[Date Prev][Date Next][Date Index]
CSLI Calendar, Monday, 21 July 2003, vol. 18:43
CSLI CALENDAR OF PUBLIC EVENTS
______________________________________________________________________
21 July 2003 Stanford Vol. 18, No. 43
______________________________________________________________________
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 22 JULY 2003 TO 1 AUGUST 2003
TUESDAY, 22 JULY 2003
11:00am Music 319: CCRMA Hearing Seminar
CCRMA Ballroom, The Knoll
"Missing-data speech recognition as a paradigm for CASA"
Dan Ellis
Columbia University
http://mambo.ucsc.edu/psl/ccrmas/ccrmas.html
Abstract below
WEDNESDAY, 23 JULY 2003
11:00am MS&E 237: Progress in Worldwide Telecommunications
McCullough 115
"ITU Standardization and its New Environment"
Houlin Zhao
Director, Telecommunication Union (ITU), Geneva Switzerland
http://www.itu.int/ITU-T/tsb-director/zhao_cv.html
For questions contact ivenak@stanford.edu or samchiu@stanford.edu
3:00pm NASA Ames Panel Discussion
Ames Visitor Center Meeting Room (N223), NASA Ames Research Center
(not on Stanford and visitors badge not required)
"New Roles for Conferences, Journals, and Online Publications in
Computing and Information Sciences: Making Research Results Timely,
Authoritative, and Widely Accessible"
Abstract below
4:00pm SRI AI Seminar Series
EJ228, SRI International
"Mental Imagery, Language, and Gesture:
Video Access to Human Communication"
Francis Quek
Vision Interfaces and System Laboratory, Computer Science &
Engineering Department, Wright State University
http://www.ai.sri.com/seminars/
Abstract below
THURSDAY, 24 JULY 2003
1:30pm SRI AI Seminar Series
EJ228, SRI International
"Knowledge Pad: Status and Findings from Work So Far"
Victoria Bellotti
PARC
http://www.ai.sri.com/seminars/
Abstract below
4:00pm PARC Forum
George Pake Auditorium at PARC
"Internet Search: Past, Present and Future"
Jan Pedersen
Chief Scientist, AltaVista
http://www.parc.com/forum/
Abstract below
THURSDAY, 31 JULY 2003
11:00am Music 319: CCRMA Hearing Seminar
CCRMA Library, The Knoll
"Music Classification and Similarity Measurements"
Adam Berenzweig
Columbia University
http://mambo.ucsc.edu/psl/ccrmas/ccrmas.html
____________
Stanford Blood Center status: critical shortage of O+ and O-; shortage
of A+, A-, B+, and B-. For an appointment:
http://bloodcenter.stanford.edu/ or call 650-723-7831. It only takes
an hour of your time.
____________
ANNOUNCEMENT
I don't normally due an early calendar but several events arrived in
my mailbox that would be missed if I waited till Wednesday.
____________
MUSIC 319: CCRMA HEARING SEMINAR
on Tuesday, 22 July 2003, 11:00am
CCRMA Ballroom, The Knoll
http://mambo.ucsc.edu/psl/ccrmas/ccrmas.html
I'm happy to announce that Dan Ellis (Columbia) will be at CCRMA on
*Tuesday* for a special Hearing Seminar.
Dan works on computational auditory scene analysis (CASA) and its
application to speech recognition problems. He's a regular at the
Hearing Seminar.
Computational auditory scene analysis (CASA) endeavors to understand
the acoustic world around us. What is a sound object? How do we
separate the myriad of sounds we hear into multiple distinct objects?
One of the most exciting developments in CASA is a new attempt to unify
bottom-up and top-down grouping cues. Many groups have investigated
low-level perceptual grouping cues such as common harmonicity and common
onsets. Dan will talk about a new effort to also group sounds with
high-level principles such as language (and speaker identification.) Dan
(and his colleagues) group distinct portions of the spectrum into a small
number of groups using the low-level cues. Then they use a speech
recognizer to figure out which groups belong together and represent a
single (speech) object.
Bring your favorite auditory scene analysis system to CCRMA on
*Tuesday* and we'll talk about how it works!
- Malcolm
"Missing-data speech recognition as a paradigm for CASA"
Dan Ellis
Laboratory for Recognition and Organization of Speech and Audio (LabROSA)
Electrical Engineering Dept., Columbia University
http://labrosa.ee.columbia.edu/
(Joint work with Jon Barker, University of Sheffield)
A machine simulation of human auditory perception must be able to
recognize and classify individual sound sources. The most successful
technique for sound classification is the statistical pattern
recognition approach employed in speech recognizers; however, in most
practical cases, this approach assumes that the entire (monaural)
signal represents the source to be classified. Realistic
`cocktail-party' scenes, composed of multiple, overlapping sources
with comparable energies, do not come close to meeting this
assumption.
A more workable assumption is to treat each time-frequency cell as
representing a single source, but to use missing-data techniques to
perform recognition using only a subset of the cells. This precludes
the use of cepstral features (which depend on every frequency
component), but is otherwise practical. The problem then becomes
finding the `present data mask' that indicates which cells are to be
considered during classification of a particular source.
This talk presents a system based on these principles, with
applications both to speech recognition in dynamic, noisy backgrounds,
and to non-speech sounds such as alarms that can occur at very poor
signal-to-noise ratios.
