[Date Prev][Date Next][Date Index]
CSLI Calendar, Wednesday, 24 September 2003, vol. 19:4
CSLI CALENDAR OF PUBLIC EVENTS
______________________________________________________________________
24 September 2003 Stanford Vol. 19, No. 4
______________________________________________________________________
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 24 SEPTEMBER 2003 TO 3 OCTOBER 2003
WEDNESDAY, 24 SEPTEMBER 2003
4:15pm EE380: Computer Systems Lab Colloquium
Gates B03 (NEC Auditorium)
"Team Overbot: 200 miles across the desert. 10 hours. No driver"
John Nagle
http://www.overbot.com/
http://ee380.stanford..edu/contents.html
Abstract below
THURSDAY, 25 SEPTEMBER 2003
4:00pm PARC Forum
George Pake Auditorium at PARC
"Angstrovision, Imaging at Tens of Nanometers:
Trevor Bacolini/Scott Mize
Angstrovision
http://www.parc.com/forum/
4:00pm UC Berkeley CIS Seminar
Soda Hall 310 (UC Berkeley)
"Multiagent Planning with Factored MDPs"
Carlos Guestrin
Intel Labs
http://www.eecs.berkeley.edu/~wainwrig/cis-seminar
Abstract below
4:15pm Fundamental Themes in Neuroscience Seminar
Munzer Auditorium, Beckman B060
"Neuronal Growth and Death: Neurofilaments, SOD1 and Lou Gehrig"
Don Clevelend, Ph.D.
UC San Diego, Ludwig Institute for Cancer Research
Seminar Host: Yanmin Yang (Neurology and Neurological Sciences)
http://www-med.stanford.edu/sbrc/calendar/
FRIDAY, 26 SEPTEMBER 2003
11:00am SRI AI Seminar Series
EJ228, SRI International
"Semantic Web for Earth and Environmental Terminology (SWEET)"
Rob Raskin
Jet Propulsion Lab
http://www.ai.sri.com/seminars/
Abstract below
12 noon UC Berkeley Oxyopia Lecture
489 Minor Hall (UC Berkeley)
"Assembly of receptive fields in cat visual cortex"
David Ferster
Smith Kettlewell, San Francisco
http://spectacle.berkeley.edu/ucbso/oxyopia/oxy_current.html
12:30pm CS547: Human-Computer Interaction Seminar
Gates B01
"Strategic Behavioral Transformations in Immersive
Collaborative Virtual Environments"
Jeremy Bailenson
Communication, Stanford University
http://www-pcd.stanford.edu/cs547/
Abstract below
2:15pm Linguistics 237D: An Introduction to Computational Word Learning
Ventura 17
first class (replaces NLP reading group)
http://www-nlp.stanford.edu/nlp/nlpgroup.html
Information below
MONDAY, 29 SEPTEMBER 2003
1:50pm 2003 CSLI-Japan Workshop on Natural Language Processing
Ventura 17
Information below
3:30pm Social Lab
Jordan Hall 420:050
Title to be announced
Marcus Feldman
Wohlford Professor In The School of Humanities and Sciences, Stanford
http://www-psych.stanford.edu/news.html#social_lab
4:00pm UC Berkeley Philosophy Lecture
Howison Library (UC Berkeley)
"Having Concepts, a brief refutation of the 20th century"
Jerry Fodor
Rutgers University
http://philosophy.berkeley.edu/html/events/philatberkeley.html
4:00pm UC Berkeley Linguistics Department Colloquium
370 Dwinelle (Berkeley)
BLC lecture
Shirley Brice Heath
http://linguistics.berkeley.edu/lingdept/Current/events.html
4:15pm CS528: Broad Area Colloquium in AI,
Geometry, Graphics, Robotics, and Vision
TCSeq 200
"High-performance imaging using dense camera arrays"
Marc Levoy
Stanford
http://graphics.stanford.edu/courses/cs528/
Abstract below (see also course announcement below)
TUESDAY, 30 SEPTEMBER 2003
10:00am SRI AI Seminar Series
EJ228, SRI International
"A Garbage Collection Algorithm and Handling Imperfect Data
in a Monitoring and Optimization System for Coal-Fired Power Plants"
Janet Murdock
http://www.ai.sri.com/seminars/
Abstract below
4:15pm Logic Seminar
Bldg. 380:381T (math corner)
"Propositional logic of continuous transformations"
G. Mints, T. Zhang
Stanford University
http://www-logic.stanford.edu/seminars.html
Abstract below
4:15pm SNRC Industry Seminar
Gates B03
"Resource Management in Utility Computing Environments"
Akhil Sahai
Hewlett Packard Laboratories, Palo Alto
http://snrc.stanford.edu/events/industry-seminar/
Abstract below
WEDNESDAY, 1 OCTOBER 2003
4:15pm EE380: Computer Systems Lab Colloquium
Gates B03 (NEC Auditorium)
"DNS Typosquatting: Method, Effects, Counteraction"
Paul Vixie
Internet Software Consortium
http://ee380.stanford.edu/contents.html
Abstract below
5:30pm Immanuel Kant Lectures
Bldg 260:113
Lecture 1: "Another I: Representing Conscious States"
Christopher Peacocke
New York University
http://www-philosophy.stanford.edu/ce.html
THURSDAY, 2 OCTOBER 2003
12:15pm CSLI CogLunch
Cordura Hall, Room 100
Title to be announced
Alison Gopnik
http://www-csli.stanford.edu/events/Coglunch/
4:00pm Biomedical Ethics Seminar
Lucille Packard Children's Hospital Auditorium
"What is Wrong with Improving Our Brains?"
