[Date Prev][Date Next][Date Index]
CSLI Calendar, 13 October 1999, vol. 15:4
C S L I C A L E N D A R O F P U B L I C E V E N T S
______________________________________________________________________
13 October 1999 Stanford Vol. 15, No. 4
______________________________________________________________________
A weekly publication of the
Center for the Study of Language and Information (CSLI)
Stanford University, Ventura Hall, Stanford, CA 94305-4115
____________
ACTIVITIES FROM 13 OCTOBER TO 22 OCTOBER 1999
WEDNESDAY, 13 OCTOBER
12 noon Stanford Computer Industry Project (SCIP)
Littlefield room L107
Startup to Speedup: From 0 to $1 Billion in 18 Months
Scott Dunlap
Director of Ecommerce, E.piphany
http://www.stanford.edu/group/scip/
12:30pm International Computer Science Institute
ICSI room 607, 1947 Center Street, Berkeley
Coherence, Reference, and the Theory of Grammar
Andrew Kehler
SRI International
http://www.icsi.berkeley.edu/talks/
4:15pm EE380: Computer Systems Laboratory Colloquium
Gates B03 (NEC Auditorium)
A Dialog with Brian Eno
Brian Eno
http://www.stanford.edu/class/ee380/contents.html
Abstract below
4:15pm CS528: Broad Area Colloquium
Gates B12
Modular Reconfigurable Robotics
Mark Yim
Xerox PARC
http://robotics.stanford.edu/ba-colloquium/
Abstract below
THURSDAY, 14 OCTOBER
11:00am CCRMA Hearing Seminar
CCRMA Library
Dichotic Pitch
Bob Dougherty
Psychology
http://www-ccrma.stanford.edu/CCRMA/Events/Events.html#hearing
Abstract below
3:15pm Experiments in Learning at Stanford
Press Warehouse, room 118
Engineering Student Learning:
Research in 'Design' and 'Knowledge Integration'
Cindy Atman
CELT
http://sll.stanford.edu/
Abstract below
4:00pm Semantics Workshop
Margaret Jacks 126
first meeting
http://www-linguistics.stanford.edu/semgroup/
4:00pm Xerox PARC Forum
George Pake Auditorium, Xerox PARC
Breaking In To Computer Systems:
Why It's Possible, How It's Done
Olin Sibert
InterTrust Technologies Corporation
http://www.parc.xerox.com/ops/projects/forum
Abstract below
4:15pm US-Japan Technology Management Center
Skilling Engineering Auditorium
The Asia Pacific Advanced Network (APAN), Its Activities
and Impact
Dr. Kilnam Chon
Chairman, Asia Pacific Advanced Network & Professor,
Department of Computer Science Korea Advanced Institute
of Science and Technology (KAIST)
http://www.stanford.edu/~viji/
Abstract below
FRIDAY, 15 OCTOBER
12 noon Logic Lunch
Room 380:383N
Recursive types with a modality
Hiroshi Nakano
Ryukoku University, visiting Stanford
http://www-logic.stanford.edu/seminars.html
Abstract below
12:30pm Seminar on People, Computers, and Design
Gates B03 (NEC classroom)
The Importance of Homes in Technology Research
Debby Hindus
http://www-pcd.stanford.edu/cs547/
Abstract below
3:15pm CS545: Infolab Seminar
188 tcSEQ (across from Gates)
Lore: A Database Management System for XML
Jennifer Widom
http://www-db.stanford.edu/dbseminar/dbseminar.html
Abstract below
MONDAY, 18 OCTOBER
7:00pm Stanford Presidential Lectures
Roble Dance Studio
Pina Bausch
Choreographer
http://prelectur.stanford.edu/calendar/
TUESDAY, 19 OCTOBER
11:00am CCRMA Hearing Seminar
CCRMA Library
Motion and music, templates and tunes
Willy Wong
http://www-ccrma.stanford.edu/CCRMA/Events/Events.html#hearing
4:00pm International Computer Science Institute
ICSI Main Lecture Hall, 1947 Center Street, Berkeley
Resource Allocation in Networks using Constraints and
Agents
Boi Faltings
Swiss Federal Institute of Technology (EPFL)
