[Next][Index][Thread]
CSLI Calendar, 23 September 1998, vol. 14:1
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
______________________________________________________________________
23 September 1998 Stanford Vol. 14, No. 1
______________________________________________________________________
A weekly publication of the
Center for the Study of Language and Information (CSLI)
Stanford University, Ventura Hall, Stanford, CA 94305-4115
____________
ACTIVITIES DURING 23 SEPTEMBER TO 2 OCTOBER 1998
THURSDAY, 24 SEPTEMBER
11:00am CCRMA Hearing Seminar
CCRMA Library
Evaluations of HRTF filters
Jyri Huopaniemi
Nokia, and formerly CCRMA and HUT in Finland
Abstract below
4:00pm Xerox PARC Forum
George Pake Auditorium, Xerox PARC
Designing Business Redux
Clement Mok
Studio Archetype
Host: Scott Minneman
Abstract below
4:15pm Mathematics Colloquium
Room 380:380W
On some connections between Hilbert's seventh and twelfth
problems
Gisbert Wustholz
ETH Zurich and MSRI
http://math.stanford.edu/html/colloquium.html
4:30pm Seminar on Computational Learning and Adaptation (SCLA)
Terman 332
Learning From What You DON'T Observe
Professor Ross Shachter
Engineering, Economic Systems & Operations Research
Stanford University
(special joint session with Decision Analysis Seminar)
Abstract below
FRIDAY, 25 SEPTEMBER
12:30pm Seminar on People, Computers, and Design
Gates B03 (NEC classroom)
Computer-Related Design at the Royal College of Art
Gillian Crampton Smith
Royal College Of Art, London
http://www-pcd.stanford.edu/cs547/
2:00pm International Computer Science Institute
Main Lecture Hall at ICSI, 1947 Center Street, Berkeley
AVALANCHE - Simulating Value Chain Coordination with
Artificial Life Agents
Boris Padovan
University of Freiburg
Abstract below
3:15pm Computer Science Infolab Seminar
Gates B-12
Rethinking Information Dissemination and Use: New Models
for Documents, Collaboration and Information Access
Robert Wilensky
UC Berkeley
Abstract below
WEDNESDAY, 30 SEPTEMBER
4:15pm AI-Vision-Robotics Division Colloquium
Gates 104
To be announced
http://www-formal.stanford.edu/aicolloq/
THURSDAY, 1 OCTOBER
4:00pm CSLI Talk
Ventura 17
Logical Pluralism
Greg Restall
Macquarie University
Abstract below
4:15pm Seminar on Computational Learning and Adaptation (SCLA)
Cordura 100 (note room change)
Predicting Lifetimes in Dynamically Allocated Memory
David Cohn
Adaptive Systems Group, Harlequin, Inc.
Abstract below
4:15pm Computer Science Talk
Gates 104
Programming Shared-Memory Multiprocessors Using the Cilk
Multithreaded Language
Charles E. Leiserson
MIT
Abstract below
FRIDAY, 2 OCTOBER
12 noon Logic Lunch
Room 380:383N
Costing Non-Classical Approaches to the Paradoxes Time
and Place
Greg Restall
Macquarie University
Abstract below
12:30pm Seminar on People, Computers, and Design
Gates B03 (NEC classroom)
A Human-Centered architecture for an Interactive
Workspace
Terry Winograd
Stanford Computer Science
http://www-pcd.stanford.edu/cs547/
3:15pm Philosophy Department Colloquium
Bldg. 90:92Q (Philosophy)
The Pragmatic Theory of Causation
Kyle Stanford
Assistant Professor, UC Irvine
http://www-philosophy.stanford.edu/general/colloqs.html
3:15pm Computer Science Infolab Seminar
Gates B-12
Building the InfoBus:
A Review of Technical Choices in the Stanford Digital
Library Project
Andreas Paepke
Stanford InfoLab
Abstract below
____________
NOTES
The new academic year is upon us and time for the start of the next
volume of the CSLI Calendar.
