[Date Prev][Date Next][Date Index]
CSLI Calendar, 2 February 2000, vol. 15:17
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
______________________________________________________________________
2 February 2000 Stanford Vol. 15, No.17
______________________________________________________________________
A weekly publication of the
Center for the Study of Language and Information (CSLI)
Stanford University, Ventura Hall, Stanford, CA 94305-4115
____________
ACTIVITIES FROM 2 FEBRUARY TO 11 FEBRUARY 2000
WEDNESDAY, 2 FEBRUARY
3:45pm Psychology Colloquium
Building 420:041
The Detection of Mentalistic Agents in Infancy
Susan Johnson
University of Pittsburgh
http://matia.stanford.edu/html/talks.html#colloq
4:15pm Broad Area Colloquium For
AI-Geometry-Graphics-Robotics-Vision
TCseq201, Lecture Hall B
Inner-loop Statistics in Automated Scientific
Discovery from Massive Datasets
Andrew Moore
Robotics Institute and Computer Science CMU
and Schenley Park Research, Inc.
http://robotics.stanford.edu/ba-colloquium/#Schedule
Abstract below
4:15pm EE380: Computer Systems Laboratory Colloquium
NEC Auditorium, Gates B03
Architecture and Performance of the Direct RDRAM
Steve Woo
Rambus
http://www.stanford.edu/class/ee380/schedule.html
Abstract below
5:15pm ME297: Design Theory and Methodology Seminar
Building 560
A School of Design: One View
Larry Leifer, Ph.D.
Stanford University
http://design.stanford.edu/Courses/me297/#2/2
Abstract below
THURSDAY, 3 FEBRUARY
12:00pm Award-Winning Teachers on Teaching
Hartley Conference Center,
Mitchell Earth Sciences Building
The Value of the Laboratory
Experience in the Sciences
Richard Zare
Stanford University
http://www-ctl.stanford.edu/events.html
12:15 Coglunch 2000: Evolution and the Mind
Cordura 101
Evolution of the Rational Faculty:
Implications for Acquisition of Scientific
and Moral Knowledge
Roger Shepard
Emeritus Professor of Psychology
Stanford University
http://www-csli.stanford.edu/events/
12:15pm Stanford Networking Seminar
Gates 104
Application Performance Pitfalls and
TCP's Nagle Algorithm
Greg Minshall
Siara Systems
http://netseminar.stanford.edu/
Abstract below
4:15pm Stanford Mathematics Colloquium
Building 380:380W
Hilbert's Tenth Problem Today:
Main Results and Open Problems
Yuri Matiyasevich
Steklov Institute of Mathematics, St. Petersburg
Abstract below
5:15pm Philosophy Colloquium
Building 60:61G
Permutation Invariance and the Generality of Logic
John MacFarlane
Philosophy, University of Pittsburgh
http://www-csli.stanford.edu/philosophy/
FRIDAY, 4 FEBRUARY
11:00am CCRMA Hearing Seminar
CCRMA Ballroom (main floor of the Knoll at Stanford)
A Not-So-Gentle Introduction to Pitch Perception
Malcolm Slaney
Stanford University
http://www-ccrma.stanford.edu/
12:30pm CS547: Human-Computer Interaction Seminar
Gates B01
Gender and Technology: A Case Study
Brenda Laurel
http://hci.stanford.edu/cs547/
Abstract below
3:00pm Applied Math Seminar
Building 380:380C
Randomized Approximate L^p-Difference
Algorithms for Massive Data Streams
Martin Strauss
AT&T Labs - Research
http://math.stanford.edu/programs/applied/seminar.html
Abstract below
3:15pm Philosophy Colloquium
Building 90:92Q
Nominalism and Epistemic Relativism
Gideon Rosen
Philosophy, Princeton University
http://www-philosophy.stanford.edu
3:30pm Linguistics Colloquium
Building 460:126
Subjects, Objects, and the Extended Projection
Principle
Howard Lasnik
University of Connecticut/CASBS,
