[Date Prev][Date Next][Date Index]
CSLI Calendar, Wednesday, 4 December 2002, vol. 18:13
CSLI CALENDAR OF PUBLIC EVENTS
______________________________________________________________________
4 December 2002 Stanford Vol. 18, No. 13
______________________________________________________________________
A weekly publication of the
Center for the Study of Language and Information (CSLI)
Stanford University, Ventura Hall, Stanford, CA 94305-4115
http://www-csli.stanford.edu/
____________
ACTIVITIES FROM 4 DECEMBER 2002 TO 13 DECEMBER 2002
WEDNESDAY, 4 DECEMBER 2002
12 noon Psychology Developmental Brownbags
Jordan Hall 420:100
One mouse, two mouses, three mice: How children might recover
from speech errors without parental feedback, innate rules or
any other explicit interventions (including those royal and divine)
Michael Ramscar
Psychology, Stanford University
http://www-psych.stanford.edu/news.html#dev_brownbag
3:00pm Reuters Foundation Digital Vision Program Seminar
Wallenberg Learning Theater, Main Quad
HP's e-Inclusion Strategy
Janiece Evans-Page and Scott Bossinger
Hewlett-Packard, Palo Alto
http://reuters.stanford.edu/seminar_speakers.html
Information below
3:00pm Computer Science Talk
Gates 104
Proving Theorems about Java and the JVM
J Strother Moore
Computer Sciences, University of Texas at Austin
Abstract below
4:00pm Fourth annual CS 248 3D video game competition
Graphics teaching labs, basement level, Sweet Hall
Information below
4:15pm EE380: Computer Systems Lab Colloquium
Gates B03 (NEC Auditorium)
Mozilla and Open Source, Our Project Dynamics and Exploitable
Technologies
Scott Collins
Mozilla
http://ee380.stanford.edu/contents.html
Abstract below
THURSDAY, 5 DECEMBER 2002
12:15pm CSLI CogLunch
Cordura Hall, Room 100
Perceiving, Comprehending and Measuring Design Activity through the
Questions Asked while Designing
Ozgur Eris
Center for Design Research, Stanford
http://www-csli.stanford.edu/events/Coglunch/
Abstract below
4:00pm Personality Seminar
Jordan Hall 420:100
Self defining memories in late adolescence
Avril Thorne
UC Santa Cruz
http://www-psych.stanford.edu/news.html#person_lab
4:00pm PARC Forum
George Pake Auditorium at PARC
Grapes, Wine & the Hilgard Project
Roger Boulton
Scott Professor of Enology and Chemical Engineering, UC Davis
http://www.parc.com/forum/
4:00pm UC Berkeley CIS Seminar
Soda Hall 310 (UC Berkeley)
Computer vision and control for soccer playing robots
Raul Rojas
Freie Universitaet Berlin
http://www.cs.berkeley.edu/~eyal/cis-seminar
Abstract below
4:00pm UC Berkeley Electronic Systems Design Seminar
DOP Center Classroom, 540AB Cory Hall (UC Berkeley)
The Design of a Formal Property-Specification Language
Moshe Y. Vardi
Rice University
http://www-cad.eecs.berkeley.edu/esd-seminar
Abstract below
4:15pm SSP: Symbolic Systems Forum
Bldg. 380:380C (Math Corner)
Linguistically Rich Statistical Models of Language
Joseph Smarr
Symbolic Systems Program
http://symsys.stanford.edu:8081/ssp-dynamic/servlet/ssp_events
Abstract below
4:15pm Networking Lecture
Packard 101
OceanStore: Toward Global Scale, Self-Repairing, Secure,
and Persistent Storage
John Kubiatowicz
Computer Science Division, UC Berkeley
Abstract Below
4:15pm Fundamental Themes in Neuroscience Seminar
Munzer Auditorium, Beckman B060
Control of Synapse Number and Strength in Developing Cortical Networks
Gina Turrigiano
Brandeis University
http://www-med.stanford.edu/sbrc/calendar/
5:30pm Stanford Phonology Workshop
Margaret Jacks Hall 460:126
A discourse-phonology interface in ASL: Pronouncing two words at once
Mary Rose
Linguistics, Stanford University
http://www-linguistics.stanford.edu/pinterest/
Abstract below
FRIDAY, 6 DECEMBER 2002
11:00am Berkeley Institute of Cognitive and Brain Seminar
Tolman 5101 (Berkeley)
Vision Concepts in Search of Improved Transportation Safety:
How the Warning Signal Tortoise Beats the Hare, and How the
Elephant Does Too, but We Can't Tell
Ted Cohn
Dept. of Bioengineering and School of Optometry, UC Berkeley
http://socrates.berkeley.edu:4247/seminar.html
12:30pm CS547: Human-Computer Interaction Seminar
Gates B01
Electronic Chalk: A tool for live and remote teaching
Raul Rojas
Freie Universitaet Berlin
http://www-pcd.stanford.edu/cs547/
Abstract below
2:15pm NLP Reading Group
Margaret Jacks Hall, 460:301
Natural Language Interaction with Intelligent Tutoring Systems
Brady Clark and Karl Schultz
Stanford University
http://www-nlp.stanford.edu/nlp/nlpgroup.html
Abstract below
3:15pm Philosophy Department Colloquium
Bldg. 60:61H (note change in place)
Coming to our Senses: Could Concept Empiricism Be True?