http://www.icsi.berkeley.edu/talks/
See CDR talk on the 20 October for abstract
4:15pm Seminar on Computational Learning and Adaptation (SCLA)
Gates 104
A Bayesian Framework for Concept Learning
Josh Tenenbaum
Psychology, Stanford University
http://www-csli.stanford.edu/cll/scla.html
Abstract below
WEDNESDAY, 20 OCTOBER
10:00am Knowledge on the Web Seminar
Gates 104
To be announced
Ramanathan Guha
Chief Technical Officer, Epinion (Co-developer of CYC)
http://www.ksl.stanford.edu/KnOWS.html
12 noon Stanford Computer Industry Project (SCIP)
Littlefield room L107
Reliability of Internet Panels
Kirthi Kalyanam
Santa Clara University
http://www.stanford.edu/group/scip/
12 noon CDR Talk
Gates 498
Resource Allocation in Networks using Constraints and
Agents
Boi Faltings
Director, AI Lab
Swiss Federal Institute of Technology (EPFL)
Abstract below
11:00am International Computer Science Institute
ICSI Main Lecture Hall, 1947 Center Street, Berkeley
Semantic Lexicon Acquisition for Learning Natural
Language Interfaces
Cynthia Thompson
CSLI, Stanford University
Abstract below
http://www.icsi.berkeley.edu/talks/
4:15pm EE380: Computer Systems Laboratory Colloquium
Gates B03 (NEC Auditorium)
To be announced
http://www.stanford.edu/class/ee380/contents.html
4:15pm CS528: Broad Area Colloquium
Gates B12
Taming the Giants and The Monsters:
Recent Developments in Data Mining
Usama Fayyad
Microsoft Corp.
http://robotics.stanford.edu/ba-colloquium/
Abstract below
THURSDAY, 21 OCTOBER
12 noon Award-Winning Teachers on Teaching
Hartley Conference Room, Mitchell Earth Sciences
Teaching Large Humanities Courses for the Frosh
Philippe Buc
History
http://www-ctl.stanford.edu/teach/awt/awtmain.html
7:30pm Stanford Phonology Workshop
Building 460, Room 126
Title TBA
Edward Flemming
Department of Linguistics, Stanford University
http://www-linguistics.stanford.edu/Linguistics/pinterest/
FRIDAY, 22 OCTOBER
12:30pm Seminar on People, Computers, and Design
Gates B03 (NEC classroom)
The Soft Machine - Design in the Cyborg Age
Marie O'Mahony
http://www-pcd.stanford.edu/cs547/
Abstract below
3:15pm CS545: Infolab Seminar
188 tcSEQ (across from Gates)
To be announced
Vishal Sikka
PatternRx
http://www-db.stanford.edu/dbseminar/dbseminar.html
4:00pm Semantics Workshop
Margaret Jacks 126
The Metaphysics of Words
Nicholas Asher
(University of Texas, Austin)
Local Prepositions Revisited
Jean Mark Gawron
(SRI International)
http://www-linguistics.stanford.edu/semgroup/
____________
EE380: COMPUTER SYSTEMS COLLOQUIUM
on Wednesday, 13 October 1999, 4:15pm to 5:30pm
NEC Auditorium (B03), Gates Computer Science Building
http://www.stanford.edu/class/ee380/contents.html
A Dialog with Brian Eno
Brian Eno
Artist
[edited from the e-mail exchange arranging the talk]
BRIAN:
I would be interested to do something with you, and now I'm wondering
what it might be. Strangely, it's always much harder when someone
generously says "talk about anything you like" - because you then have
the responsibility of wondering what on earth that might be. I must
add also that I'm a little timid about talking to a very specialized
and, I'm sure, highly intelligent group about anything even
approaching their own discipline. So, though I work with computers a
lot and have strong and occasionally iconoclastic feelings about them,
I can't really imagine there's anything much I could say on the
subject that would be news to such an audience.