Art Caplan, Ph.D
Director, Center of Bioethics
University of Pennsylvania
http://scbe.stanford.edu//events/grand_rounds.html
4:00pm SRI AI Seminar Series
Enterprise Room, SRI International
"Adventures in Artificial Intelligence"
Nils Nilsson
Computer Science, Stanford
http://www.ai.sri.com/seminars/
Abstract below
4:00pm PARC Forum
George Pake Auditorium at PARC
"Sonoluminescence: The Star in a Jar"
Seth Putterman
Physics Department, UCLA
http://www.parc.com/forum/
4:00pm UC Berkeley CIS Seminar
Soda Hall 310 (UC Berkeley)
"The difficulty of making a robot go straight"
Sebastian Thrun
Stanford
http://www.eecs.berkeley.edu/~wainwrig/cis-seminar
Abstract below
4:00pm UC Berkeley Philosophy Lecture
Howison Library (UC Berkeley)
'context'
Jerry Fodor
Rutgers University
http://philosophy.berkeley.edu/html/events/philatberkeley.html
4:15pm Fundamental Themes in Neuroscience Seminar
Munzer Auditorium, Beckman B060
"Multimodal Feedback Integration in Sensorimotor Control and Learning"
Philip Sabes, Ph.D.
University of California at San Francisco, Physiology
Seminar Host: Krishna Shenoy (Department of Electrical Engineering)
http://www-med.stanford.edu/sbrc/calendar/
5:30pm Immanuel Kant Lectures
Bldg 260:113
Lecture 2: "Another I: Representing Conscious States"
Christopher Peacocke
New York University
http://www-philosophy.stanford.edu/ce.html
FRIDAY, 3 OCTOBER 2003
12:30pm CS547: Human-Computer Interaction Seminar
Gates B01
"Designing Technology"
Bill Moggridge
Ideo
http://www-pcd.stanford.edu/cs547/
3:15pm Immanuel Kant Lectures
Bldg 260:113
Discussion: "Another I: Representing Conscious States"
Christopher Peacocke
New York University
http://www-philosophy.stanford.edu/ce.html
SATURDAY, 4 OCTOBER 2003
All day Workshop on the History of Artificial Life
http://www.stanford.edu/dept/HPS/ALworkshop/
be warned, web page has too much dependence on flash
SUNDAY, 5 OCTOBER 2003
All day Workshop on the History of Artificial Life
http://www.stanford.edu/dept/HPS/ALworkshop/
be warned, web page has too much dependence on flash
____________
Stanford Blood Center status: critical shortage of O+ and O-; shortage
of A-, A+, and B+. For an appointment:
http://bloodcenter.stanford.edu/ or call 650-723-7831. It only takes
an hour of your time.
____________
COURSE ANNOUNCEMENTS
CS528: BROAD AREA COLLOQUIUM IN AI,
GEOMETRY, GRAPHICS, VISION AND ROBOTICS
Mondays, 4:15pm, TCSeq 200
http://graphics.stanford.edu/courses/cs528/
The Broad Area Colloquium series begins again this quarter on Monday
afternoons. This colloquium is intended to bring established and
senior researchers from the fields of AI, Geometry, Graphics,
Robotics, and Vision to discuss and explain broad considerations and
high-level tasks which the relevant communities are addressing. The
talks are intended to create awareness and interest for all of the
members of these communities, hopefully bridging gaps and creating
collaborations.
Those who would like to receive weekly reminder announcements of the
seminar may subscribe to the mailing list
broad-area-cs-colloquium@lists. To be added, send e-mail to
majordomo@lists.stanford.edu with a blank subject line and "subscribe
broad-area-cs-colloquium" in the body.