Jesse Prinz
University of North Carolina, Chapel Hill
http://www-philosophy.stanford.edu/ce.html
3:15pm Friday Cognitive Seminar
Jordan Hall 420:100
Title to be announced
Elisabetta Zibetti
http://www-psych.stanford.edu/news.html#frisem
3:30pm Linguistics Department Colloquium
Margaret Jacks Hall 460:126
Evolutionary Optimality Theory
Gerhard Jaeger
Universitaat Potsdam
http://www-linguistics.stanford.edu/colloq/
Abstract below
MONDAY, 9 DECEMBER 2002
2:00pm Special University Oral Examination
Gates 260
Efficient Algorithms for Option Pricing
Donald Aingworth
Computer Science, Stanford University
URL
Abstract below
4:00pm Berkeley Linguistics Department Colloquium
182 Dwinelle (Berkeley)
Coding Strategies in Comparative Bantu and Implications for
Diachronic Syntax
Yukiko Morimoto
University of Duesseldorf
http://linguistics.berkeley.edu/lingdept/Current/events.html
Abstract below
TUESDAY, 10 DECEMBER 2002
3:00pm BISC Seminar
405 Soda Hall, Berkeley
SEAL Soft, Elementary And Logical approach for pattern recognition
Bojan Novak
University of Maribor, Slovenija
http://buffy.eecs.berkeley.edu/Seminars/
4:15pm SNRC Industry Seminar
Packard 101
Approaches to 4G Access Network Architectures
James Kempf
DoCoMo Labs - USA
http://snrc.stanford.edu/events/industry-seminar/
7:00pm Emerging Technology Group
Cubberley Community Center, A-3, 4000 Middlefield, Palo Alto
Creating Applications with Mozilla
Ian Oeschger
Netscape
WEDNESDAY, 11 DECEMBER 2002
3:45pm Psychology Department Colloquium
Jordan Hall 420:041
Culture and Point of View
Richard Nisbett
University of Michigan
http://www-psych.stanford.edu/news.html
THURSDAY, 12 DECEMBER 2002
4:00pm PARC Forum
George Pake Auditorium at PARC
Nexus Analysis: Expanding the Circumference of Discourse Analysis
Ron Scollon and Suzanne Scollon
Linguistics, Georgetown University
http://www.parc.com/forum/
4:15pm CSLI Seminar on Computational Learning and Adaptation
Cordura Hall, room 100
Hierarchies of Models: Toward Understanding Planetary Nebulae
Kevin Knuth
Computational Sciences Department, Code IC NASA Ames Research Center
http://cll.stanford.edu/scla.html
Abstract below
4:15pm Fundamental Themes in Neuroscience Seminar
Munzer Auditorium, Beckman B060
Synaptic Mechanisms for Cortical Map Plasticity in Rat Somatosensory
Cortex
Daniel Feldman
UC San Diego
http://www-med.stanford.edu/sbrc/calendar/
____________
Stanford Blood Bank status: Critical shortage of O- and a shortage
of O+ and A-. For an appointment: http://bloodcenter.stanford.edu/
or call 650-723-7831. It only takes an hour of your time.
____________
REUTERS FOUNDATION DIGITAL VISION PROGRAM SEMINAR
on Wednesday, 4 December 2002, 3:00pm - 4:30pm
Wallenberg Learning Theater, Wallenberg Hall, Main Quad.
http://reuters.stanford.edu/seminar_speakers.html
HP's e-Inclusion Strategy
Janiece Evans-Page and Scott Bossinger
Hewlett-Packard, Palo Alto
HP is committed to making the social and economic benefits of the
digital age accessible to all people through our e-Inclusion strategy.
The majority of our strategic grants initiatives are aligned with our
e-Inclusion strategy. Through these initiatives, HP is providing the
technology resources, tools, and solutions to creatively address
important issues in underserved communities.
____________
COMPUTER SCIENCE TALK
on Wednesday, 4 December 2002, 3:00pm - 4:30pm
Gates 104
Proving Theorems about Java and the JVM
J Strother Moore
Computer Sciences, University of Texas at Austin
I will explain how the latest version of the Boyer-Moore theorem
prover, ACL2, is used to prove theorems about Java methods. ACL2 may
be thought of as a theorem prover for functional Common Lisp. We
model the Java Virtual Machine (JVM) operationally, e.g., as a Lisp
function. We prove theorems about Java methods by compiling the
methods with javac and proving theorems about the execution of the
resulting bytecode by the JVM model. I will discuss the general
approach and show some examples. The examples will deal with Java's
int arithmetic, simple control, including recursion, object creation
and modification in the heap, thread creation, and mutual exclusion
via monitors.