At the moment I'm in a making rather than a theorizing phase - which
is to say I don't have a new, current talk in my mind to be adapted to
your circumstances. It might turn out that this is therefore not the
most ideal time for me to do this....but perhaps you have an idea. I'm
a good improviser.
DENNIS:
Our students will soon be out there trying to create the future (in
the sense that Stewart Brand wrote about the MIT Media Lab folks). We
tend to be a nerdy, technologically-oriented lot with relatively
little understanding of art and design. I read and learned from your
book, A Year with Swollen Appendices. In your narrative it was not the
"theorizing" that was captivating, but the laying out of the "doing"
with occasional reflections about theory and society. Perhaps a
similar approach could be the framework for a talk.
You might approach this audience by explaining a bit about who you are
and what you are trying to accomplish artistically and
intellectually. You certainly should talk about your relationship with
computers, good and bad. I would expect your iconoclastic feelings to
show and to see computers relegated to their proper status--tools--and
not elevated to any god-like status or role.
Personally, I would like to hear you talk about how you see the
process of "doing". How will art and design fare in an increasingly
automated and computer based world? An informal, improvised, off-hand
talk along these lines would be completely appropriate.
BRIAN:
My talk won't be a very highly structured talk, but in the past I've
done OK by the seat of my pants, so I don't mind trying again. Hope
those pants can still take it...
I've found that a dialogue with the audience often produces more
surprises than the talk. I'm not doing the lazy-speakers trick of
saying "Let's just have questions", but I'd like have think of a way
that the questions can start earlier, or be more integrated.
I did a talk in Paris a couple of months ago where I invited people
(before the talk started) to submit questions on paper and then used
those as the basis for my talk. I had them laid out before me and just
used them as a skeleton around which i could build the flesh of the
talk. The good part was that I could just ignore the questions that
weren't pointing anywhere interesting. It was a successful format -
in that it gave me the reassurance that at least ONE person wanted to
know what I was talking about, and that it led the talk into some
interesting tangents which I probably wouldn't have gone down
otherwise.
About the speaker: Brian Eno is an artist, composer, musician, and
author. He was born in 1948 in Woodbridge, Suffolk. His father was a
postman; his mother a Belgian immigrant. He grew up in Woodbridge
where much of his extended family still resides. He is the product of
a Catholic grammar school and the Ipswitch Art School. He holds a
diploma in Fine Art from the Winchester School of Art.
In 1969 he moved to London, joined Roxy Music and began making and
producing records. In the late 1970's, he again began doing visual
arts, making installations utilizing light, video, slides, sound, and
an occasional computer. His book, A Year With Swollen Appendices, his
diary for 1995, is a fascinating read. [-dra]
____________
BROAD AREA COLLOQUIUM FOR
AI-GEOMETRY-GRAPHICS-ROBOTICS-VISION
on Wednesday, 13 October 1999, 4:15pm
Gates B12
http://robotics.stanford.edu/ba-colloquium/
Modular Reconfigurable Robotics
Mark Yim
Xerox PARC
Modular, self-reconfigurable robots are those that are made up of a
large number of modules, but a small number of module types. As the
number of modules increases, these systems show promise of great
versatility, robustness and low cost. However, to make this realizable
there are many computational and manufacturing issues that must be
addressed.
We will show the progress of two modular reconfigurable robot systems;
PolyBot and Proteo, and present some of the issues in applying them to
a search and rescue task and a shape configuration task respectively.
These tasks are rich in interesting problems in motion planning in
unstructured environments, distributed computation and control, robust
redundant actuation and control, computational geometry, and image
understanding/sensor fusion.