Stanford University
http://www-linguistics.stanford.edu/Linguistics/
Abstract below
MONDAY, 7 FEBRUARY
3:30pm Social Lab Talk
Building 420:050
On Coincidences
Perci Diaconis
Stanford University
http://matia.stanford.edu/html/talks.html#social_lab
4:15pm Probability & Stochastic Processes Seminar
Sequoia Hall:200
A New Approach to Perfect Sampling From
Nasty Distributions
Mark Huber
IEOR, Cornell
http://www-stat.stanford.edu/~amir/prob-seminar/
Abstract below
TUESDAY, 8 FEBRUARY
10:00am Special Linguistics Talk
Building 460:126
V1 Phenomena And The Syntax Prosody Interface
David Adger
University of York
http://www-csli.stanford.edu/events/
3:15pm Stanford Learning Lab Presentation
Press Warehouse, Staff Training Room
A Real-time Group Communication System
Using Immersive Natural Metaphors
Steve DiPaola
Onlive Group Communities.com
http://sll.stanford.edu/
Abstract below
4:15pm Special Logic Seminar
Building 380:381T
Metatheorems and the Logicists
Saul Kripke
Princeton University
WEDNESDAY, 9 FEBRUARY
4:15pm Broad Area Colloquium For
AI-Geometry-Graphics-Robotics-Vision
TCseq201, Lecture Hall B
Stereo Algorithms and Representations for
Image-Based Rendering
Richard Szeliski
Vision Technology Group
Microsoft Research
http://robotics.stanford.edu/ba-colloquium/#Schedule
Abstract below
4:15pm EE380: Computer Systems Laboratory Colloquium
NEC Auditorium, Gates B03
Transmeta's Crusoe Processor
Dave Ditzel
Transmeta Corp.
http://www.stanford.edu/class/ee380/schedule.html
4:15pm ME297: Design Theory and Methodology Seminar
Building 560
Designing a Consumer Experience
Bill Cockayne
Co-Founder, Scout Electromedia
http://design.stanford.edu/Courses/me297/#2/2
THURSDAY, 10 FEBRUARY
12:00pm Award-Winning Teachers on Teaching
Hartley Conference Center,
Mitchell Earth Sciences Building
Soap Bubbles, Thermodynamics, and
Engineering Science: Teaching the Ideas
Behind all the Mathematics
Dean Lynn Orr
School of Earth Sciences
http://www-ctl.stanford.edu/events.html
12:45pm Stanford Networking Seminar
Gates 104
Developments in Optical Networks
Andreas V. Bechtolsheim, Cisco Systems
http://netseminar.stanford.edu/
4:00pm Xerox PARC Forum
George Pake Auditorium at Xerox PARC
10 Years on 2 Wheels":
A Photographer's Journey Around the World
Helge Pedersen
Globeriders
http://www.parc.xerox.com/ops/projects/forum/
FRIDAY, 11 FEBRUARY
12:15 Coglunch 2000: Evolution and the Mind
Cordura 101
The Good, the Bad and the Outrageous:
Theories of Language Evolution and their
Implications for Linguistic Nativism
Fiona Cowie
Division of Humanities and Social Sciences,
California Institute of Technology
http://www-csli.stanford.edu/events/
12:30pm CS547: Human-Computer Interaction Seminar
Gates BO1
In Support of Multimedia Conversations
Greg Wolff
Ricoh Silicon Valley
http://www-pcd.stanford.edu/cs547/
Abstract below
3:00pm Applied Math Seminar
Building 380:380C
Estimating Deformations of Stationary Processes
Maureen Clerc
Stanford Statistics
http://math.stanford.edu/programs/applied/seminar.html
3:15pm Friday Cognitive Seminar
Building 420:100
Memory for Motion Events
Meredyth Krych
http://matia.stanford.edu/html/talks.html#frisem
3:30pm Information Packaging Seminar
Building 460:126
The Role of Secondary Predication in
Information Packaging: The French
Presentational Relative Construction
Knud Lambrecht
University of Texas
http://www-linguistics.stanford.edu
Abstract not included due to technical constraints.