CSLI had an interesting Summer; we were flooded by a broken water pipe
in the attic in early July (approx. 150,000 gallons of water,
according to Stanford Facilities, took a short cut through the
ceilings of several offices and the networking closets). Many thanks
to V&M Restoration which got us running again within 5 days though we
still have a few finishing touches to go even now. See
http://www-csli.stanford.edu/csli/flood/ for some pictures.
____________
CCRMA HEARING SEMINAR
on Thursday, 24 September 1998, 11:00am
CCRMA Library, The Knoll
http://www-ccrma.stanford.edu/CCRMA/Events/Events.html
Objective and Subjective Evaluation of
Head-Related Transfer Function Filter Design
(also to be presented at the
AES 105th Convention, San Francisco, Sept. 26-30, 1998)
Jyri Huopaniemi
Nokia Research Center
Helsinki, Finland
mailto:jyri.huopaniemi@research.nokia.com
(the work was carried out while the author was a
visiting scholar at CCRMA, Stanford University)
In this presentation, the problem of modeling head-related transfer
functions (HRTFs) is addressed. Traditionally, HRTFs are approximated
in real-time applications using minimum-phase reconstruction and
various digital filter design techniques, yielding FIR or IIR
structures.
In this work, binaural auditory modeling has been applied to HRTF filter
design analysis, and design methods have been compared from the auditory
perception point of view. This paper presents applicable perceptually valid
smoothing and filter design techniques and discusses listening test results
for localization and timbre degradation using individualized HRTFs.
____________
XEROX PARC FORUM
on Thursday, 24 September 1998, 4:00pm - 5:00pm
George Pake Auditorium, Xerox
http://www.parc.xerox.com/ops/projects/forum/
"Designing Business Redux"
Clement Mok
Studio Archetype
Rules are changing and business and marketing plans seems to be
revised on an on-going basis. How does an organization keep focus
without being swept into trends-du-jour? Design Business Redux is a
presentation on how a culture of engineering and design can co-exist
to create better value for your clients and for your business during
this period of madness.
Biography: Clement Mok is the founder of Studio Archetype and is the
Chief Creative Officer of Sapient. He has also started two successful
software product companies: CMCD (a CD publishing group that produced
the award-winning "Visual Symbols Library") and NetObjects (an
internet software company that produced "NetObjects Fusion"). Clement
has received numerous design awards over the years, including being
named one of ID Magazine's forty most influential designers in 1994.
His recent book, "Designing Business: Multiple Media, Multiple
Disciplines," won the Financial Times Best How-To Business Book of the
Year for 1997.
____________
SPECIAL JOINT SESSION OF DECISION ANALYSIS SEMINAR AND
SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION
on Thursday, 24 September 1998, 4:30pm to 5:45pm
Terman 332
http://www-csli.stanford.edu/cll/scla.html
Learning From What You DON'T Observe
Professor Ross Shachter
Engineering, Economic Systems & Operations Research
Stanford University
The process of diagnosis involves learning about the state of a system
from various observations of symptoms or findings about the system.
Sophisticated Bayesian algorithms have been developed to revise and
maintain beliefs about the system as observations are made.
Nonetheless, diagnostic models have tended to ignore some common sense
reasoning exploited by human diagnosticians. In particular, one can
learn from which observations have NOT been made, in the spirit of
conversational implicature.
There are two concepts that we describe to extract information from
the observations not made. First, some symptoms, if present, are more
likely to be reported before others. Second, most human diagnosticians
and expert systems are economical in their data-gathering, searching
first where they are more likely to find symptoms present. Thus, there
is a desirable bias toward reporting symptoms that are present. We
develop a simple model for these concepts that can significantly
improve diagnostic inference.
(joint work with Mark A. Peot)
____________
BERKELEY INTERNATIONAL COMPUTER SCIENCE INSTITUTE
on Friday, 25 September 1998, 2:00 - 3:00 p.m.
