[Date Prev][Date Next][Date Index]
CSLI Calendar, Wednesday, 4 June 2003, vol. 18:36
CSLI CALENDAR OF PUBLIC EVENTS
______________________________________________________________________
4 June 2003 Stanford Vol. 18, No. 36
______________________________________________________________________
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 JUNE 2003 TO 13 JUNE 2003
WEDNESDAY, 4 JUNE 2003
12 noon Psychology Developmental Brownbags
Bldg. 380:381U
"Isn't it ironic?: Young children's understanding of
situational and verbal irony in similar story contexts"
Jennifer Dyer
UC Santa Cruz
http://www-psych.stanford.edu/news.html#dev_brownbag
1:00pm Special University Oral Examination
Durand 450
"New Algorithms, Data Structures, and User Interfaces for
Machine Learning of Large Datasets with Applications"
Jorge Moraleda
Aerospace Robotics Laboratory, Stanford
Abstract below
2:00pm EE144/245 Wireless Lab Student Projects Demonstrations/Posters
SEQ Plaza between Packard and SEQ Teaching Center
till 4:30pm
3:00pm Reuters Foundation Digital Vision Program Seminar
email for location mailto:digitalvision@csli.stanford.edu
Jerry Michalski
founder and president, Sociate
http://www.sociate.com/
http://reuters.stanford.edu/seminar_speakers.html
3:45pm Psychology Department Colloquium
Jordan Hall 420:041
"Is There Nothing So Practical as a Good Theory?: Lewin's
Exhortation and the Study of Health Behavior Change"
Alex Rothman
University of Minnesota
http://www-psych.stanford.edu/news.html
4:15pm EE380: Computer Systems Lab Colloquium
Gates B03 (NEC Auditorium)
"Computers versus Common Sense: An Engineering Approach to AI"
Doug Lenat
Cycorp
http://ee380.stanford.edu/contents.html
Abstract below
THURSDAY, 5 JUNE 2003
11:00am Music 319: CCRMA Hearing Seminar
CCRMA Library, The Knoll
"Effects of the medial olivocochlear reflex on cochlear mechanics"
Duck Kim
University of Connecticut
http://mambo.ucsc.edu/psl/ccrmas/ccrmas.html
Abstract below
4:00pm PARC Forum
George Pake Auditorium at PARC
"The Evolution of the Networking Industry"
Mike Volpi
Cisco Systems
http://www.parc.com/forum/
Abstract below
4:15pm CSLI Seminar on Computational Learning and Adaptation
Cordura Hall, room 100
"Boosting as a Metaphor for Algorithm Design"
Eugene Nudelman
Stanford
http://cll.stanford.edu/scla.html
Abstract below
4:15pm SSP10: Symbolic Systems Forum
Bldg. 380:380C (Math Corner)
Presentation of Senior Honors Theses
Symbolic Systems Senior Honors Candidates
http://symsys.stanford.edu:8081/ssp-dynamic/servlet/ssp_events
Information below
5:30pm Stanford Phonology and Syntax Workshops
Margaret Jacks Hall 460:126
"Syntactic Asymmetries in Phonological Domain Formation:
Phrases, Compounds, Derivatives"
Michael Wagner
MIT/Stanford
http://www-linguistics.stanford.edu/pinterest/
FRIDAY, 6 JUNE 2003
2:15pm NLP Reading Group
Wallenberg 323
"A* Parsing: Fast Exact Viterbi Parse Selection"
Dan Klein
Stanford University
http://www.stanford.edu/~danklein/papers/pcfg-astar.pdf
http://www-nlp.stanford.edu/nlp/nlpgroup.html
Abstract below
3:15pm Friday Cognitive Seminar
Jordan Hall 420:100
Title to be announced
Nicolas Davidenko
http://www-psych.stanford.edu/news.html#frisem
MONDAY, 9 JUNE 2003
3:30pm Berkeley Intel Research Seminar
Intel Research Berkeley, 2150 Shattuck, Ste. 1300
"Tech-Based ARt -