About the speaker: J Strother Moore holds the Admiral B.R. Inman
Centennial Chair in Computing Theory at the University of Texas at
Austin. He is also chair of the department. He is the author of many
books and papers on automated theorem proving and mechanical
verification of computing systems. Along with Boyer he is a co-author
of the Boyer-Moore theorem prover and the Boyer-Moore fast string
searching algorithm. With Matt Kaufmann he is the co-author of the
ACL2 theorem prover. Moore got his PhD from the University of
Edinburgh in 1973 and his BS from MIT in 1970. Moore was a founder of
Computational Logic, Inc., and served as its chief scientist for ten
years. He and Bob Boyer were awarded the 1991 Current Prize in
Automatic Theorem Proving by the American Mathematical Society. In
1999 they were awarded the Herbrand Award for their work in automatic
theorem proving. Moore is a Fellow of the American Association for
Artificial Intelligence.
____________
FOURTH ANNUAL CS 248 3D VIDEO GAME COMPETITION
on Wednesday, 4 December 2002, 4:00pm-6:00pm
Graphics teaching labs, basement level, Sweet Hall
At 4:00pm on Wednesday, December 4, a judging will be held to select
the best 3D video game produced by a current student (or team of
students) in CS 248 - Introduction to Computer Graphics.
The jury will consist of:
BRENT IVERSON, Chief Technical Office (CTO) of Electronic Arts's
Redwood Shores studio. Other companies Brent has been affiliated with
include Jane's Combat Simulations and Sonalysts, Inc. Game credits at
Electronic Arts include Rumble Racing, Sub Command, World War II
Fighters, U.S. Navy Fighters, Chuck Yeager's Air Combat, and LHX:
Attack Chopper.
LANG BEECK, software design engineer at Microsoft and a Stanford
alumnus (MSCE '84). Pre-Microsoft game credits include Aces of the
Deep, Cyberstorm, ProPilot, and Red Baron II. Credits at Microsoft
include Flight Simulator 2000, Combat Flight Simulator 2, Microsoft
Flight Simulator 2002, and Combat Flight Simulator 3.
DENNIS ("THRESH") FONG, winner of the 1997 Red Annihilation Quake
Tournament. Still undefeated in tournament play, Dennis is generally
acknowledged to be the best Quake/Doom player in the world. He
co-authored the Official Quake II Strategy Guide, and co-founded the
gamer's portal "gamers.com". He also writes a monthly column in PC
Gamer magazine.
GARY KING, of NVIDIA's game developer relations team. Gary is also a
Stanford alum and a veteran of the first CS248 game competition.
ZAK MIDDLETON, Master's student in Computer Science and a stalwart
teaching assistant in CS 248 this year.
Non-voting member:
GEORGE PETSCHNIGG, Master's student in Electrical Engineering, winner
of last year's videogame competition, and a stalwart teaching
assistant in CS 248 this year.
While grades for the assignments in CS 248 are based mainly on
"technical merit", entries in the video game competition will be
judged on technical merit, compelling game play, and originality.
Students are not required to participate in this competition.
Here is the schedule of events:
Wednesday, December 4:
9:00 - 3:30 Grading of video games (course students only)
3:30 - 4:00 Professor and TAs meet to choose 7-8 finalist teams
4:00 Public part of video game competition begins
4:00 - 5:30 Finalists present their games to the jury
5:30 - 5:45 Jury retires to consider their decision
5:45 Announcement of winners
5:45 - 6:30 Continued heavy partying
There will be one grand prize - an all-expenses-paid trip to Siggraph
2003 in San Diego next summer, and one second-place prize - dinner for
two at Il Fornaio in Palo Alto. In addition, every member of a
finalist team will receive a current video game title for the PC
platform, generously donated by Electronic Arts. Finally, there will
also be a special prize given for the wackiest or most daring
submission - a Microsoft Xbox, generously donated by NVIDIA. If the
grand prize is won by a team, it must be split among the team members.
The second-place prize will be duplicated as necessary to cover the
team. Only one Xbox will be awarded.
Refreshments will be served beginning at 4:00pm. Finalists' entries
will be "hung" on the PCs in the graphics labs and will be available
for viewing throughout the judging and party.
____________
EE380: COMPUTER SYSTEMS LAB COLLOQUIUM
on Wednesday, 4 December 2002, 4:15pm to 5:30pm
NEC Auditorium (B03), Gates Computer Science Building
http://ee380.stanford.edu/contents.html
Mozilla and Open Source,
Our Project Dynamics, and Exploitable Technologies
Scott Collins
mozilla.org and AOL/Netscape
This presentation is an introduction to Open Source development on a
large scale from the perspective and experiences of the Mozilla
project. It covers Open Source, the strategies and roles that have
enabled successful distributed development in this project, and a
breakdown of the resulting exploitable machinery and technologies in
mozilla. This is a technical talk that will be useful to people who
want to better understand what Open Source is, how to succeed in an
Open Source project of their own, or who want to hack on mozilla code
or UI, either as a contributor, or for their own projects (or both!).