Please go to:
http://www-linguistics.stanford.edu/Linguistics/semgroup/lambrecht.html
____________
BROAD AREA COLLOQUIUM FOR
AI-GEOMETRY-GRAPHICS-ROBOTICS-VISION
on Wednesday, 2 February 2000, 4:15pm
TCseq201, Lecture Hall B
http://robotics.stanford.edu/ba-colloquium/#Schedule
Inner-loop Statistics in Automated Scientific
Discovery from Massive Datasets
Andrew Moore
Robotics Institute and Computer Science CMU
and Schenley Park Research, Inc.
Intensive statistical analysis of massive data sources ("data mining")
has been embraced as one of the final areas with a need for massive
computation beyond that available on a $2000 computer or $200
videogame. We begin this talk with two examples of software, instead
of hardware, giving 1000-fold speedups over traditional
implementations of statistical algorithms for prediction, density
estimation, and clustering. We then pause to examine directions in
which these software solutions seemed blocked when faced with Physics,
Biology and commercial scientific data discovery problems. The primary
blocks were a curse of dimensionality and limitations on machine main
memories. This is followed by four examples of new pieces of research
that circumvent these barriers: lazy cached sufficient statistics,
exact accelerated k-means, multiresolution ball-trees for very high
dimensional real-valued data, and filament identifiers. We then reveal
the reason for our new-found respect for super-computation: when an
algorithm you previously ran overnight executes in seconds, you find
yourself wanting to run it ten thousand times. We show the impact of
being able to run intensive statistics as an inner loop has had on our
analysis of cosmology data (preliminary data from the Sloan Digital
Sky Survey) and biotoxin identification, where desirable but
hopelessly extravagant operations such as model selection,
bootstrapping, backfitting, randomization and graphical model design
now become somewhat non-hopeless. Joint work with Andy Connolly (U
Pitt Physics), Artur Dubrawski (Schenley Park Research), Geoff Gordon
(Auton Lab), Paul Komarek (Auton Lab), Bob Nichol (CMU Physics), Dan
Pelleg (Auton Lab) and Larry Wasserman (CMU Statistics).
About the Speaker:
Andrew Moore (www.cs.cmu.edu/~awm) is the A. Nico Haberman Associate
Professor of Computer Science and Robotics at CMU. He received a Phd
in Computer Science from the University of Cambridge in 1991 (thesis
topic: Robot Learning). He has worked with robots that learn,
factories than learn and supply chains that learn. His research
interests include: statistical foundations, autonomous learning
systems for manufacturing, efficient algorithms for machine learning
from massive data and reinforcement learning, finite production
scheduling, and machine learning applied to optimization. He is the
co-owner and CTO of Schenley Park Research Inc---a 12 person
Pittsburgh-based AI startup supplying data mining and decision theory
products and solutions to manufacturing, business-to-business and
biotechnology clients.
____________
EE380: COMPUTER SYSTEMS LABORATORY COLLOQUIUM
on Wednesday, 2 February 2000, 4:15pm
NEC Auditorium, Gates B03
http://www.stanford.edu/class/ee380/schedule.html
Architecture and Performance of the Direct RDRAM
Steve Woo
Rambus
Over the past decade, rising clock speeds and advances in processor
design such as superscalar processing and simultaneous multithreading
have led to dramatic increases in CPU performance. Over the same time
period memory performance has increased more slowly, with increases
coming more from advances in manufacturing rather than from
revolutionary architectural improvements. Consequently, the
processor-memory performance gap has grown wider. If these trends
continue, the performance of future computing platforms will be
constrained more and more by the performance of their memory systems.