as Seen from the Trenches and as Seen from the Sky"
Michael Naimark
Independent Media Artist and Researcher
http://buffy.eecs.berkeley.edu/Seminars/
TUESDAY, 10 JUNE 2003
7:00pm Emerging Technology Group
Cubberley Community Center, H-1, 4000 Middlefield, Palo Alto
"The Past, Present, and Future of Instant Messaging"
Ashvil D'Costa
i3Connect
http://www.sdforum.org/
(there is a fee for non-SDForum members, see web page)
Abstract below
WEDNESDAY, 11 JUNE 2003
3:00pm Reuters Foundation Digital Vision Program Seminar
email for location mailto:digitalvision@csli.stanford.edu
Stephen R. Du Mont
Cisco
http://reuters.stanford.edu/seminar_speakers.html
THURSDAY, 12 JUNE 2003
12 noon RNI/Stanford Seminar on Theoretical Neuroscience
CISX 101, Allen Center for Integrated Systems (NOTE LOCATION)
"The Dynamical Brain"
Terrence Sejnowski
Salk Institute
http://www.rni.org/seminar.html
Abstract below
4:00pm PARC Forum
George Pake Auditorium at PARC
"Enabling Productivity with Large Displays"
Gary Starkweather
Microsoft Research
http://www.parc.com/forum/
4:15pm CSLI Seminar on Computational Learning and Adaptation
Cordura Hall, room 100
"Learning DFA from Simple Examples"
Rajesh Parekh
Blue Martini Software
http://cll.stanford.edu/scla.html
Abstract below
____________
Stanford Blood Center status: shortage of O-, O+, A-, B+, and AB-.
For an appointment: http://bloodcenter.stanford.edu/ or call
650-723-7831. It only takes an hour of your time.
____________
WORKSHOP ANNOUNCEMENT
"Barwise and Situation Semantics"
Stanford, California
Thursday, 26 June 2003
A workshop co-located with CONTEXT '03
http://www.cs.tcd.ie/Tim.Fernando/sa-b.html
The late Jon Barwise was, among many other things, the first director
of the Center for the Study of Language and Information at Stanford.
His book with John Perry, Situations and Attitudes, appeared some
twenty years ago (in 1983), launching situation semantics, a semantic
framework that analyzes context in terms of situations.
This workshop provides a forum for works addressing the following
question: What problems, issues and/or insights connected with
situation semantics and Barwise motivate your research today? And how?
A special issue of Research on Language and Computation and/or Journal
of Applied Logic is projected, based on the workshop.
More information is available on the web site:
http://www.cs.tcd.ie/Tim.Fernando/sa-b.html
____________
ANNOUNCEMENT
Internet Law Program, Summer 2003
Stanford Law School
30 June 2003 - 4 July 2003
http://cyberlaw.stanford.edu/ilaw2003
Registration is now open for this summer's Internet Law Program (ILaw)
in Stanford, CA, sponsored by the Stanford Law School Center for
Internet and Society and the Berkman Center for Internet & Society at
Harvard Law School. On the agenda: recent reforms in
intellectual-property systems, privacy versus security on the Net, the
changing shape and role of ICANN, "open" versus "proprietary" software
systems, regulating pornography, jurisdictional problems, cybercrime,
addressing the digital divide, and more. No previous experience with
Internet law is necessary to enroll. The program is designed for
lawyers, policymakers, business and technology professionals,
government and non-profit executives, and journalists who write about
technology. International participation is encouraged.