Biography: Mark Yim has been a researcher at the Xerox Palo Alto
Research Center (PARC) since 1996. Currently, he leads a government
funded project building a modular, reconfigurable robot system at
PARC. He has recently authored a book chapter on robots for kids. He
has published in journals and conferences in the areas of mobile robot
planning, distributed robotics, optimal control, MEMS, and haptic
devices. He has authored over 20 patents. His work on MEMS and
robotics has been featured in the New York Times, USA Today, Discovery
Channel, BBC news and MSNBC. He has been nominated as one of the
TR100, the top 100 young innovators by Technology Review Magazine. He
received his PhD in mechanical engineering from Stanford University in
1994.
____________
CCRMA HEARING SEMINAR
on Thursday, 14 October 1999, 11:00am
CCRMA Library, The Knoll
http://www-ccrma.stanford.edu/CCRMA/Events/Events.html
Dichotic Pitch: The "magic ear"
Bob Dougherty
Stanford
Two patterns of appropriately filtered acoustic white noise can be
binaurally fused by the human auditory system to extract pitch and
location information that is not available to either ear alone. This
phenomenon, called dichotic pitch, is the auditory analog of the
random-dot stereogram. I will present a new algorithm for generating
dichotic pitch which produces stimuli that are more effective and
useful than those of previous algorithms. This new method allows the
psychophysical assessment of dichotic pitch detection thresholds and
also allows manipulation of several key parameters of the dichotic
pitch stimulus. I will demonstrate the effects of interaural coherence
and interaural time difference on the dichotic pitch percept. I will
also show that dichotic pitch detection is significantly impaired in
some individuals with developmental dyslexia. These results suggest a
low-level auditory deficit associated with dyslexia and also
demonstrate the potential value of these dichotic pitch stimuli for
assessment of auditory processing.
Biography: Bob Dougherty is currently a PostDoc in the vision group
here at Stanford. He did his work on dichotic pitch while he was a
PostDoc at Univ. of British Columbia. Bob has been coming to the
Hearing Seminar for a year now and I'm very happy to be able to
finally learn more about his own work. Check out the web sites at the
end of his abstract to hear some sample sounds. (Be sure you have a
stereo audio system (and headphones) hooked to your computer.)
Demos of dichotic pitch stimuli can be found at:
http://www.stanford.edu/~bobd/audDemos.html
Matlab source code for creating dichotic pitch stimuli can be found at:
http://www.stanford.edu/~bobd/demos/DPCode.html
____________
EXPERIMENTS IN LEARNING
on Thursday, 14 October 1999, 3:15pm
Stanford Press Warehouse, room 118
http://sll.stanford.edu/
Engineering Student Learning:
Research in 'Design' and 'Knowledge Integration'
Cynthia J. Atman
Director, Center for Engineering Learning and Teaching
Associate Professor, Industrial Engineering
University of Washington
Jennifer Turns
NSF Postdoctoral Fellow
Center for Engineering Learning and Teaching
University of Washington
The newly created Center for Engineering Learning and Teaching (CELT)
has two main goals - to conduct basic research on engineering student
learning and to help facilitate engineering student learning at the
University of Washington. In this talk, we will describe how CELT is
designed to tightly couple these two goals, and we will describe two
current programs that contribute to achieving these goals. In the
first half of the talk, we will focus on basic research to understand
engineering student design processes. Specifically, we will survey our
six years of research using verbal protocol analysis to explore
student approaches to engineering design. In the second half of the
talk, we will focus on efforts to facilitate learning through an
instructional project currently underway in civil and environmental
engineering - the Knowledge Integration Across the Curriculum (KIAC)
project. Specifically, we will describe the design and on-going
evaluation of a course to help students integrate knowledge gained
through prior classes and experiences into an engineering knowledge
framework. In addition, we will demonstrate how this project has led
to the identification of more basic research questions, consistent
with the first goal of CELT. We will conclude by returning to the CELT
research model and briefly describe other projects currently underway.
