physical, social, and biological sciences. In this theory,
world knowledge is represented as a collection of stable and
autonomous relationships called "mechanisms", and actions are treated
as local modifications of those relationships. The theory permits us
to: predict the ramifications of unanticipated eventualities, process
sentences in which actions appear as modalities (e.g., do(p),
"increase taxes", "make him laugh") comprehend sentences phrased
counterfactually (e.g., "B would have been different if A were true").
and compute probabilities of such sentences. Comparisons to the causal
theories of Simon (1953), Good (1961), Suppes (1970) and Lewis (1973)
will be provided, and solutions to long-standing problems in
statistics, epidemiology and economics will be demonstrated.
____________
END MATERIAL
The CSLI Calendar appears weekly on Wednesdays throughout the academic
year. Announcements, abstracts, and other information to appear in
the Calendar should be submitted to the editor, who reserves the right
to decide what does or does not go in the calendar
mailto:incalendar@csli.stanford.edu.
Requests to be added to the mailing list should be sent to
mailto:majordomo@csli.stanford.edu. With the lines in the body of
the text of either
subscribe csli-calendar
for the long form or
subscribe csli-short-calendar
for the short form. Problems with subscribing or unsubscribing should
be sent to owner-csli-calendar@csli.stanford.edu.
The full current issue is at
http://www-csli.stanford.edu/csli/Archive/calendar/current.html
and the archives at
http://www-csli.stanford.edu/csli/Archive/calendar/.
People on most of the CSLI computers can type 'help csli-calendar' to
see the current issue.
The CSLI Calendar is also posted each week to
news://nntp-csli.stanford.edu/csli.bboard.
Information about CSLI's research program is available at
http://www-csli.stanford.edu/csli/.
For maps to the Stanford University campus see
http://www.stanford.edu/home/visitors/maps.html.
____________