Learn what mozilla can do for you, both now and when you're looking
for a job!
About the speaker: Scott Collins has been a participant and
contributor on the mozilla project since its inception. Some of his
initial contributions to the project were chronicled in the PBS
documentary "Code Rush"[2] as Netscape undertook efforts to put their
browser into the public domain. He is currently the lead (technical)
evangelist of the "Mozilla University"[3] effort that is underway to
promote the use, understanding, research and development of mozilla
technologies in a variety of educational settings.
____________
CSLI COGLUNCH
on Thursday, 5 December 2002, 12:15pm-1:30pm
Cordura Hall, Room 100
http://www-csli.stanford.edu/events/Coglunch/
Perceiving, Comprehending and Measuring Design Activity
through the Questions Asked while Designing
Ozgur Eris, Ph.D.
Center for Design Research, Mechanical Engineering, Stanford
Designing is question intensive. However, our understanding of what
designers accomplish by asking questions is rather limited. The
research I will present in this talk treats question asking while
designing as a process, and examines its key aspects. The theoretical
part of the research involves the development of a taxonomy of
questions asked while designing. The contribution of the
taxonomy--apart from proving to be a comprehensive analysis
framework--is its ability to differentiate between Deep Reasoning
Questions (DRQs), and Generative Design Questions (GDQs).
The empirical part of the research involves designing and conducting
experiments to test hypotheses generated from field observations. The
more significant hypotheses postulate relationships between question
asking processes of teams and their design processes, and between
their combined DRQ+GDQ asking rates and performance. Both hypotheses
were verified. The findings also demonstrated DRQ+GDQ utilization to
be a mechanism designers rely on for managing divergent and convergent
modes of thinking. Special consideration was given to laying out the
foundations of a unified design theory, which integrates the findings
on question asking with existing understandings on decision making in
design contexts.
____________
UC BERKELEY CIS SEMINAR
on Thursday, 5 December 2002, 4:00pm
Soda Hall 310 (UC Berkeley)
http://www.cs.berkeley.edu/~eyal/cis-seminar
Computer vision and control for soccer playing robots
Raul Rojas
Freie Universitaet Berlin
http://page.inf.fu-berlin.de/~rojas
http://www.fu-fighters.de/
We have built a team of autonomous soccer playing robots that has
participated in the last four RoboCup tournaments (the FU
Fighters). Soccer robots differ from industrial robots mainly in their
speed: decisions must be taken in fractions of a second in a very
dynamic and unpredictable environment. In this talk, I will explain
the main challenges that must be faced when building and programming
robots for this task. Our robots have a reactive control architecture
organized in layers. A hierarchy of virtual sensors triggers a
hierarchy of behaviors. Behavior modi can excite or inhibit other
behaviors. I will also explain how we process the images from the
video camera in order to localize the robots in the field. In one of
our teams, there is a camera above the field, in another team every
robot carries its own camera. I will explain how to calibrate the
geometry and the colors automatically. Finally, I will talk about our
current and future work applying reinforcement learning to the control
architecture. I will show a video with scenes of the last tournament
in Japan.
The FU Fighters have won three times second place at the RoboCup world
championship (Stockholm 1999, Melbourne 2000, and Fukuoka 2002) and
are two-time winners of the most important European Championship.
____________
UC BERKELEY ELECTRONIC SYSTEMS DESIGN SEMINAR
on Thursday, 5 December 2002, 4:00pm - 5:00pm
DOP Center Classroom (540AB), Cory Hall (UC Berkeley)
http://www-cad.eecs.berkeley.edu/esd-seminar
The Design of a Formal Property-Specification Language
Moshe Y. Vardi
Rice University
In recent years, the need for formal specification languages is
growing rapidly as the functional validation environment in
semiconductor design is changing to include more and more validation
engines based on formal verification technologies. In particular, the
usage of Formal Equivalence Verification and Formal Property
Verification is growing, new symbolic simulation engines are
introduced and hybrid environments of scalar and symbolic simulators
are developed. To facilitate the use of these new-generation
validation engines - properties, checkers and reference models need to
be developed in a formal language.
In this talk we describe the design of the ForSpec Temporal Logic
(FTL), the new temporal logic of ForSpec, Intel's new formal
property-specification language, which is today part of Synopsis
OpenVera hardware verification language (http://www.open-vera.com).
The key features of FTL are: it is a linear temporal logic, based on
Pnueli's LTL, it enables the user to define temporal connectives over
time windows, it enables the user to define regular events, which are
regular sequences of Boolean events, and then relate such events via
special connectives, and it contains constructs that enable the user
to model multiple clock and reset signals, which is useful in the
verification of globally asynchronous and locally synchronous hardware
designs. The focus of the talk is on design rationale, rather than a
detailed language description.