This talk will describe the Direct Rambus DRAM (Direct RDRAM), a novel
architecture designed to help reduce the processor-memory performance
gap. The Direct RDRAM was designed to meet the needs of several
computing environments, including home entertainment units, desktop
personal computers, laptops, and workstations/servers. Some of the
important design considerations for these computing environments will
be discussed, along with how they ultimately manifested themselves in
the Direct RDRAM architecture. A discussion of the performance
advantages of systems that utilize Direct RDRAMs compared to systems
that use other commercially-available alternatives will be included.
About the speaker:
Steven Woo is a Member of the Technical Staff in the Logical
Architecture Group of Rambus Inc., a company that designs, develops,
and licenses its high-speed chip-to-chip interface technology to
enhance the performance and cost-effectiveness of computing
systems. Since joining Rambus Inc. in 1996, he has been a member of
the Direct RDRAM team focusing on the development of detailed
event-driven simulators and performance analysis. Prior to joining
Rambus, he worked in the Neural Network Technology Center at Hughes
Aircraft Company, Ground Systems Group. He received BS Engineering
(1986) and Master of Engineering (1988) degrees from Harvey Mudd
College, and MS (1991) and PhD (1996) degrees in Electrical
Engineering from Stanford University.
____________
ME297: DESIGN THEORY AND METHODOLOGY SEMINAR
on Wednesday, 2 February 2000, 5:15pm
Building 560
http://design.stanford.edu/Courses/me297/#2/2
A School of Design: One View
Larry Leifer, Ph.D.
Stanford University
For the past two and a half year, Larry Leifer has been applying
principles of Design to the problem of designing new forms of improved
learning experiences for Stanford students as the Director of the
Stanford Learning Lab. He has, though, been developing new learning
experiences at Stanford for several decades as a faculty member in the
Design Division of the School of Engineering. Some of the courses he
has taught include the ME101: Visual Thinking and ME210: Team-Based
Design with Corporate Partners sequences, and is currently teaching a
freshman seminar in design, ME122N: Designing the Human
Experience - An Exploration into the Theory and Practice of Design
Thinking. In this informal presentation, Prof. Leifer will share some
of his thoughts - garnered from his extensive experience and
conversations with others - on a School of Design. This session
will invite the members of the audience to express ideas and interact
freely with the speaker throughout the session. In addition to being
an opportunity to explore one view of the design school concept, it
will be an opportunity for members of the seminar cohort to sound some
of their own ideas in preparation for the papers they will be
presenting at the end of the quarter. All are invited to attend and
join in the discussion.
Larry Leifer is the Director of the Center for Design Research and the
Stanford Learning Lab. His research interests include rehabilitation
engineering, design methodology, global learning, and programmable
electromechanical systems, among others. He has published in the areas
of diagnostic electrophysiology, functional assessment of voluntary
movement, human operator information processing, rehabilitation
robotics, design team protocol analysis, design knowledge capture and
concurrent engineering.
____________
STANFORD NETWORKING SEMINAR
on Thursday, 3 February 2000, 12:15pm to 1:45pm
Gates 104
http://netseminar.stanford.edu/
Application Performance Pitfalls and TCP's Nagle Algorithm
Greg Minshall
Siara Systems
Performance improvements to networked applications can have unintended
consequences. In a study of the performance of the Network News
Transport Protocol (NNTP), the initial results suggested it would be
useful to disable TCP's Nagle algorithm for this application. Doing so
significantly improved latencies. However, closer observation revealed
that with the Nagle algorithm disabled, the application was
transmitting an order of magnitude more packets. We found that proper
application buffer management significantly improves performance, but
that the Nagle algorithm still slightly increases mean latency. We
suggest that modifying the Nagle algorithm would eliminate this cost.
About the speaker:
Greg Minshall is on the technical staff of Siara Systems, a networking
startup located in Silicon Valley. Previously, he was at Ipsilon
Networks, Novell, Kinetics, and the University of California,
Berkeley. He has been involved in the design and implementation of
internet protocols, at the routing and transport layers, since the
early 1980s. At one time in his life, he worked as a hardware
technician/engineer on the Illiac IV SIMD computer.