Questions? Please contact Harvard's Robyn Mintz at:
rmintz@cyber.law.harvard.edu
____________
SPECIAL UNIVERSITY ORAL EXAMINATION
on Wednesday, 4 June 2003, 1:00pm
Durand 450
"New Algorithms, Data Structures, and User Interfaces for
Machine Learning of Large Datasets with Applications"
Jorge Moraleda
Aerospace Robotics Laboratory, Stanford
Predictive modeling, the use of prior data to make decisions or
predict performance and future trends, is useful in many domains:
Biology: gene and protein function, binding sites, interactions, pathways
Medicine: diagnosis, drug effects and interactions
Finance: stock market, credit risk
Marketing: customer response
Manufacturing: quality control
Industrial Plants: fault detection and diagnoses
Many good paradigms for data analysis can be used for predictive
modeling, such as Bayesian networks, neural networks, kernel methods,
Markov models, etc. Existing algorithms and data structures based on
these paradigms work well for datasets of reasonable size. However,
because of recent advances in data gathering technology
(e.g. automated data collection systems, inexpensive storage space),
much larger amounts of data are being collected. Existing methods need
to be extended in order to analyze and display the information
contained in these larger datasets.
In this talk I will present a new system capable of learning
predictive models from large datasets and displaying and interacting
with these large models in a user friendly manner. The new system has
been developed within the Bayesian networks framework because of their
transparency, their ability to deal with uncertain domains, and the
fact that their model structure reflects the underlying structure of
the domain. These characteristics make them especially well suited to
deal with many real world problems.
Learning Bayesian network structure from data is NP-hard, and thus
heuristic search is necessary to find good models. I will present two
methods of improving this search: the AD+Tree, a data structure that
caches counts from the dataset very efficiently, enabling very fast
evaluation of different models; and Queue Learning, a new algorithm
for learning Bayesian network structure that produces better models
early in the search than existing techniques. I will also present a
new user interface that enables easy interaction with these larger
models.
____________
EE380: COMPUTER SYSTEMS LAB COLLOQUIUM
on Wednesday, 4 June 2003, 4:15pm to 5:30pm
NEC Auditorium (B03), Gates Computer Science Building
http://ee380.stanford.edu/contents.html
"Computers versus Common Sense
An Engineering Approach to AI"
Doug Lenat
Cycorp, Inc.
Computers today are idiot-savants. They may manage bits flawlessly and
furiously, but they have no understanding of what those bits signify.
And they have poor models of themselves and of the human beings they
serve and represent. To break that "brittleness bottleneck," we need a
new software layer that contains the millions of things the average
person knows about the world. Some of this is factual, such as how
often a U.S. Presidential election is held, or even ephemeral, such as
the name of the current President; but most of the needed content is
more like rules of thumb, such as why one should carry a glass of
water open-end up. In terms of a newspaper or book, we are talking
about codifying the white space - the things the authors don't need to
bother saying (e.g., the White House is in Washington, D.C.; tables
have flat horizontal tops; appliances stop working when turned off.)
Since 1984, my team has spent the seven person-centuries necessary to
build that artifact. In this talk, I'll describe what we did, and why,
and some of the lessons we learned about representing commonsense
knowledge, and doing reasoning in huge knowledge-based systems. In
particular, I'll explain why we took an empirical, engineering
approach to the problem, rather than a theoretical, scientific
approach. I'll also discuss some current and future commercial
applications of our technology (CYC).
About the speaker: Douglas B. Lenat received hi Ph.D. in Computer
Science at Stanford in 1976; his thesis was a heuristic program called
AM that made hundreds of small creative discoveries in mathematics --
a theorem proposer, rather than a theorem prover -- for which he was
awarded the biannual IJCAI Computers & Thought Award in 1977.
Dr. Lenat was named one of the original Fellows of the AAAI (American
Association for Artificial Intelligence). A prolific author, he has
been a professor at CMU and Stanford, a founder of Teknowledge, and
the only individual ever to serve on the technical advisory boards of
both Apple and Microsoft. His interest and experience in national
security has led him to regularly consult for several U.S. agencies
and the White House. From 1984 through 1994, Dr. Lenat directed the
Cyc common sense knowledge base and reasoning project for MCC, the
USA's first high-technology research consortium. He is the President
of Cycorp, the company he founded in 1994 to carry on the development
and commercialization of the Cyc technology.