____________
XEROX PARC FORUM
on Thursday, 14 October 1999, 4:00pm - 5:00pm
George Pake Auditorium, Xerox
http://www.parc.xerox.com/ops/projects/forum/
Breaking In To Computer Systems:
Why It's Possible, How It's Done
Olin Sibert
InterTrust Technologies Corporation
Ever since computer systems have included controls on their resources,
people have figured out ways to bypass those controls, or "break in"
to the computer. Although computer security has become vastly more
visible to the public recently, the techniques for breaking in, and
the reasons that they work, actually haven't changed very much in over
thirty years. Market pressures, networks, personal computers, and new
technologies have simply allowed the same old software development
practices to produce the same old types of flaws in brand new
products. The major difference today is the number of people who can
figure out how exploit the flaws, and the speed with which they do so.
This talk, which includes a collection of examples from all across the
computing world, discusses the types of technical flaws that computer
systems tend to exhibit, the ways those flaws are exploited, and how
people go about finding them. It also discusses the patterns that
lead organizations to deliver products incorporating security flaws,
and how technology can contribute to breaking that cycle. Will the
situation always be as bad as it is today? Let's hope not; to
paraphrase the late Senator Everett Dirksen, break into a few million
computers here and a few million there, and pretty soon you're doing
real damage.
Biography: Olin Sibert has been breaking into computer systems off and
on for over 25 years. He spent some formative years working for
Honeywell and MIT on the Multics project, which produced one of the
few commercial computer systems for which security was a major goal.
After that, he spent 12 years as an independent consultant advising a
variety of computer vendors and the National Security Agency and
working with their secure operating system development and evaluation
programs. At present, Mr. Sibert is Vice President, Strategic
Technologies at InterTrust Technologies Corporation, where since late
1994 he has been responsible for defining the security and product
architecture of InterTrust's digital rights management E-commerce
products.
____________
TRANSFORMATION OF R & D IN EAST ASIA & JAPAN
on Thursday, 14 October 1999, 4:15pm-5:30pm
Skilling Engineering Auditorium
http://www.stanford.edu/~viji/
The Asia Pacific Advanced Network (APAN),
Its Activities and Impact
Dr. Kilnam Chon
Chairman, Asia Pacific Advanced Network &
Professor, Department of Computer Science
Korea Advanced Institute of Science and Technology (KAIST)
http://cosmos.kaist.ac.kr/salab/prof.html
http://www.apan.net/
Second generation networking started with various experiments on
gigabit testbeds around the world in the 1980's and 1990's. It is now
moving into the research and development production phase.
In Asia, various gigabit testbeds have been developed among many
countries and regions including Australia, China, Japan, Korea,
Malaysia, Singapore, Taiwan, and Thailand. Asia-Pacific Advanced
Network Consortium (APAN) was formed in 1997 with four founding
members (Australia, Japan, Korea and Singapore) to interconnect these
national and regional testbeds, as well as other high performance
networks, to offer a production network service for the research and
education community of the Asia-Pacific region with global
interconnectivity.
A specific example of APAN regional networking activities in Korea is
explained along with the network infrastructure for the research and
education community.
____________
LOGIC LUNCH
on Friday, 15 October 1999, 12 noon
Math Corner 380:383N
http://www-logic.stanford.edu/seminars.html
Recursive Types with a Modality
Hiroshi Nakano
Ryukoku University, visiting Stanford
Recursive types, or self-referential types, are indispensable to
understand some important features of programming languages, such as
inductive (or coinductive) data structures, and self-referential
objects in object-oriented programming. In particular,
object-oriented languages requires even types with negative
self-references. A model for such unrestricted recursive types was
invented by MacQueen, Plotkin and Sethi [1983], and has been widely
adopted as a basis for object-oriented typing systems.
On the other hand, it is well known that logical formulae with
unrestricted self-references introduce a contradiction called
Russell's paradox. Through the "formulae as types" notion, this
paradox corresponds to the fact that every type is inhabited by a
diverging program which does not provide any information, and implies
that even with the model mentioned above, one can regard types only as
partial specifications of programs, and must discuss their convergence
outside of the typing system.
In this talk, a modal typing system and its realizability
interpretation are proposed to capture recursive types, including such
paradoxical ones, in the formulae as types notion, where the
convergence of well-typed programs is ensured. Furthermore, this
typing system gives such a concise description of fixed point
combinators, such as Y, that one can construct recursive programs
using fixed point combinators referring only to their type.