About the Speaker: Moshe Y. Vardi is Karen Ostrum George Professor
in Computational Engineering and Chair of Computer Science at Rice
University. His interests focus on applications of logic to computer
science, including database theory, finite-model theory, knowledge in
multi-agent systems, computer-aided verification and reasoning, and
teaching logic across the curriculum.
____________
SSP: SYMBOLIC SYSTEMS FORUM
on Thursday, 5 December 2002, 4:15pm
Bldg. 380:380C (Math Corner)
http://symsys.stanford.edu:8081/ssp-dynamic/servlet/ssp_events
Linguistically Rich Statistical Models of Language
Joseph Smarr
Symbolic Systems Program
Nearly every prominent vision of the future includes people conversing
with computers using natural language. Realizing this vision requires
solving both "general AI problems" (knowledge representation, common
sense reasoning, etc.) and language-specific problems (coping with the
complexity, ambiguity, and flexibility of natural language). Much
progress has been made on this latter front in the last half-century
by two rather disconnected fields of research: Theoretical Linguistics
and Natural Language Processing.
Theoretical Linguistics seeks to build rich logical representations of
language that explicate its structure and meaning, and in so doing,
describe the broad sets of valid and invalid utterances. NLP in
contrast has tended to settle for simpler representations that allow
robust processing of everyday language use, often relying on a
probabilistic (rather than categorical) view of language.
Now there is a historic convergence of the two fields taking
place-Theoretical Linguists are building applied systems using
techniques from NLP like statistical disambiguation, and NLP
researchers are adopting richer, more linguistically sophisticated
models for traditional NLP tasks like Information Extraction. This
talk will describe in more detail the circumstances and substance of
this convergence, highlighting recent work in service of the unified
goal of building linguistically rich statistical models of language.
____________
NETWORKING LECTURE
on Thursday, 5 December 2002, 4:15pm
Packard EE 101
OceanStore: Toward Global Scale, Self-Repairing, Secure,
and Persistent Storage
John Kubiatowicz
Computer Science Division, UC Berkeley
The peer-to-peer revolution seems to promise that the vast, untrusted,
and unreliable resource of the Internet can lead to desirable
properties such as security, availability, and 1000-year durability.
In this talk, we will examine how this might be possible. We will
examine the mechanisms of OceanStore, a utility infrastructure
designed to span the globe and provide continuous access to persistent
information. The OceanStore model involves hundreds or thousands of
service providers cooperating to provide service.
To achieve its properties, OceanStore exploits the power of aggregates
-- many elements working together. Since OceanStore is comprised of
untrusted servers, data is protected through redundancy and
cryptographic techniques. To improve performance, data is allowed to
be cached anywhere, anytime. Its routing infrastructure is
self-repairing and utilizes path redundancy and continuous monitoring
to adapt to regional outages and denial of service attacks.
Monitoring also enhances performance through pro-active movement of
data. This talk will describe the mechanisms of OceanStore and
discuss the status of its implementation.
____________
STANFORD PHONOLOGY WORKSHOP
on Thursday, 5 December 2002, 5:30pm
Margaret Jacks Hall 460:126
http://www-linguistics.stanford.edu/pinterest/
A discourse-phonology interface in ASL:
Pronouncing two words at once
Mary Rose
Stanford University
Many signs in ASL (and other signed languages) require two hands. In
discourse, however, they are often pronounced with just one hand.
One-handed variants often occur while one hand holds the final segment
of a sign, or 'perseverates', and the other continues signing the
utterance. This talk will focus on perseveration that helps to
structure signed discourse, what I am calling discourse perseveration,
or DP. In DP, the perseverated sign serves as a reminder of its
referent, a 'conceptual landmark' for the ongoing discourse (Liddell,
in press; Dudis 2002). This requires the signer to pronounce two signs
at once, but this demand must fundamentally change the shape of
two-handed signs in the utterance. The talk will introduce the
phonetics of DP by outlining which features of either of the two
simultaneously produced signs adapt to the situation. I will describe
how the handshapes, places of articulation, palm orientation, and
movements of either sign varies in DP from its usual form. In some
two-handed words, the articulators are mirror images of each other, so
when they get pronounced with just one hand they lose only the
redundancy encoded in the second hand's features. Not all two-handed
signs can adapt so easily, however, because the hands do not always
exactly mirror each other. For example, if the two-handed sign
requires contact between the two hands, that contact is usually
preserved in discourse perseveration, but at the same time it is
manipulated so that the perseverating hand can keep doing its job. It
is not only the two-handed sign that changes its handshape, movement
or location, but those features of the perseverated sign may also
change. Discourse perseveration, then, causes a complex interaction
between two signs that are essentially being pronounced at the same
time, yet must be made intelligible to the interlocutor as instances
of their respective lexemes. I will briefly give some background on
ASL phonology and the structure of two-handed signs and then explain
the spatial aspects of ASL discourse strategies that establish the
context for discourse perseveration. I then turn to the phonetic
details of DP, laying out what features of each sign can or cannot
change to allow for the simultaneous production of the other. In
general, the interaction places quite strict constraints on the
phonetics of the perseverating hand, while the two-handed sign is
relatively phonetically 'flexible', so that it is sometimes
articulated radically differently from its citation form to allow the
perseveration to continue.