_____________
STANFORD MATHEMATICS COLLOQUIUM
on Thursday, 3 February 2000, 4:15pm
Building 380:380W
Hilbert's Tenth Problem Today:
Main Results and Open Problems
Yuri Matiyasevich
Steklov Institute of Mathematics, St. Petersburg
Among the 23 problems stated by David Hilbert in 1900 we find:
10. Determination of the Solvability of a Diophantine Equation.
Given a diophantine equation with any number of unknown quantities and
with rational integral numerical coefficients: {To devise a process
according to which it can be determined by a finite number of
operations whether the equation is solvable in rational integers.}
The problem was shown to be undecidable in 1970. Since that time over
300 papers have been published about simplifications, improvements,
and applications of this result in different branches of mathematics.
In the lecture I plan to survey the main achievements in this area and
discuss some important related questions which still remain open.
_____________
CS547: HUMAN-COMPUTER INTERACTION SEMINAR
on Friday, 4 February 2000, 12:30pm to 2:00pm
Gates B01
http://hci.stanford.edu/cs547/
Gender and Technology: A Case Study
Brenda Laurel
This presentation will review research conducted by Interval Research
in collaboration with Cheskin Research that led to the formation of
Purple Moon, a company devoted to interactive media for preteen
girls. Purple Moon was formed in 1996, with initial product launch in
the fall of 1997.
The company was acquired by Mattel in 1999. During its life, ongoing
research played a major role in the formation of new product concepts
and designs. This talk will discuss research methods and key findings,
how research was translated into design and marketing decisions, and
the other cultural forces that influenced the company's products and
its fate.
Brenda Laurel is a designer, researcher and writer. Her work focuses
on interactive narrative, human-computer interaction, and cultural
aspects of technology. Her career in human-computer interaction spans
over twenty years. She holds an M.F.A. and Ph.D. in theatre from the
Ohio State University. Her doctoral dissertation was the first to
propose a comprehensive architecture for computer-based interactive
fantasy and fiction. Brenda was one of the founding Members of the
research staff at Interval Research Corporation in Palo Alto,
California, where she coordinated research activities exploring gender
and technology, and where she co-produced and directed the Placeholder
Virtual Reality project. She was also one of the founders and
VP/Design of a spinoff company from Interval - Purple Moon - formed to
market products based on this research. Purple Moon was acquired by
Mattel in 1999. In 1990 she co-founded Telepresence Research, Inc. to
develop virtual reality and remote presence technology and
applications. She has worked as a software designer, producer, and
researcher for companies including Atari, Activision, and
Apple. Brenda has published extensively on topics including
interactive fiction, computer games, autonomous agents, virtual
reality, and political and artistic issues in interactive media. She
is editor of the book, The Art of Human-Computer Interface Design
[Addison-Wesley 1990] and author of Computers as Theatre
[Addison-Wesley 1991; 2nd edition 1993], and a collection of essays
entitled Severed Heads.
____________
APPLIED MATH SEMINAR
on Friday, 4 February 2000, 3:00pm
Building 380:380C
http://math.stanford.edu/programs/applied/seminar.html
Randomized Approximate L^p-Difference
Algorithms for Massive Data Streams
Martin Strauss
AT&T Labs - Research
A computer with severely limited memory is given an
arbitrarily-ordered stream of elements from two finite sequences, such
as a_7 = 5, a_4 = 3, b_1 = 2, a_5 = 1, b_7 = 2, ... and the computer's
goal is to approximate the sum of |a_i - b_i| ^ p with high
probability. This problem has applications in web search engines,
catalogs of high dimensional data sets (such as images), and others.
We survey a number of recent results and techniques for this and
related problems. A goal of this talk is highlight a variety of tools
and theories that the design and analysis of randomized algorithms
shares with applied mathematics, including:
Bounding of large deviations---Chernoff and Chebychev Random variables
of limited independence Algebraic coding theory Pseudorandom number
generators Estimating parameters of probability distributions Stable
distributions Self-similar measures
No background in theoretical or applied computer science will be
assumed; the tiny bit needed will be presented.