____________
MUSIC 319: CCRMA HEARING SEMINAR
on Thursday, 5 June 2003, 11:00am
CCRMA Library, The Knoll
http://mambo.ucsc.edu/psl/ccrmas/ccrmas.html
Sunil Puria and I are very happy to introduce Duck Kim to the CCRMA
Hearing Seminar community. Duck is a well-known researcher studying
active and non-linear processes in the cochlea. Neely and Kim (1983)
introduced us to the first active cochlear model (about two years
before the discovery of outer hair cell motility by Brownell.)
We've talked about many pieces of cochlear puzzle over the years. The
auditory system has exquisite sensitivity and a very large dynamic
range. This sensitivity and the dynamic range of the system are
controlled by a feedback system that includes the inner hair cells
(detectors), up into the brain, and then back down through the outer
hair cells. It's pretty hard to tear this system apart and understand
all of its behavior.
One approach is based on measuring the acoustic distortion products
produced by a healthy ear. Put two tones into a non-linear system and
new tones are produced at sums and differences of the original
frequencies. The otoacoustic emissions from our own ears are
generally not heard, but they can tell us something about the
nonlinearities in the cochlea. Their time course tells us how the
feedback loop is affected by the longer pathways up through the brain.
Duck Kim, a distinguished scientist from U of Connecticut and a good
speaker, will be talking about some of his recent modeling efforts.
How do the distortion products change with time and level, and what is
changing that causes the measured emissions? What is the purpose of
controlling the outer hair cells via the medial efferents? Come to
CCRMA to find out more. This is a good opportunity to hear one of the
principle researchers in the field.
Bring your own otoacoustic emissions and we'll talk about where they
come from at the next CCRMA Hearing Seminar.
- Malcolm Slaney
"Effects of the medial olivocochlear reflex on cochlear mechanics:
experimental and modeling studies of DPOAE"
Duck O. Kim
Neuroscience, U. Connecticut Health Center, Farmington, CT
http://penguin.uchc.edu/~kimdolab/
Time-dependent changes (or adaptation) in level of distortion product
otoacoustic emission (DPOAE) are partly effects of ipsilateral reflex
of the medial olivocochlear (MOC) system. This study presents
experimental data of DPOAE in cats and modeling results, and addresses
the following questions. How does the MOC reflex change DPOAE? Why is
large adaptation of DPOAE associated with a trough of a DPOAE
input/output (I/O) function? Why does DPOAE level either increase or
decrease with time depending upon the stimulus condition? We
hypothesize that the fast component of DPOAE adaptation is produced by
a shift in the operating point of the I/O nonlinearity (Boltzmann
function) of the outer hair cell (OHC) hair bundle, which in turn is
produced by the MOC reflex through changes in the mean potential and
length of the OHC and a rotation of the OHC hair bundle. The ability
of the model to reproduce salient behaviors of DPOAE supports the
hypothesis.
____________
PARC FORUM
on Thursday, 5 June 2003, 4:00pm - 5:00pm
George Pake Auditorium, PARC
http://www.parc.com/forum/
"The Evolution of the Networking Industry"
Mike Volpi
Senior Vice President/General Manager,
Routing Technology Group, Cisco Systems, Inc.
There are some fundamentals in the networking industry that have
evolved over the last decade, as it has become a very different place
to compete. For example, financial strength provides a significant
advantage in the ability to out innovate the competition on a regular
and consistent basis. To enter the networking industry today, a large
initial investment is required. In order to survive, the investment
has to be spread across volumes. Only those with scale distribution
and volume sales will break away and succeed.
In this presentation, Mike Volpi will discuss the evolution of the
networking industry and how innovation is directly affected by these
changes.
About the Speaker: Mike Volpi is the Senior Vice President and General
Manager for Cisco Systems' Routing Technology Group. In this
capacity, he is responsible for all of Cisco's routing and access
products including the core, midrange, and access routers, content
networking mobile wireless, and cable and video solutions.
Previously, Volpi was Senior Vice President of the Internet Switching
and Services Group where he was responsible for developing switching
products for data centers and distribution applications, as well as
creating a services infrastructure for security and content
networking.