____________
CS547: HUMAN-COMPUTER INTERACTION SEMINAR
on Friday, 6 December 2002, 12:30-2:00pm
Gates B01
http://www-pcd.stanford.edu/cs547/
Electronic Chalk: A tool for live and remote teaching
Raul Rojas
Freie Universitaet Berlin
http://www.e-chalk.de/
E-Chalk is a system to record live lectures and transmit them over the
Internet. The lecturer writes directly on a large touch sensitive
surface (whiteboard or retro projection system) talking directly to
the class. Three streams are transmitted over the Internet: audio,
video, and board contents. The system is completely written in Java,
therefore the potential audience needs only an Internet browser for
receiving the three streams. It is possible to do postproduction of
the lectures by erasing, rearranging, or recording new segments.
The philosophy of E-Chalk is to provide "smart" tools for enhancing
the classroom experience. The lecturer can paste images on the board
directly from bookmarks. He or she can also call an algebraic server
to solve equations or plot functions (Mathematica from Wolfram
Research). It is also possible to access web sites directly, send a
textual query, and get a textual answer. It is possible, for example,
to ask for a translation, a calendar, or any information provided by
Web sites and integrate this into the board contents on the fly
. E-Chalk is intended to be controlled using handwriting
recognition. A first prototype of mathematical handwriting recognition
has been written.
E-Chalk can be used with a variety of hardware: touch sensitive
whiteboards, large touch screens, ultrasound digitizers, PDAs and
the new tablet PCs. It is also possible to use E-Chalk in
combination with video conferencing systems. In the talk, I will
explain what is our vision for the future and how pen based
computing fits into that picture.
About the speaker: Raul Rojas is a Professor of Artificial
Intelligence at the Department of Mathematics and Computer Science,
Freie Universitaet Berlin He is the author of several books about
Neural Networks and the History of Computing. He is also a member of
the editorial board of several journals and has published extensively
about theory and applications of Neural Networks. Rojas received his
PhD and "Habilitation" (an additional German degree after the PhD)
from Freie Universitaet Berlin. He was a Professor of Computer Science
at the Technical University of Vienna and at the University of Halle,
before moving to the FU Berlin in 1997. His research interests include
topics in AI, robotics, and multimedia tools for the
classroom. Prof. Rojas is the team leader of the FU-Fighters, a
robotic soccer team that has won three times second place at the
RoboCup robotic soccer world championship ( http://www.fu-fighters.de/
). He was also advisor to the team that developed the handwriting
recognition system for the Bundespost. The Bundespost mail-sorting
machines are now handling several million letters a day. The E-Chalk
system won the Entrepreneurs Multimedia Prize 2001 from the Ministry
of Economics and Research in Germany.
____________
NLP READING GROUP
on Friday, 6 December 2002, 2:15pm
Margaret Jacks Hall, Room 460:301
http://www-nlp.stanford.edu/nlp/nlpgroup.html
Natural Language Interaction with Intelligent Tutoring Systems
Brady Clark and Karl Schultz
Stanford University
Tutorial interaction poses a hard problem for dialogue systems
research: tutorial dialogue is mixed-initiative (tutors and students
ask questions) and multi-modal (tutors and students speak and point).
We have addressed these problems by developing an intelligent tutoring
system - using the Architecture for Conversational Intelligence
developed at CSLI - which allows speech and gestural input and
output. This work is novel in that our tutor is the first spoken
language intelligent tutoring system, whereas all other existing
tutors utilize text-based input.
In this talk we will give a brief overview of the research questions
we are currently addressing (broadly, how can natural language be used
to advantage in intelligent tutoring systems?) and the architecture of
the tutor we have built for shipboard damage control. We will end with
a live demo.
A draft paper is available at:
http://www-csli.stanford.edu/semlab/muri/its-architecture-CLASS.ps
- please do not cite or distribute it, since it is a draft!
____________
LINGUISTICS DEPARTMENT COLLOQUIUM
on Friday, 6 December 2002, 3:30pm
Margaret Jacks Hall, Room 460:126
http://www-linguistics.stanford.edu/colloq/
Evolutionary Optimality Theory
Gerhard Jaeger
Universitaat Potsdam
In the talk I will propose a variant of Boersma's Gradual Learning
Algorithm for Stochastic Optimality Theory. While in the original
version the learner is always (or tries to become) a speaker, I assume
that the learner is both speaker and hearer. This learning theory is
applied to the OT system from Aissen (2000), which was developed to
explain the typology of differential case marking. It can be shown
that the constraint sub-hierarchies that Aissen assumes to be
universal follow from the statistical patterns of language use that
have been uncovered in several corpus studies, if one adopts the
bidirectional learning approach.