____________
LINGUISTICS COLLOQUIUM
on Friday, 4 February 2000, 3:30pm
Building 460:126
http://www-linguistics.stanford.edu/Linguistics/
Subjects, Objects, and the Extended Projection
Principle
Howard Lasnik
University of Connecticut/CASBS,
Stanford University
There is strong evidence for 'object shift' in English. That is, just
as there is an 'EPP position' high in the sentence where subjects and
derived subjects wind up, there is a similar 'EPP position' in the VP
region where objects and ECM subjects (among other categories) wind
up. The evidence comes from several phenomena. First, for various
binding purposes ECM subjects act as if they are higher than elements
in the higher clause. (1) is a representative example involving
Condition A satisfaction:
(1) The DA proved [two men to have been at the scene of the crime]
during each other's trials Other parallel examples involve weak
crossover and negative polarity item licensing. Under the standard
assumption that c-command is involved in all of these phenomena, and
given that the adverbial clauses containing the item that needs to be
licensed is in the higher clause, the acceptability of examples like
(1) indicates that the ECM subject is in the higher clause as well.
In the kinds of paradigms just mentioned, the behavior of ECM subjects
is comparable to that of transitive objects. The following example
parallels (1):
(2) The DA accused two men during each other's trials Note that
under reasonable (though not universally accepted) assumptions about
clause structure, even direct object is not high enough to c-command
into an adverbial adjunct. So, under those assumptions, even object
raises. The null hypothesis is that object and ECM subject raise to
the same higher position.
An additional argument for raising of an object or an ECM subject has
to do with the Pseudogapping ellipsis construction. This construction
is exemplified in (3).
(3) Mary hired John, and Susan will [hire] Bill. Plausibly, the
construction involves VP ellipsis, with the remnant having escaped
from the ellipsis site via a movement operation, object shift.
The clausal phrase structure proposed by Chomsky (1991) provides a
possible target for the raising motivated by the above phenomena:
[Spec, AgrO], where AgrO is a functional head just above VP. Chomsky
had suggested that such raising exists, but that it is covert,
happening in the LF component. However, there are good reasons, which
I will summarize, to think that the raising is actually overt with
raising of V to a still higher position. Covert movement, on the other
hand, will typically involve formal features alone, hence does not
create any relevant binding or licensing configurations. The
'split-VP' hypothesis of Koizumi (1993) and Koizumi (1995), which I
adopt in its essentials, provides the needed structure for overt
raising. Within such an approach, it is natural to assume that the
'EPP' feature driving raising to 'subject position' resides in Agr,
hence is also responsible for raising to 'object position', under the
plausible assumption of Chomsky (1991) that 'AgrS' and 'AgrS' are
merely mnemonic, there really being just Agr, which can occur in
various places in the structure. This result constitutes part of a
promising reduction of an apparent asymmetry between subject and
object. However, unlike the situation with 'subject shift', object
shift is not obligatory. First, extraction out of an object is much
more acceptable than extraction out of a subject (the CED effect):
(4) Who was [a picture of t] selected
(5) Who did you select [a picture of t]
As Branigan (1992) points out, if object and subject both necessarily
raise overtly, to [Spec, AgrO] and [Spec, AgrS] respectively, whatever
constraint is responsible for CED effects cannot distinguish (4) from
(5). I will explore a number of interactions between extraction and
'high' binding effects, all of them indicating that when the object or
ECM subject is high, extraction out of that NP is degraded, in accord
with the CED.
Verb-particle constructions provide additional evidence. Johnson
(1991) persuasively argues that the order V-NP-prt arises from the
raising of the NP from its base position, and the further raising of
the V portion of the 'particle-verb'. Pairs like the following, then,
indicate that the raising of the NP is optional:
(6) Mary called up friends of John
(7) Mary called friends of John up
When the NP precedes the particle, extraction out of the NP is
seriously degraded, as now expected:
(8) Who did Mary call up friends of
(9) Who did Mary call friends of up
Finally, I will examine a very interesting verb-particle construction
first discussed by Kayne (1985) and later analyzed by Johnson (1991)
as involving overt raising of the ECM subject 'John'.