Students who wish to attend this series and receive credit (1 unit)
may enroll in CS 528. Those who miss not more than one of the talks
given during the quarter will receive credit.
The first lecture of the quarter will be Monday, 9/29/03.
Check the web site for details:
http://graphics.stanford.edu/courses/cs528/ Questions may be sent to
bac-coordinators@cs.stanford.edu
CS359: Little Engines of Proof
MW 4:15-5:30pm GATES 400
Lecturers: L. de Moura, H. Ruess, N. Shankar, and A. Tiwari
From SRI International
We survey the state of the art in automated deduction through a
collection of useful algorithms and semi-algorithms. Topics include
propositional satisfiability, ground decision procedures for equality
and inequality, combination methods, quantifier elimination,
rewriting, and proof search. The emphasis is on simple and unified
theoretical foundations, efficient implementation, and novel
applications. Our presentation includes recent research results and
complements the contents of the Fall Quarter Logic Seminars. A basic
background in logic and computation is adequate. The course should be
of interest to graduate students in logic, computer science, and
philosophy.
ChE459: FRONTIERS IN INTERDISCIPLINARY BIOSCIENCES
Clark Center Auditorium
Thursdays, 4:15pm
http://biox.stanford.edu/courses/459_announce.html
An introduction to cutting-edge research involving interdisciplinary
approaches to bioscience and biotechnology, intended for both
specialists and non-specialists. Organized and sponsored by the
Stanford Bio-X Program. Three seminars each quarter address a broad
set of scientific and technical themes related to interdisciplinary
approaches to important issues in bioengineering, medicine, and the
chemical, physical and biological sciences.
Leading investigators from Stanford and throughout the world present
the latest breakthroughs and endeavours that cut broadly across many
core disciplines. Pre-seminars introduce basic concepts and provide
background for non-experts. Registered students attend all
pre-seminars in advance of the primary seminars, others welcome.
Prerequisite: keen interest in all of science, engineering, and
medicine with particular interest in life itself. Recommended: basic
knowledge of mathematics, biology, chemistry, and physics. 1 unit,
To receive seminar reminders and course information, please subscribe
to the "Frontiers in Interdisciplinary Biosciences" email list. To
subscribe to the list, please send an email to:
majordomo@lists.stanford.edu
In the body of the email type:
subscribe 459_course
LOGICAL METHODS IN THE HUMANITIES WORKSHOP
This is a Stanford Humanities Center workshop, funded through the
Mellon foundation. It is open to all interested faculty and students.
The primary focus of the workshop will be to discuss research in logic
and the history and philosophy of the exact sciences. We will have
regular meetings where faculty and graduate students will present
their work, as well as talks by several visitors.
Our regular meetings will be held on Friday afternoons, from 3-5pm.
Any exceptions will be announced. Our first meeting will be on October
10th.
Email ayap@stanford.edu if you are interested in receiving
announcements about our meetings. This will be our main way of
communicating meeting times, topics, etc.
Our list of invited speakers include:
Harvey Friedman (Ohio State)
Paolo Mancosu (UC Berkeley)
Michael Detlefsen (Notre Dame)
Kenneth Manders (Pitt)
Douglas Jesseph (North Carolina State)
If you have any questions, feel free to contact Audrey Yap
(ayap@stanford.edu) or the faculty coordinator, Andrew Arana
(aarana@stanford.edu).
____________
EE380: COMPUTER SYSTEMS LAB COLLOQUIUM
on Wednesday, 24 September 2003, 4:15pm to 5:30pm
NEC Auditorium (B03), Gates Computer Science Building
http://ee380.stanford.edu/contents.html
"200 Miles across the desert.
10 hours. No Driver."
John Nagle
Team Overbot
http://www.overbot.com/
Team Overbot is building a robot vehicle designed to travel 250 miles
across the Mojave Desert for a Department of Defense
competition. There's nobody on board, no remote control, and the route
is not disclosed in advance. It's the most ambitious ground robotics
project today. We'll describe how we're doing it and are bringing the
robot vehicle.
About the speaker: John Nagle (Stanford CS '85) keeps a low profile.