Prior to his role as head of the Internet Switching and Services
Group, Volpi was the Chief Strategy Officer. As CSO, Volpi was
responsible for Cisco's corporate strategy as well as business
development, strategic alliances, advanced Internet projects, legal
services, and government affairs. Volpi directed Cisco's expansion
into new markets and technologies through a balance of internal
development, partnerships and acquisitions.
Cisco acquired more than 70 companies during his tenure, and developed
Cisco's acquisition and integration processes, which have been a
significant driver of Cisco growth into new markets.
____________
CSLI SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION
on Thursday, 5 June 2003, 4:15pm to 5:30pm
Cordura 100
http://cll.stanford.edu/scla.html
"Boosting as a Metaphor for Algorithm Design"
Eugene Nudelman
Computer Science Department, Stanford University
It is well known that NP-Hard problems can often be solved easily in
practice, and algorithms exhibit highly different behaviour when
presented with different instances. There has been a lot of work
recently in trying to understand the empirical complexity of hard
problems. Our approach has been to use machine learning to construct
statistical models of hardness. I'll present the application of this
approach to understanding the hardness of the winner determination
problem (WDP) for combinatorial auctions.
I'll also present two applications of this methodology. One is to
construct algorithm portfolios that outperform their constituent
algorithms. Our case studies indicate that this leads to improvements
in our ability to solve WDP and SAT instances. I'll describe a way to
use the hardness models to induce harder (or easier) problem instance
distributions.
This is joint work with Kevin Leyton-Brown, Galend Andrew, Jim
McFadden, and Yoav Shoham, also of the CS department.
____________
SSP10: SYMBOLIC SYSTEMS FORUM
on Thursday, 5 June 2003, 4:15pm
Bldg. 380:380C (Math Corner)
http://symsys.stanford.edu:8081/ssp-dynamic/servlet/ssp_events
Presentations of Senior Honors Projects
SSP Honors Students, Class of '03
4:15pm Micah Boster,
"Market Liberalization and Global Cellular Phone Propagation"
(Advisor: David Abernethy, Political Science)
4:25pm Rory Berry,
"The Effects of Multiple Instructors and Presentation Modality
on Learning, Perceived Learning, and User Satisfaction in E-Learning"
(Advisor: Cliff Nass, Communication)
4:35pm Jed Rose,
"VSG's and the Deaf: A novel visual symbol game to help the
deaf learn to read complex English sentences"
(Advisor: Daniel Schwartz, Education)
4:45pm Ash Brown,
"Persuasive CALL: Using Persuasion to Make Computer Assisted
Language Learning More Effective"
(Advisor: B.J. Fogg, Computer Science)
4:55pm Kiely Martinez,
"Animal play: biological and philosophical perspectives"
(Advisor: Stuart Thompson, Biological Sciences)
5:05pm Bayle Shanks,
"Collaborative encyclopaedias covering neuroscience and A.I research"
(Advisor: John Gabrieli, Psychology)
5:15pm Sara Wampler,
"A Network Analysis of the 20 July 1944 Conspiracy"
(Advisor: Elizabeth Bernhardt, German Studies)
5:25pm Alexis Battle,
"The Probabilistic Discovery of Overlapping Gene Modules and
Their Regulation"
(Advisor: Daphne Koller, Computer Science)
5:35pm Hilary Spencer,
"Visualization and Graph-Drawing Techniques for
Protein-Protein Interaction Networks"
(Advisor: Patrick Hanrahan, Computer Science)
Refreshments will be served afterward.