(10) Mary made John out to be a fool
Observe that the raising seen in (10) is optional. For most speakers,
(12) is an acceptable alternative to (10).
(11) Mary made out John to be a fool
Thus, we have yet another instance of optional object shift. The
scope properties of the construction prove particularly interesting,
so I will examine them in some detail.
____________
PROBABILITY & STOCHASTIC PROCESSES SEMINAR
on Monday, 7 February 2000, 4:15pm
Sequoia Hall:200
http://www-stat.stanford.edu/~amir/prob-seminar/
A New Approach to Perfect Sampling From
Nasty Distributions
Mark Huber
IEOR, Cornell
Visiting - Stanford, Statistics Department
The problem of how to generate samples from high dimensional
distributions has applications in a wide variety of areas, from
statistical mechanics to Bayesian statistics. Commonly, Monte Carlo
Markov chain techniques are used, where a Markov chain is run "for a
long time". Our approach to these problems is fundamentally
different, and requires no knowledge of how long to run the chain.
Unlike other perfect sampling techniques, our method is not bound to
the classic Markov chains for these problems, and so is the first to
achieve linear time algorithms for sampling from these distributions.
_____________
STANFORD LEARNING LAB PRESENTATION
on Tuesday, 8 February 2000, 3:15pm to 4:30pm
Press Warehouse, Staff Training Room
http://sll.stanford.edu/
A Real-time Group Communication System
Using Immersive Natural Metaphors
Steve DiPaola
Onlive Group Communities.com
OnLive's Internet-based real-time group communication system allow
groups of people to socialize with each other by navigating through 3D
spaces, meeting others and talking with their own voices through
emotive, lip-syncing 3D head avatars.
Our design goal was to develop a virtual community system that
emulates natural social paradigms, where the participants sense a
tele-presence that they are really there in the virtual space with
other people. This collective sense of "being-there" does not happen
over the phone or with teleconferencing; it is a new and emerging
phenomenon, unique to immersive virtual communities. We borrow from
disciplines such as group dynamics, facial animation, architectural
design, virtual reality and cognitive sciences, which allows the
system to draw from the natural social neural programming inherent in
all of us rather than creating artificial, social-enabling user
interface mechanisms.
One might assume, given our natural emulation approach that our design
choice is to strive to make avatars and spaces increasingly
realistic. However this is not the case. We have tried to emulate
natural paradigms just enough to achieve recognition of
familiarity. Once this familiarity is achieved; we have learned it is
better to use the given resources to achieve another natural metaphor
instead. It is the combination of several interconnected natural
metaphors that brings this immersive "sense of presence" and
engagement. Once this level of virtual engagement is achieved, we
believe an enhanced level of socialization, learning, and
communication is achievable.
Bio: Steve DiPaola has been involved with computer based character
systems for many years starting back in 1984 when he was a senior
member of the computer animation research group at the New York
Institute of Technology. He specialized in 3D character animation R&D
as well as producing animation for film, TV and his Fine Art work. His
main area of expertise at NYIT was 3D Facial Animation and has
published several papers and book excerpts on the subject. He is
currently Director of Development at Communities.com's OnLive Group,
where he leads a team of artists, architects, UI designers and
musicians in designing and developing 3D avatars and virtual spaces He
co-headed the San Francisco office of Saatchi and Saatchi's innovation
arm called Darwin Digital as Creative Director. Darwin Digital was
mandated to explore state of the art new media and interactive
projects including several Internet based characters projects.