Not all case marking patterns are learnable by the Bidirectional
Gradual Learning Algorithm (BiGLA), and some patterns are easier to
learn than others. If learning with limited resources is repeated over
several generations, one can distinguish stable and instable language
types, and certain tendencies for language change emerge. In the
second part of the talk I will present and discuss some experimental
findings on the basis of this evolutionary approach.
____________
SPECIAL UNIVERSITY ORAL EXAMINATION
on Monday, 9 December 2002, 2:00pm-3:00pm
Gates 260
(Refreshments served at 1:45 PM)
Efficient Algorithms for Option Pricing
Donald Aingworth
Computer Science, Stanford University
Over the past 30 years, there has been considerable interest in
pricing options. Even before the internet craze began, options have
been popular, both as a speculative tool, and for hedging risk. The
work of Fisher Black, Myron Scholes, and R. C. Merton provided a
common framework for the pricing of options, but did not provide an
efficient solution for many popular types of options. The discrete
model put forth by John Cox, Stephen Ross, and Mark Rubenstein allowed
for effective pricing for many more option types, but still leaves
open the pricing of many important option types. For example,
path-dependent options, and models of stock movement more
sophisticated than geometric Brownian motion remained hard to solve.
My research provides several improved techniques for the pricing of
options in the lattice model.
I define a new technique that augments traditional Monte-Carlo
methods, showing that the existence of additional data can lead to
improvements on random sampling. In particular, in many
computer-science applications, the probability of an event can easily
be observed when the event itself if observed. This allows one to
"condense" observations, decreasing the variance of an estimator.
The technique is general, and is easily applied to option pricing,
where sophisticated Monte-Carlo methods are often the best-known
method of pricing options.
Path-dependent options are often particularly intractable. Here,
options have memory, and their value is a function of their history.
As a result, even in the discrete world, efficient, deterministic
algorithms are not known for many common option types. I define a
framework within which certain path-dependent options can be
approximated. This framework is applied to Asian options, also known
as average-price options. The technique works for both European, and
American options, and is the first known algorithm to give guaranteed
error bounds in polynomial time.
One weakness of many option-pricing algorithms is the assumptions that
are made on the underlying. For example, stocks do not precisely
follow geometric Brownian motion because extreme events are more
likely than would be implied ("fat-tails"). I define a model where
volatility switches between states in a Markov process. As the number
of states increases, this model converges on the stochastic volatility
model, a popular framework for which there is no known algorithm for
pricing American options. The algorithms in the new model are
efficient, and can easily be calibrated against empirical work.
____________
BERKELEY LINGUISTICS DEPARTMENT COLLOQUIUM
on Monday, 9 December 2002, 4:00pm
182 Dwinelle (Berkeley)
http://linguistics.berkeley.edu/lingdept/Current/events.html
Coding Strategies in Comparative Bantu
and Implications for Diachronic Syntax
Yukiko Morimoto
University of Duesseldorf
That linguistic change is always preceded by social, regional, or
stylistic variation is a fundamental notion in variation studies since
the beginning of modern sociolinguistic theory (Labov 1966). In
diachronic typology (Greenberg 1969, 1978), so-called "exceptions" to
synchronic universals are interpreted as unstable transitory stages in
linguistic change. If synchronic variation--both within and across
languages--and linguistic change are so inextricably linked, then we
should not only expect both types of variation to be equally
systematic, but also exist within a single typological space (see
Bresnan and Deo 2001). Yet, much of the research on linguistic
typology and (language-internal) variation (or "optionality")
continues without much reference to each other. Using stochastic
ranking of universal, violable constraints in Optimality Theory
(Prince and Smolensky 1993; see for example, Boersma 1997, 1998 for
Stochastic OT), my goal in the present work is to relate these two
types of variation to linguistic change in the domain of agreement in
Bantu languages.
Majority of Bantu languages encode subjects by head-marking and
objects by word order. This reflects a point in the historical
process whereby the positional licensing of objects becomes obligatory
due to the loss of inflectional morphology (= object agreement on the
verb). As such, we observe considerable variation both across and
within languages in the use of head-marking morphology for
objects. Moreover, this variation in object agreement displays a
striking parallelism with differential object marking (DOM) in case
marking languages (cf. Bossong 1985). In recent work, Aissen (2000)
provides an Optimality Theoretic account of DOM under a single
generalization in (1).
(1) The higher in prominence a direct object the more likely it is to
be overtly case marked--where the dimensions along which
prominence is assessed include animacy (1a) and definiteness (1b).
a. Animacy: Human > Animate > Inanimate
b. Definiteness: Pronoun > Name > Definite >
Indef.Specific > Non-specific
Among Bantu languages, Makua requires object marking (OM) for humans
(Stucky 1981, 1983). In KiSwahili, OM is optional for inanimates but
obligatory for animates (Bokamba 1981; also Wald 1979). Kihung'an
requires OM for definite object NPs (Takizala 1973); Zulu exhibits the
same pattern (Wald 1979). In Kiyaka (Kidima 1984) OM is obligatory for
personal names but optional for definite NPs. In Kichaga, OM is
obligatory for (morphologically free) personal pronouns (Bresnan and
Moshi 1993:52). In Chichewa and many other Bantu languages, OM is used
exclusively for topic-anaphoricity and is in complementary
distribution with any clause-internal object NP (cf. Morolong and
Hyman 1977).