____________
NASA AMES TALK
on Wednesday, 23 July 2003, 3:00pm - 4:00pm
Ames Visitor Center Meeting Room (N223) (NASA Ames Research Center)
Visitors Badges are not required for this part of the Center
Note this is not at Stanford
"New Roles for Conferences, Journals, and Online Publications in
Computing and Information Sciences: Making Research Results Timely,
Authoritative, and Widely Accessible"
Panelists (and their areas of research):
Willem Visser, RIACS/NASA Ames, Ph.D. (Manchester)
automated software engineering - validation and verification
Charlotte Linde, NASA Ames, Ph.D. (Columbia)
work practice analysis - organizational narrative
Jeremy Frank, NASA Ames, Ph.D. (UC Davis)
autonomy and robotics - planning and scheduling
Moderator:
Gregory Aist, RIACS/NASA Ames, Ph.D. (Carnegie Mellon)
natural language processing - intelligent tutoring systems
Agenda:
Introduction and Overview of the Challenge
Panel Discussion
Summary
Open Discussion with Audience
Scientific publications take many forms: informal "white papers",
technical reports, workshop and conference articles, invited or
refereed journal articles, book chapters, books, software/databases,
and websites.
What are the various merits of each of these forms of publication? How
do (or should) they factor into the evaluation of working scientists
for hiring, retention, and promotion? How are various publications
viewed by peer groups as shown in the awarding of grants and the
conferring of awards? These are perennial issues in science.
Computer scientists often face particular challenges since many
prestigious venues are in fact conferences (e.g. SIGGRAPH). While
many CS conferences are stringently reviewed, conference publications
are sometimes devalued by larger review bodies (e.g. tenure review,
grant review panels) relative to journal articles. Unfortunately, the
publication timeframe of some journals may not support the appearance
of new results in a timely manner.
This panel will address these questions drawing on the publication
practices of the panelists, and engage the audience in a discussion of
the role of various forms of publications, both now and in the future.
Finally, we'll touch briefly on issues of visibility to upper
corporate, government, and academic management, and outreach to the
public.
____________
SRI AI SEMINAR SERIES
on Wednesday, 23 July 2003, 4:00pm - 5:30pm
EJ228, SRI International
http://www.ai.sri.com/seminars/
"Mental Imagery, Language, and Gesture:
Video Access to Human Communication"
Francis Quek
Vision Interfaces and System Laboratory,
Computer Science & Engineering, Wright State University
Much video data involves recording of humans engaged in communication.
One may loosely classify the venues of such communications as meetings
with varying degrees of formality for such purposes as planning,
conflict resolution, negotiation, collaboration, confrontation,
information exchange, and gossip. We argue that the understanding of
human multimodal communicative behavior, and how witting or unwitting
visual displays relate to such communication is key to any approach to
the analysis of such data. We need to address two questions: how do we
bridge video and audio processing with the realities of human
multimodal communication, and how information from the different modes
may be fused. One path from multimodal behavior to language is bridged
by the underlying mental imagery. This visuospatial imagery, for a
speaker, relates not to the elements of syntax, but to the units of
thought that drive the expression (vocal utterance and visible
display). The basic idea is that mental imagery is integral to
language production, and non-verbal behavior (gesture, gaze, facial
expression) informs us of this imagery. Hence, we have a handle to
extract information on human discourse from video. The question
becomes what computable features of behavior are informative about
imagery and the organization of the discourse. We present the
Catchment Feature Model (CFM) our two key questions. We motivate the
CFM from psycholinguistic research, and present the Model. In contrast
to 'whole gesture' recognition, the CFM applies a feature
decomposition approach that facilitates cross-modal fusion at the
level of discourse planning and conceptualization. We shall discuss
the CFM-based experimental framework, and cite concrete examples of
Catchment Features (CF).
About the speaker: Francis Quek is currently an Associate Professor in
the Department of Computer Science and Engineering at the Wright State
University. He is director of the Vision Interfaces and Systems
Laboratory (VISLab) which he established for computer vision, medical
imaging, vision-based interaction, and human-computer interaction
research. He performs research in multimodal verbal/non-verbal
interaction, vision-based interaction, facial modeling, multimedia
databases, medical imaging, collaboration technology, computer vision,
human computer interaction, and computer graphics.
____________
SRI AI SEMINAR SERIES
on Thursday, 24 July 2003, 1:30pm
EJ228, SRI International
http://www.ai.sri.com/seminars/
"Knowledge Pad: Status and Findings from Work So Far"
Victoria Bellotti
Senior Member of Research Staff, PARC
PARC is currently engaged to conduct research towards the definition
of a DARPA Broad Area Announcement calling for research into the
design of a task-management tool called the Knowledge Pad (Kpad for
short). The announcement will specify the requirements for Kpad and
the evaluation criteria for the design efforts. The idea for the tool
is to use AI techniques leveraging task ontologies to provide support
for managing to-dos and executing tasks. Ongoing fieldwork is helping
to ground the requirements for Kpad in real task-management practices
and has exposed some questionable assumptions in the original vision
of how people manage their tasks and how Kpad might support them. I
will outline the fieldwork so far and its implications for Kpad.
____________
PARC FORUM
on Thursday, 24 July 2003, 4:00pm - 5:00pm
George Pake Auditorium, PARC
http://www.parc.com/forum/
"Internet Search: Past, Present and Future"
Jan Pedersen
Chief Scientist, AltaVista
First, the talk will survey the short history of Internet Search
Engines, from first generation systems that indexed a few million
pages to today's very sophisticated services that access billions of
documents in less than a second. Second, it will describe the
economics that drive the evolution of Internet Search services and the
analyze their recent business success. Finally, the talk closes with
a discussion of particular technologies responsible for the high
performance of these services and speculate on future technology
directions.
About the speaker: Jan Pedersen is currently Chief Scientist,
AltaVista. In the early 90s he was area manager of QCA at PARC.
____________
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
____________