____________
BROAD AREA COLLOQUIUM FOR
AI-GEOMETRY-GRAPHICS-ROBOTICS-VISION
on Wednesday, 9 February 2000, 4:15pm
TCseq201, Lecture Hall B
http://robotics.stanford.edu/ba-colloquium/#Schedule
Stereo Algorithms and Representations for
Image-Based Rendering
Richard Szeliski
Vision Technology Group
Microsoft Research
In this talk, I will review a number of stereo matching algorithms and
representations I have developed in the last few years. The talk
focuses on techniques that are especially well suited for image-based
rendering applications such as novel view generation and the mixing of
live imagery with synthetic computer graphics. I will begin by
reviewing some recent approaches to the classic problem of recovering
a depth map from two or more images. I will then describe a number of
newer representations (and their associated reconstruction
algorithms), including volumetric representations, layered
plane-plus-parallax representations, and multiple depth maps. Each of
these techniques has its own strengths and weaknesses, which I will
address.
About the Speaker:
Richard Szeliski is a Senior Researcher in the Vision Technology Group
at Microsoft Research, where he is pursuing research in 3-D computer
vision, video scene analysis, and image-based rendering. His current
focus is on constructing photorealistic 3D scene models from multiple
images and video. He received a Ph. D. degree in Computer Science from
Carnegie Mellon University, Pittsburgh, in 1988. He joined Microsoft
Research in 1995. Prior to Microsoft, he worked at Bell-Northern
Research, Schlumberger Palo Alto Research, the Artificial Intelligence
Center of SRI International, and the Cambridge Research Lab of Digital
Equipment Corporation.
Dr. Szeliski has published over 60 research papers in computer vision,
computer graphics, medical imaging, neural nets, and parallel
numerical algorithms, as well as the book Bayesian Modeling of
Uncertainty in Low-Level Vision. He served as co-chair of the SPIE
Conferences on Geometric Methods in Computer Vision, the 1999 Vision
Algorithms Workshop, and as an Associate Editor of the IEEE
Transactions on Pattern Analysis and Machine Intelligence.
_____________
CS547: HUMAN-COMPUTER INTERACTION SEMINAR
on Friday, 11 February 2000, 12:30pm to 2:00pm
Gates BO1
http://www-pcd.stanford.edu/cs547/
In Support of Multimedia Conversations
Greg Wolff
Ricoh Silicon Valley
This presentation describes a set of research prototypes designed to
improve the effectiveness of computer mediated communication. Most
electronic communication today lacks the richness of face-to-face
interaction. Email users quickly learn about the misunderstandings
that can occur with humorous or sarcastic messages (notwithstanding
the familiar "smiley-face" emoticon :-) and even experienced
correspondents have difficulty conveying the proper urgency or
IMPORTANCE of a message. In addition to the emotional content,
face-to-face communication also provides a shared context that allows
participants to easily reference topics of discussion. Speakers
naturally point to images or pick out particular lines from a
spreadsheet they may be talking about. Such actions have no
counterparts in existing electronic messaging systems.
Our prototypes address the deficiencies of electronic communication
through a number of novel human-computer interaction techniques that
use naturally recorded speech to convey emotional cues and provide a
context for referring to photos, documents, and other multimedia
objects. The portable "StoryTrack" device, acts as a kind of digital
photo album that explicitly supports the creation of stories or
narratives illustrated by the digital photos. Another prototype runs
in a more standard application environment and uses a "point & talk"
interaction model for easily composing and viewing multimedia
messages. Preliminary usage results will be presented that demonstrate
the effectiveness of these designs for particular types of
conversations.
Greg Wolff leads the information appliances research group at Ricoh
Silicon Valley. After receiving degrees in Cognitive Science from MIT
(BS) and Carnegie Mellon (MS), Greg developed one of the first WYSIWYG
hypertext markup language authoring tools in 1989 while at IBM's Human
Factors lab. At Ricoh, he has made contributions to the field of
machine learning and automatic speech reading. Current interests
include an open source project that will enable non-programmers to
develop Web applications and methods for enriching computer mediated
communication.
____________
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.
and
news://nntp.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.
____________