____________
UC BERKELEY CIS SEMINAR
on Thursday, 25 September 2003, 4:00pm-5:00pm
Soda Hall 310 (UC Berkeley)
http://www.eecs.berkeley.edu/~wainwrig/cis-seminar
"Multiagent Planning with Factored MDPs"
Carlos Guestrin
Intel Labs
Many real-world tasks require multiple agents to coordinate in order
to achieve a common goal. Examples include multi-robot systems,
network routing and supply chain management. Unfortunately, these
large-scale problems are often quite complex, involving many states
and multiple decision makers. Factored Markov decision processes
(MDPs) allow us to represent a complex transition model compactly
using a dynamic Bayesian network. This representation often allows an
exponential reduction in the representation size, but the complexity
of exact solution algorithms for such MDPs grows exponentially in the
number of variables describing the state of the system and in the
number of agents.
This talk presents a framework for approximate planning that can
exploit structure in a factored MDP to solve problems with many
trillions of states and actions. The talk will focus on three key
elements:
* Factored Linear Programs -- A novel LP decomposition technique,
using ideas from inference in Bayesian networks, which can yield
exponential reductions in planning time.
* Distributed Coordination -- A distributed multiagent decision
making algorithm, where the coordination structure arises naturally
from the system dynamics.
* Generalization in Relational MDPs -- A method for learning general
solutions from solved tasks, that allows us to act in new scenarios
without replanning.
We demonstrate our approach on the task of multiagent coordination in
a real strategic computer war game.
____________
SRI AI SEMINAR SERIES
on Friday, 26 September 2003, 11:00am
EJ228, SRI International
http://www.ai.sri.com/seminars/
"Semantic Web for Earth and Environmental Terminology (SWEET)"
Rob Raskin
Jet Propulsion Lab
SWEET is an effort to improve discovery of Earth science data and
information resources. The project consists of a collection of
ontologies and an associated search tool that eliminates the need for
exact keyword matches. The ontologies are stored in a Postgres DBMS
and are converted on demand into OWL XML format. The ontologies
include both orthogonal concepts (physical property, substance, space,
time, etc.) and integrative concepts (e.g. phenomena) which span
multiple ontology spaces. A numerical ontology was built to provide an
underlying base for the space, time, and physical property ontologies.
We also developed a tool that can query external databases (such as
gazetteers) in response to a request, and then incorporate the new
knowledge within the ontologies. SWEET enables a common semantic
framework for defining, classifying, discovering, sharing, and
accessing Earth science knowledge. This semantic framework is being
incorporated into other NASA-led initiatives that benefit from
semantic interoperability, including the: Earth Science Markup
Language (ESML), Earth Science Modeling Framework (ESMF), and Earth
Science Information Partners Federation Interactive Network for
Discovery (ESIP-FIND).
About the Speaker: Rob Raskin is a Senior Technical Staff Member in
the Earth Science Data Systems Section at NASAs Jet Propulsion
Laboratory. He is Principal Investigator of the NASA-funded SWEET
project and is Manager of Data Access Services at JPLs Physical
Oceanography Distributed Active Archive Center (PO.DAAC). He is Chair
of the Information Technology & Interoperability Committee of the
Earth Science Information Partner (ESIP) Federation. Prior to coming
to JPL, he was a Visiting Fellow at the National Center for Geographic
Information & Analysis at UC-Santa Barbara. He holds a Ph.D. in
Atmospheric Science from the University of Michigan.
____________
CS547: HUMAN-COMPUTER INTERACTION SEMINAR
on Friday, 26 September 2003, 12:30-2:00pm
Gates B01
http://www-pcd.stanford.edu/cs547/
"Strategic Behavioral Transformations in
Immersive Collaborative Virtual Environments"
Jeremy Bailenson
Communication, Stanford
http://communication.stanford.edu/faculty/bailenson.html
Over time, our mode of remote communication has evolved from written
letters to telephones, email, internet chat rooms, and
video-conferences. Similarly, collaborative virtual environments
(CVEs) promise to further change the nature of remote
interaction. CVEs are systems which track verbal and nonverbal signals
of multiple interactants and render those signals onto avatars,
three-dimensional, digital representations of people in a shared
digital space. In this talk, I describe a series of projects that
explore the manners in which CVEs can qualitatively change the nature
of remote communication. Unlike telephone conversations and
video-conferences, interactants in CVEs have the ability to
systematically filter the physical appearance and behavioral actions
of their avatars in the eyes of their conversational partners,
amplifying or suppressing features and nonverbal signals in real-time
for strategic purposes. These transformations can have a drastic
impact on interactants' persuasive and instructional
abilities. Furthermore, using CVEs, behavioral researchers can use
this mismatch between performed and perceived behavior as a tool to
examine complex patterns of nonverbal behavior with nearly perfect
experimental control and great precision. Implications for
communications systems, marketing strategies, and behavioral science
research will be discussed..