____________
SEMINAR ON PEOPLE, COMPUTER, AND DESIGN
on Friday, 15 October 1999, 12:30-2:00pm
Gates B03 (NEC Classroom)
http://www-pcd.stanford.edu/cs547/
The Importance of Homes in Technology Research
Debby Hindus
Interval Research Corporation
In this talk, I will argue for the importance of home-related research
on technology. Several important differences between researching homes
and researching workplaces are described, and several issues in
conducting home-related research are discussed in the context of
specific research efforts. Ways to advance home-related research as a
discipline are presented, including an existing course on technology
design with a home focus.
Biography: Debby Hindus has been a Member of the Research Staff at
Interval Research Corporation in Palo Alto, CA, since 1992. Her
current research interests include broadband applications in the home
and wireless technologies. Ms. Hindus has co-authored several studies
of novel communications technology for workplaces and homes. In 1999,
Ms. Hindus taught a new Stanford course on The Design of Domestic and
Consumer Technologies.. Earlier research addressed a new kind of
computer-mediated communication, the audio space, and the design of
user interactions within an audio space. Ms. Hindus holds an MS degree
from the MIT Media Lab and a BSCS degree from the University of
Michigan. While in the Media Lab's Speech Research group, her work
focused on innovative speech applications for interacting with
computers.
____________
CS545: INFOLAB SEMINAR
on Friday, 15 October 1999, 3:15pm - 4:30pm
188 tcSEQ (across from Gates)
http://www-db.stanford.edu/dbseminar/dbseminar.html
Lore: A Database Management System for XML
Jennifer Widom
Database Group Faculty
Over the past several years database researchers have explored issues
in managing schema-less "semistructured" data: data that may be
irregular or incomplete, and whose structure may evolve rapidly and
unpredictably. The Lore system at Stanford, under research and
development for the past 4 years, is a full-featured DBMS designed
specifically for semistructured data. Fortuitously, Lore's original
data model, called the Object Exchange Model (OEM), is very similar to
W3C's new eXtensible Markup Language (XML), and we have just completed
migrating Lore to full XML compliance.
I will provide an overview of the Lore system and will briefly
highlight some of its more challenging and novel aspects: Lore's
expressive OQL-based query language, indexing capabilities, cost-based
query optimizer, dynamic structural summaries, and proximity search
capabilities. The talk will be followed by a demonstration of Lore's
stand-alone capabilities, and (time permitting) a demonstration of
Lore serving as an XML engine behind a Microsoft prepackaged XML demo.
To find out more about Lore or experiment with the online demos please
visit http://www-db.stanford.edu/lore
The talk presents joint work with Jason McHugh and Roy Goldman.
Biography: Jennifer Widom received her Bachelors degree from the
Indiana University School of Music in 1982 and her Ph.D. from Cornell
University in 1987. From 1987-88 she was a Visiting Assistant
Professor in the Computer Science Department at Cornell. She spent
five years as a Research Staff Member at the IBM Almaden Research
Center before joining the Stanford faculty in 1993.
____________
SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION (SCLA)
on Tuesday, 19 October 1999, 4:15pm to 5:30pm
Gates 104
(notice change in time and place)
http://www-csli.stanford.edu/cll/scla.html
jointly with NOBOTS
http://robotics.Stanford.EDU/groups/nobots/
A Bayesian Framework for Concept Learning
Josh Tenenbaum
Department of Psychology
Stanford University
I will describe a Bayesian framework aimed at understanding human
concept learning in computational terms and bringing machine concept
learning systems closer to the potential of human learners. I will
present theoretical analysis and data from several experiments with
human subjects that address three specific questions. How are people
often able to generalize a concept from only a small number of
positive examples? How does people's prior knowledge interact with
the examples they observe to guide the generalizations that they make?
Why does generalization appear in some cases to be based on abstract
rules and in other cases to be based on similarity to exemplars? I
will also discuss some implications for machine concept learning
systems, with a focus on the problem of example-based database
retrieval.