____________
NLP READING GROUP
on Friday, 6 June 2003, 2:15pm - 3:30pm
Wallenberg 323
http://www-nlp.stanford.edu/nlp/nlpgroup.html
"A* Parsing: Fast Exact Viterbi Parse Selection"
Dan Klein
Stanford University
http://www.stanford.edu/~danklein/papers/pcfg-astar.pdf
We present an extension of the classic A* search procedure to tabular
PCFG parsing. The use of A* search can dramatically reduce the time
required to find a best parse by conservatively estimating the
probabilities of parse completions. We discuss various estimates and
give efficient algorithms for computing them. On average-length Penn
treebank sentences, our most detailed estimate reduces the total
number of edges processed to less than 3% of that required by
exhaustive parsing, and a simpler estimate, which requires less than a
minute of pre-computation, reduces the work to less than 5%. Unlike
best-first and finite-beam methods for achieving this kind of
speed-up, an A* method is guaranteed to find the most likely parse,
not just an approximation. Our parser, which is simpler to implement
than an upward-propagating best-first parser, is correct for a wide
range of parser control strategies and maintains worst-case cubic
time.
____________
EMERGING TECHNOLOGY GROUP
on Tuesday, 10 June 2003, 7:00pm
Cubberley Community Center, H-1, 4000 Middlefield, Palo Alto
http://www.sdforum.org/
(there is a fee for non-SDForum members, see web page)
"The Past, Present, and Future of Instant Messaging"
Ashvil D'Costa
i3Connect
We start with evaluating communication as critical to business. Then
establish the power that presence information adds to the value of
communication, establish the value for IM in business. Then we talk
about various other applications of IM in business infrastructure and
get down to the details and talk about the state of the industry and
standards (open source etc.). And then take the discussion to the
future and trends that we see taking shape.
About the Speaker: Ashvil D'Costa is the president and founder of
i3Connect and has 11 years of experience in developing quality
software products and providing technology leadership with companies
such as Think Systems (now i2), ADP, Intel and Wells Fargo. Ashvil
has driven the vision of i3Connect since its inception.
____________
RNI/STANFORD SEMINAR SERIES IN THEORETICAL NEUROSCIENCE
on Thursday, 12 June 2003, 12 noon - 1pm
CISX 101, 330 Serra Mall
http://www.rni.org/seminar.html
"The Dynamical Brain"
Terrence Sejnowski
Salk Institute
http://www.salk.edu/faculty/faculty/details.php?id=48
Even in the absence of sensory stimuli, neurons in the cortex are
spontaneously active and can fire coherently. Experiments and models
suggest that these coherent patterns are modulated by top-down
influences that allow us to expect, attend and flexibly respond.
About the Speaker: Terrence J. Sejnowski, professor and head of the
Computational Neurobiology Laboratory at the Salk Institute, is a
pioneer in the field of computational neuroscience. He has a B.S. in
physics from Case Western and a Ph.D. in physics from Princeton
University. He received the Presidential Young Investigator Award for
1984-89, Fairchild Distinguished Scholar award in 1992-93, and Wright
Prize in 1996. Among other things, Sejnowski is interested in the
hippocampus, believed to play a major role in learning and memory; and
the cerebral cortex, which holds our knowledge of the world and how to
interact with it.
____________
CSLI SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION
on Thursday, 12 June 2003, 4:15pm to 5:30pm
Cordura 100
http://cll.stanford.edu/scla.html
"Learning DFA from Simple Examples"
Rajesh Parekh
Business Intelligence Group
Blue Martini Software
Efficient learning of DFA is a challenging research problem in
grammatical inference. It is known that both exact and approximate (in
the PAC sense) identifiability of DFA is hard. Pitt has posed the
following open research problem: Are DFA PAC-identifiable if examples
are drawn from the uniform distribution, or some other known simple
distribution?. I will discuss the learnability under the
Solomonoff-Levin universal distribution (m) of the class of DFA whose
canonical representations have logarithmic Kolmogorov complexity. I
will show that the class of DFA is efficiently learnable under the
PACS (PAC learning with simple examples) model wherein positive and
negative examples are sampled according to the universal distribution
conditional on a description of the target concept. Further, I will
address the relationship between models for learning in helpful
environments and explain how unnatural collusion between the teacher
and learner can potentially trivialize the task of learning in such
environments.
This is joint work with Vasant Honavar, Department of Computer
Science, Iowa State University.
____________
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
____________