About the Speaker: Jeremy Bailenson received his Ph.D. in Cognitive
Psychology from Northwestern University in 2000, where he developed a
mathematical model of persuasive argumentation as well as studied
cultural differences in reasoning and categorization. For the next
four years he worked as a Post-Doctoral Fellow in the Research Center
for Virtual Environments and Behavior at the University of California,
Santa Barbara, studying nonverbal behavior in Immersive Virtual
Reality (IVR). Currently, he is an assistant professor in the
Department of Communication at Stanford University. His research is
funded by the National Science Foundation.
____________
LINGUISTICS 237D: An Introduction to Computational Word Learning
on Friday, 26 September 2003, 2:15pm - 3:30pm
Ventura 17
http://www-nlp.stanford.edu/nlp/nlpgroup.html
Introduction to computational methods for learning the syntax and
semantics of words and multiword expressions based on evidence from
language resources (such as dictionaries and thesauri) and actual
language usage in text corpora.
The NLP reading group for the fall quarter is going to be run as a
special course / seminar on Computational Word Learning, by Tim and
myself. We hope to introduce some of the most interesting approaches
in recent years, and to discuss current research and ongoing
developments. We will be delighted to consider any suggestions of
material on this topic which you would like to present or would like
us to make sure we cover.
The first meeting will be next Friday (26th) at 2.15pm, though we are
considering moving this to Fridays at 11am for subsequent weeks,
because Friday afternoons have proved troublesome for some of us in
the past. If you can't make it either on Fridays at 11am or 2.15pm,
and would like to be able to attend, please let us know and your vote
will be counted.
The traditional NLP reading group will resume in winter quarter, so
please have your cutting-edge research and favorite papers ready for
then.
____________
2003 CSLI-JAPAN WORKSHOP ON NATURAL LANGUAGE PROCESSING
on Monday, 29 September 2003, 1:50pm-5:00pm
Ventura 17
The following informal workshop will showcase some of the more recent
research to take place at CSLI and affiliate sites in Japan. Note
that space is limited in Ventura 17.
Program:
13:50-14:15 Isao Goto (NHK),
"NLP research at NHK"
I will introduce NLP research at NHK (Japan Broadcasting
Corporation) and describe my research on statistical
transliteration.
14:15-14:40 Dominic Widdows (CSLI),
"Orthogonal Negation in Vector Spaces for Modelling
Word-Meanings and Document Retrieval"
Standard IR systems can process queries such as "web NOT
internet", enabling users who are interested in arachnids to
avoid documents about computing. The documents retrieved for
such a query should be irrelevant to the negated query
term. Most systems implement this by reprocessing results
after retrieval to remove documents containing the unwanted
string of letters. This paper describes and evaluates a
theoretically motivated method for removing unwanted meanings
directly from the original query in vector models, with the
same vector negation operator as used in quantum
logic. Irrelevance in vector spaces is modelled using
orthogonality, so query vectors are made orthogonal to the
negated term or terms. As well as removing unwanted terms,
this form of vector negation reduces the occurrence of
synonyms and neighbours of the negated terms by as much as 76%
compared with standard Boolean methods. By altering the query
vector itself, vector negation removes not only unwanted
strings but unwanted meanings.
14:40-15:05 Hiromi Nakaiwa (ATR),
"Corpus-Centered Computation - Spoken Language Translation
Research at ATR-SLT"
For translation technology suitable for speech-to-speech
translation (S2S), a new methodology named Corpus-Centered
Computation, (abbreviated to C3 and pronounced "C-cubed") is
introduced. Different from conventional approaches adopted by
machine translation systems for written language C3 places
corpora in the center of technology, for example, translation
knowledge is extracted from corpora, translation quality is
gauged by referring to corpora, while corpora themselves are
normalized by paraphrase or filtered automatically.
15:05-15:20 BREAK
15:20-15:45 Chris Manning (CS, Stanford), Title TBA
15:45-16:10 Takaaki Tanaka (NTT),
"Extending and Applying a Precision Japanese Grammar"
I will describe the results of a recent effort to extend the
coverage of the JACY HPSG-based grammar and subsequently use
it to treebank Japanese dictionary glosses.
16:10-16:35 Timothy Baldwin (CSLI),
"Road-testing the English Resource Grammar over the
British National Corpus"
I will describe the results of the application of the English
Resource Grammar -- a linguistically-precise HPSG-based
grammar -- to data taken from the British National
Corpus. Particular focus is given to error analysis of
sentences the grammar was unable to parse, and the
ramifications for future grammar development.