____________
CDR TALK
on Wednesday, 20 October 1999, 12 noon
Gates 498
Resource Allocation in Networks
using Constraints and Agents
Boi Faltings
Director, AI Lab
Swiss Federal Institute of Technology (EPFL)
http://liawww.epfl.ch/
Modern networking technologies require dynamic allocation of network
resources to traffic such as to ensure a certain service quality. For
most data networking technologies, this problem is NP-complete.
We present a novel abstraction technique that allows modelling network
resource allocation as a constraint satisfaction problem. It thus
becomes possible to apply a large body of heuristic techniques which
significantly outperform existing techniques.
We furthermore show how the abstraction admits an agent structure that
allows implementing distributed resource allocation with a low
communication overhead.
RSVP to Brenda Ackerman <mailto:ackerman@isl.stanford.edu>
____________
BERKELEY INTERNATIONAL COMPUTER SCIENCE INSTITUTE
on Wednesday, 20 October 1999, 12:30-2:30pm
Room 607, ICSI, 1947 Center Street, Sixth Floor, Berkeley
http://www.icsi.berkeley.edu/talks/
Semantic Lexicon Acquisition for
Learning Natural Language Interfaces
Cynthia Thompson
CSLI, Stanford University
A long-standing goal for the field of artificial intelligence is to
enable computer understanding of human languages. A core requirement
in reaching this goal is the ability to transform individual sentences
into a form better suited for computer manipulation. This ability,
semantic parsing, requires several knowledge sources, such as a
grammar, lexicon, and parsing mechanism.
Building natural language parsing systems by hand is a tedious,
error-prone undertaking. We build on previous research in automating
the construction of such systems using machine learning techniques.
The result is a combined system that learns semantic lexicons and
semantic parsers from one common set of training examples. The input
required is a corpus of sentence/representation pairs, where the
representations are in the output format desired. A recent system,
WOLFIE, learns semantic lexicons to be used as background knowledge by
a previously developed parser acquisition system, CHILL. The combined
system is tested on a real world domain of answering database queries.
We also compare this combination to a combination of CHILL with a
previously developed lexicon learner, demonstrating superior
performance with our system. Finally, we show the ability of the
system to learn to process natural languages other than English.
____________
BROAD AREA COLLOQUIUM FOR
AI-GEOMETRY-GRAPHICS-ROBOTICS-VISION
on Wednesday, 20 October 1999, 4:15pm
Gates B12
http://robotics.stanford.edu/ba-colloquium/
Taming the Giants and The Monsters:
Recent Developments in Data Mining
Usama Fayyad
Microsoft Research
Knowledge Discovery in Databases (KDD) and Data Mining are concerned
with the extraction of interesting structure from databases,
especially large stores. Following a brief overview of this rapidly
growing area of research and applications, I'll focus on data mining
methods. These methods have their origins in statistics, pattern
recognition, learning, visualization, databases, optimization, and
parallel computing.
I'll discuss some classification and clustering methods and how they
are scaled to large databases. I'll present results from our recent
work to demonstrate that the methods can be effectively scaled to work
with large databases with only limited memory resources. I'll outline
the research challenges and opportunities posed by the problem of
extracting models from massive data sets. Operating under such
scalability constraints poses interesting problems for how models can
be built and what methods are practical. Some applications will be
used to motivate and illustrate the techniques.
Biography: Usama Fayyad is a Senior Researcher at Microsoft Research
(http://research.microsoft.com/~fayyad). His research interests
include scaling data mining algorithms to large databases, learning
algorithms, and statistical pattern recognition, especially
classification and clustering. After receiving the Ph.D. degree from
The University of Michigan, Ann Arbor in 1991, he joined the Jet
Propulsion Laboratory (JPL), California Institute of Technology, where
(until 1996) he headed the Machine Learning Systems Group and
developed data mining systems for automated science data analysis. He
received the 1994 NASA Exceptional Achievement Medal and the JPL 1993
Lew Allen Award for Excellence in Research for his work on developing
data mining systems to solve challenging science analysis problems in
astronomy and remote sensing. He remains affiliated with JPL as a
Distinguished Visiting Scientist. He is a co-editor of Advances in
Knowledge Discovery and Data Mining (AAAI/MIT Press, 1996) and is an
Editor-in-Chief of the journal: Data Mining and Knowledge Discovery.