Preliminary quantitative results from the Helsinki corpus of Swahili
suggest that the presence/absence of OM is not categorical even with
relatively marked animate object NPs (based on Seidl and Dimitriadis's
1997 study). Of the 312 sentences with overt object NPs in
spoken/written text, OM co-occurs with 72% of the animate object NPs
(104/312), and with 12% of the inanimates (20/312). The systematic
variability in the morphosyntax of a single language indicated by
these quantitative results challenges the traditional conception of
generative syntax, in which parameter values are either active or
inactive throughout the whole grammar of a given language.
The present talk outlines a stochastic OT approach to DOM in Bantu and
highlights the idea that variation across and within languages operate
within a single typological space and can be related to linguistic
change in a principled manner.
____________
SNRC INDUSTRY SEMINAR
on Tuesday, 10 December 2002, 4:15pm
Packard 101
http://snrc.stanford.edu/events/industry-seminar/
Approaches to 4G Access Network Architectures
James Kempf
Research Fellow, DoCoMo Labs - USA
With 3G wireless network deployments proceeding at a slow pace, it may
seem premature to start examining 4G access networks, but with ITU-R
beginning discussions for international 4G spectrum allocation, some
consideration of how 4G access networks should be architected seems in
order. In this talk, we will examine some of the problems plaguing
current 3G deployments specifically related to access network
architecture. With the Internet Protocol as the unifying abstraction,
4G access networks must achieve lower cost per km of coverage and
higher aggregate bandwidth than 3G networks, while accommodating the
physical layer properties that characterize the most likely 4G
spectrum. An intermediate step of an IP friendly WAN wireless
protocol, such as the newly organized IEEE 802.16 effort, is an
attractive option.
About the speaker: Dr. James Kempf graduated from University of
Arizona in 1983 with a Ph.D. in Systems Engineering, and moved to
Silicon Valley to work at Hewlett-Packard. In 1989, James moved to Sun
Microsystems, where he worked on a variety of product and research
projects, including a SPARC-powered tablet computer with a wireless
networking connection using a first generation, frequency hopping
802.11 card.
Since 2001, James has been a Research Fellow at DoCoMo USA Labs,
primarily working on next generation radio access networks. James is
active in IETF, co-chairing the Seamoby and SEND working groups, and
is currently serving on IAB.
____________
CSLI SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION
on Thursday, 12 December 2002, 4:15pm to 5:30pm
Cordura 100
http://cll.stanford.edu/scla.html
Hierarchies of Models: Toward Understanding Planetary Nebulae
Kevin H. Knuth
NASA Ames Research Center,
Computational Sciences Department, Code IC
Stars like our sun (initial masses between 0.8 to 8 solar masses) end
their lives as swollen red giants surrounded by cool extended
atmospheres. The nuclear reactions in their cores create carbon,
nitrogen and oxygen, which are transported by convection to the outer
envelope of the stellar atmosphere. As the star finally collapses to
become a white dwarf, this envelope is expelled from the star to form
a planetary nebula (PN) rich in organic molecules. The physics,
dynamics, and chemistry of these nebulae are poorly understood and
have implications not only for our understanding of the stellar life
cycle but also for organic astrochemistry and the creation of
prebiotic molecules in interstellar space.
We are working toward generating three-dimensional models of planetary
nebulae (PNe), which include the size, orientation, shape, expansion
rate and mass distribution of the nebula. Such a reconstruction of a
PN is a challenging problem for several reasons. First, the data
consist of images (taken from a single viewpoint) obtained over time
from the Hubble Space Telescope (HST) and ground-based long-slit
spectra. Second, the fact that we have two disparate data types
requires that we utilize a method that allows these data to be used
together to obtain a solution. To address these first two challenges
we employ Bayesian model estimation using a parameterized physical
model that incorporates much prior information about the known physics
of the PN. As we have found that the forward problem of this
comprehensive model is extremely time consuming, we have introduced a
method relying on a hierarchical set of models, where each model in
the hierarchy allows us to estimate increasingly more detail. In this
talk, we describe these analysis techniques and explore the advantages
and disadvantages of employing such a set of hierarchical models.
This talk describes joint work with Arsen R. Hajian.
____________
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
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 (i.e., no abstracts). 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/Archive/calendar/current.shtml
and the archives at
http://www-csli.stanford.edu/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.
and
news://news.stanford.edu/su.events
Information about CSLI's research program is available at
http://www-csli.stanford.edu/
For maps to the Stanford University campus see
http://www.stanford.edu/home/visitors/maps.html
____________