16:35-17:00 Ichiro Yamada (NHK),
"Meta Data Generation using Closed Captions"
Meta Data is a data for the data. I suggested a method to
generate Meta Data of a TV program using closed captions. This
method extracts some important events from a soccer game for
Meta Data. I will also introduce the research I want to do
while at CSLI.
____________
CS528: BROAD AREA COLLOQUIUM FOR
AI-GEOMETRY-GRAPHICS-VISION-ROBOTICS
on Monday, 29 September 2003, 4:15pm
TCSeq 200
http://graphics.stanford.edu/courses/cs528/
"High-performance imaging using dense camera arrays"
Marc Levoy
Computer Science Department, Stanford
The advent of inexpensive digital image sensors, combined with the
decreasing costs of computing and memory, have led to widespread
interest in building sensing systems that incorporate large numbers of
cameras. To explore this area, the Stanford Computer Graphics
Laboratory has designed a scalable architecture for arrays of
miniature video cameras. Our current implementation contains 128
custom, CMOS-based cameras, each of which measures 30mm on a side. The
system is designed to record 128 synchronized, slightly compressed
video streams through four PCs to a striped disk array.
Multi-camera systems can function in many ways, depending on the
spacing and arrangement of the cameras. If the cameras are packed
closely together, then the system functions as a
single-center-of-projection synthetic camera, which we can configure
to provide high performance along one or more imaging dimensions, such
as spatial resolution, signal-to-noise ratio, dynamic range, depth of
field, frame rate, or spectral sensitivity. If the cameras are placed
farther apart, then the system functions as a
multiple-center-of-projection camera, and the data it captures is
called a (video) light field. If the cameras are placed at an
intermediate spacing, then the system functions as a single camera
with a large synthetic aperture. Such a system has the ability to
focus narrowly on one depth plane, making objects in front or behind
that plane so blurry that they almost disappear. This allows us to see
through partially occluding environments, like foliage or crowds. In
this talk, I will describe the architecture of our system, the
hardware of our current 128-camera implementation, and some of the
applications we are developing for it.
About the Speaker: Marc Levoy is an Associate Professor of Computer
Science and (jointly) Electrical Engineering at Stanford
University. He received a Bachelor's and Master's in Architecture from
Cornell University in 1976 and 1978, and a PhD in Computer Science
from the University of North Carolina at Chapel Hill in 1989. In the
1970's Levoy worked on computer animation, developing an early
computer-assisted cartoon animation system. In the 1980's Levoy worked
on volume rendering, a family of techniques for displaying sampled
three-dimensional functions such as medical scanner data. In the
1990's he worked on technology and algorithms for digitizing
three-dimensional objects. His current interests include sensing and
display technologies, image-based modeling and rendering, and
applications of computer graphics in art history, preservation,
restoration, and archaeology. Levoy received the NSF Presidential
Young Investigator Award in 1991 and the SIGGRAPH Computer Graphics
Achievement Award in 1996 for his work in volume rendering.
____________
SRI AI SEMINAR SERIES
on Tuesday, 30 September 2003, 10:00am
EJ228, SRI International
http://www.ai.sri.com/seminars/
"A Garbage Collection Algorithm and Handling Imperfect Data in a
Monitoring and Optimization System for Coal-Fired Power Plants"
Janet Murdock
During my time at Praxis Engineers (now part of GE), I was primarily
responsible for one module in a product called Coalogic. Coalogic is a
software package that optimizes operation of coal-fired power plants.
In this talk, I'll describe the general solution provided by Coalogic
and describe my own work on two interesting problems, one involving
algorithm development and the other dealing with imperfect plant data.