He was program co-chair of KDD-94 and KDD-95 (the First International
Conference on Knowledge Discovery and Data Mining) and is general
chair of KDD-96 and KDD-99. He co-chaired the 1997 Workshops on the
role of KDD in Visualizations held at KDD-97 and IEEE Vis-97
conferences.
____________
SEMINAR ON PEOPLE, COMPUTER, AND DESIGN
on Friday, 22 October 1999, 12:30-2:00pm
Gates B03 (NEC Classroom)
http://www-pcd.stanford.edu/cs547/
The Soft Machine - Design in the Cyborg Age
Marie O'Mahony
The term cyborg was coined by Manfred E. Clynes and Nathan S. Kline in
1960 to refer to an enhanced man machine hybrid who could survive in
extra-terrestrial environments. Technology and thinking have moved a
long way since the first cyborg (a white rat) and, since it has
started to become part of our design and popular culture, it is time
to assess the implications of what is becoming the cyborg age.
This slide presentation will move between such diverse disciplines as
Art, Medicine, Science Fiction, Design and Science. This juxtaposition
of images and ideas from very different areas reflects the current
trend towards closer links between Art and Science.
Biography: Marie O'Mahony is an independent consultant specializing in
textiles and technology based in London. She has worked for companies
and institutions advising on projects, preparing reports and
organizing workshops, symposiums and exhibitions. Clients include The
Netherlands Design Institute, Interval Research Corporation, Ove Arup
and Partners, Stedelijk Museum of Modern Art, Interstoff at Messe
Frankfurt and Zaha M Hadid. She is co-author of TechnoTextiles and
currently researching a book to be titled The Soft Machine - Design in
the Cyborg Age for publication in Spring 2001.
____________
CSLI IAP CONFERENCE
ON HUMAN-COMPUTER INTERACTION
November 10-12, 1999
Cordura Hall, Room 100
Stanford University
http://www-csli.stanford.edu/Tutorials/schedule.shtml
The latest research on how humans and computers interact will be the
subject of a conference November 10 to 12, sponsored by the Stanford
Center for the Study of Language and Information (CSLI). Planned for
businesses and other organizations concerned with this topic, the
conference is free and open to the public; however, space is limited
so registration is required by October 27, 1999.
Sessions will run from 9am to 12:30pm and 1:30pm to 5pm on November 10
and 11 in Room 100 of Cordura Hall, at the corner of Panama Street and
Campus Drive West. On November 12, the conference will be in the same
location but finish by 2:30pm. Between speakers, demonstrations of
research prototypes in progress will be offered.
To register or for further information, go to the web page
http://www-csli.stanford.edu/Tutorials/schedule.shtml
or contact
Michele King
Industrial Affiliates Program
Center for the Study of Language and Information (CSLI)
Ventura Hall
Stanford, CA 94305-4115
Tel:(650) 723-3084
Fax:(650) 723-0758
Email: mking@csli.stanford.edu
____________
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. Problems with subscribing or unsubscribing should
be sent to owner-csli-calendar@csli.stanford.edu.
The full current issue is at
http://www-csli.stanford.edu/csli/Archive/calendar/current.html
and the archives at
http://www-csli.stanford.edu/csli/Archive/calendar/.
People on most of the CSLI computers can type 'help csli-calendar' to
see the current issue.
The CSLI Calendar is also posted each week to
news://nntp-csli.stanford.edu/csli.bboard.
Information about CSLI's research program is available at
http://www-csli.stanford.edu/.
For maps to the Stanford University campus see
http://www.stanford.edu/home/visitors/maps.html.
____________