In operating a coal-fired power plant, an optimization problem arises
because coals vary in physical properties, while electricity demand
varies over the course of a day, from day to day, and over the
seasons. Different coals produce different amounts of heat (per pound
of coal) when burned as well as different amounts of byproducts such
as ash, sulfur dioxide, and various nitrogen oxide compounds. The goal
is to minimize the cost of coal and the cost of processing and/or
disposing of byproducts while meeting the electrical demand and
staying within legal limits for emitted pollutants. Coalogic provides
a three-part solution. First, it augments the plant's on-line data
with a simulation model of coal flows through piles, silos, and
bunkers in the coal yard. This simulation model, known as Coal
Tracking, calculates the instantaneous coal properties for various
locations in the coal yard. (Coal Tracking was my primary
responsibility.) The second part of the solution, Blend Advisor,
calculates the best blend of available coals to burn over the next 8
to 24 hours to meet the expected electricity demand. Blend Advisor
uses current state data from Coal Tracking as well as simulation
models to project how to flow coal from various piles, silos, and
bunkers so that the right blend of coal appears at the boilers at the
right time. The third part of the solution is a boiler operations
optimizer called Optifire. Optifire calculates optimal instantaneous
boiler settings such as air ratios and steam flows. Optifire relies on
Coal Tracking's calculated coal properties for the coal that is
currently entering the boiler. Optifire contains a boiler simulator
that runs a number of different boiler settings to identify optimal
settings. An interesting algorithmic problem that arose while
transitioning the Coal Tracking module from prototype to product was
the need to garbage collect coal receipts. Coal Tracking internally
generates an object (an instance of a C++ class) to hold the coal
properties for each coal delivery. These objects are called coal
receipts and contain properties such as ash and sulfur content for
that particular coal. The coal itself is represented by separate
objects that contain the mass of the coal and a pointer to the
corresponding receipt. The coal from a single delivery gets fragmented
and mixed with coals from other deliveries as it is moved from piles,
through gates, onto conveyors, into hoppers, silos, and bunkers. At
any point in time, thousands of coal objects pointing to the same
receipt may be scattered through the various equipment simulators in
Coal Tracking. After some period of time (ranging from days to months)
all the coal from a particular delivery leaves the coal yard. At this
point the receipt can be garbage collected. A number of tradeoffs and
complications made the use of a simple pointer/reference count scheme
infeasible. Complicating factors included usage of complex simulators
for silos and bunkers. These are large bodies of code that are
difficult and risky to change. Other factors were speed issues, the
need to periodically write out coal yard state to files, and the
impact on customers of any bugs introduced by adding garbage
collection. I'll describe these tradeoffs and the resulting hybrid
garbage collection algorithm. Since Coal Tracking runs 24 hours/day
updating its simulation models every 60 seconds, it is subject to the
inconsistencies and imperfections of real data. The simulation models
have some ability to correct themselves when inconsistencies are
detected. For each power plant, we conducted an engineering analysis
to attempt to discover in advance likely data quality issues and
configure simulation models to handle the inconsistencies. In spite of
these efforts, 50-75% of our customer support effort resulted from
"bad" data. I'll describe three troubleshooting scenarios to
illustrate problems that we solved and problems we left open. One
scenario provided specific insight that led to (at least a partial)
solution for previously known shortcomings. A post-mortem analysis of
problems solved vs. problems not solved reveals a mixture of business
and technical factors driving the decision to fix or not fix.
____________
LOGIC SEMINAR
on Tuesday, 30 September 2003, 4:15pm-5:30pm
Math Corner 380:381T
http://www-logic.stanford.edu/seminars.html
The first meeting of logic seminar in AY 2003/2004
At the beginning we discuss organizational questions. The theme for
this quarter and possibly next is decidable theories. We'll begin
with detailed reviews of new papers on decision procedures for
arithmetical formulas with small number of quantifiers (Shih Peng
Tung) as well as decidable fragments of set theory (Harvey Friedman).
There will be occasional special presentations on other topics by
visitors.
"Propositional logic of continuous transformations"
G.Mints, T. Zhang
Stanford University
Dynamical topological logic studies models of the form (X,T), where X
is a topological space, T a transformation on X. Application in
computer science include hybrid systems, there is a hope to treat
mathematically interesting models. Propositional formulas are
constructed from variables (atomic formulas) by Boolean connectives,
necessity \Box and a monadic operation o. Variables are interpreted by
subsets of X, Boolean connectives act in a natural way, \Box is the
interior and o is the pre-image under operation T. Under this
interpretation the axiom schema o\Box A->\Box oA called (C) expresses
continuity of T. J. Davoren proved completeness of S4+C [including
o(A&B)=oA&oB, o(~A)=~oA] for the class of all topological spaces, in
particular for finite spaces derived from Kripke models.
We prove completeness of S4+C for Cantor space. The proof uses a
continuous and open map W from the Cantor space onto a suitable Kripke
model (K,T) [with T continuous in order topology on K] and a
continuous map S on Cantor space satisfying condition: WS=TW.
____________
SNRC INDUSTRY SEMINAR
on Tuesday, 30 September 2003, 4:15pm
Gates B03
http://snrc.stanford.edu/events/industry-seminar/
"Resource Management in Utility Computing Environments"
Akhil Sahai
Hewlett Packard Laboratories, Palo Alto
Currently, there is a trend towards managing data center
infrastructure and services within a utility model that provides
resources on demand. HP's Utility Data Center product, IBM's
"on-demand" computing initiative , Sun's N1 vision, and Microsoft's
DSI initiative are examples of this trend. In addition, much of the
work in the Global Grid Forum is targeted at developing a web-services
based infrastructure that can offer resources (e.g. servers, network,
software, storage etc.) to users when needed. The intent in these
initiatives is to create systems that provide automated provisioning,
configuration, and lifecycle management of a wide variety of
infrastructure resources. These environments are heterogeneous,
dynamic and plug-n-play in nature. Resource management and
virtualization play an important role in these utility computing
environments. The talk will describe various issues and their
solutions that are involved in managing large set of heterogenerous
resources in such an environment.
About the Speaker: Dr. Akhil Sahai is a senior scientist at HP
Laboratories, Palo alto. He is currently researching policy
management in utility computing environments. He has been researching
service management, solutions and architectures and was one of the
initial members of the e-speak team that shaped Hewlett Packard's
pioneering web services technology. He has led research earlier at
Multi-Media Systems group at Kent Ridge Digital Labs (earlier ISS)
Singapore. He has about 8 years of industrial experience. He has a
ph.D is computer science from INRIA-IRISA France and Master's degree
in computer science from Indian Institute of Science Bangalore. He has
published widely in distributed systems, network/system/service
management and mobile computing.
____________
EE380: COMPUTER SYSTEMS LAB COLLOQUIUM
on Wednesday, 1 October 2003, 4:15pm to 5:30pm
NEC Auditorium (B03), Gates Computer Science Building
http://ee380.stanford.edu/contents.html
"DNS Typosquatting: Methods, Effects, and Counteraction
What Verisign's Sitefinder Means for the Internet Domain System"
Paul Vixie
Internet Software Consortium
On September 15 2003, VeriSign added a DNS wildcard to the COM and NET
zones. On September 17, ISC released a patch to BIND that allows our
users to filter out the effects of these wildcards. In this
extemporaneous talk, Paul Vixie of ISC will explain what happened,
who's upset about it, and why.
About the speaker: Paul Vixie was the last maintainer of BIND4, the
main author of BIND8, and an architect of BIND9. After cofounding MAPS
(an anti-spam company) and having some some excellent adventures in
DotBombLand, he returned to ISC in May 2002 as its full time
president. Paul holds the record for "most CERT advisories due to a
single author", and no longer writes code that other people will see.
____________
SRI AI SEMINAR SERIES
on Thursday, 2 October 2003, 4:00pm - 5:30pm
Enterprise Room, SRI International
http://www.ai.sri.com/seminars/
"Adventures in Artificial Intelligence"
Nils Nilsson
Robotics Lab., Computer Science, Stanford University
I'll reveal details about my three most exciting scientific-technical
amorous adventures over the last forty years. Minos, a neural network
at SRI, was my first date. Shakey the Robot was a more serious affair,
and many prominent progeny resulted. My current infatuation is with
Teleo-Reactive programs. This talk is a repeat of my lecture at the
2003 International Joint Conference on Artificial Intelligence---given
on the occasion of my receiving the IJCAI "Research Excellence award.
About the Speaker: Nils J. Nilsson, Kumagai Professor of Engineering
(Emeritus) in the Department of Computer Science at Stanford
University, received his PhD degree in Electrical Engineering from
Stanford in 1958. He spent twenty-three years at the Artificial
Intelligence Center of SRI International working on statistical and
neural-network approaches to pattern recognition, co-inventing the A*
heuristic search algorithm and the STRIPS automatic planning system,
directing work on the integrated mobile robot, SHAKEY, and
collaborating in the development of the PROSPECTOR expert system. He
has published five textbooks on artificial intelligence. Professor
Nilsson returned to Stanford in 1985 as the Chairman of the Department
of Computer Science, a position he held until August 1990. Besides
teaching courses on artificial intelligence and on machine learning,
he has conducted research on flexible robots that are able to react to
dynamic worlds, plan courses of action, and learn from experience.
____________
UC BERKELEY CIS SEMINAR
on Thursday, 2 October 2003, 4:00pm-5:00pm
Soda Hall 310 (UC Berkeley)
http://www.eecs.berkeley.edu/~wainwrig/cis-seminar
"The difficulty of making a robot go straight"
Sebastian Thrun
Computer Science, Stanford University
On May 30, 2003, a team of researchers sent an autonomous mobile robot
into an abandoned coal mine near Pittsburgh, PA. The robot's task was
simply to acquire a 3-D map of a mile-long supply corridor of the
mine, autonomously, outside the communication range of the control
station. This talk will report our experiences with this and a number
of related subterranean expeditions. The speaker will discuss Gaussian
Markov random field techniques for robotic mapping and exploration,
presenting algorithms and theorems. He will show how these techniques
work in subterranean worlds, where robots face all sorts of
challenges.
____________
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
____________