Main Lecture Hall, ICSI, 1947 Center Street, Sixth Floor, Berkeley
http://www.icsi.berkeley.edu/talks/
AVALANCHE -
Simulating Value Chain Coordination with Artificial Life Agents
Boris Padovan
University of Freiburg
mailto:padovan@telematik.iig.uni-freiburg.de
The increasing use of information technology within and between
companies yields to changes in the predominant coordination
mechanisms. On the one hand it is argued that we witness an overall
shift towards market-like coordination of business transactions,
whereas others argue that we face more hierarchy-like coordination. In
this presentation a multi-agent system is described that globally
coordinates a multiple step value chain either way by using local
optimization rules of self-interested autonomous agents. In further
step it will be analyzed how changes in setup parameters (e.g.
transaction costs, co-operational behavior) will effect the
predominant coordination mechanism among the artificial life agents.
____________
COMPUTER SCIENCE INFOLAB SEMINAR
on Friday, 25 September 1998, 3:15pm - 4:30pm
Gates B12
http://www-db.stanford.edu/dbseminar/dbseminar.html
Rethinking Information Dissemination and Use:
New Models for Documents, Collaboration and Information Access
Robert Wilensky
UC Berkeley
Information technology can provide for the enhancement of services of
the sort we have come to expect from libraries. In addition, it offers
the opportunity for new services beyond those associated with
traditional libraries. In our view, the full potential of the
technology will be realized when expanded beyond the collections
building, information organization, and user access services, to
embrace the entire lifecycle of information use, from creation and
publishing to access, use, annotation and collaboration.
In this talk, we discuss some elements that facilitate this vision.
"Multivalent Documents" is a new model of documents, which embodies a
network-centric rather than desktop metaphor. The multivalent document
model is (i) highly open, meaning that is supports an open-ended
variety of document formats and functions, (ii) highly extensible,
meaning that it can be extended and customized in novel ways and to
meet particular user needs, and (iii) highly distributed, meaning that
components of a document may exist as separate networked resources,
which are combined dynamically into a coherent documents. A
particularly attractive aspect of the model is the manner in which it
supports "spontaneous collaboration", the ability of a user to
annotate web pages, scanned images, and other networked, resources for
which that user has no privileged relation. We will demonstrate
prototypes operating on a variety of document types, and on geographic
information data.
Multivalent documents address some issues in manipulating on-line
resources. Finding those resources is still problematic. "Automatic
content analysis" is the set of techniques for analyzing the content
of information objects so as to facilitate their subsequent access. We
present some recent developments in this area for accessing document
images, photographs, and text.
The work described is being conducted within the UC Berkeley Digital
Library Project, a multiterabyte collection of multimedia information
pertaining to the California environment.
Biography: Robert Wilensky received his B.A. and his Ph.D. from Yale
University. In 1978, he joined the faculty of the University of
California at Berkeley, where he is now Professor in the Division of
Computer Science, and in the School of Information Management and
Systems.
He has served as Chair of the Computer Science Division, the director
of BAIR, the Berkeley Artificial Intelligence Research Project, and
the director of Berkeley Cognitive Science Program.
Professor Wilensky has published numerous articles and books in the
area of artificial intelligence, planning, knowledge representation,
natural language processing, computer programming and digital
information systems. He is currently Principal Investigator of UC
Berkeley's Digital Library Project.
____________
CSLI TALK
on Thursday, 1 October 1998, 4:00pm
Ventura Hall, Room 17
http://www-csli.stanford.edu/csli/Coglunch/
Logical Pluralism
Greg Restall
Macquarie University
Coauthor: JC Beall, University of Tasmania
http://www.mq.edu.au/~phildept/staff/grestall/files/pluralism.ps
A widespread assumption in contemporary philosophy of logic is that
there is one true logic, that there is one and only one correct answer
as to whether a given determinate argument is deductively valid. We
reject this account of logic, and in this paper we propose an
alternative view, which we call logical pluralism. According to
logical pluralism, there is not exactly one true logic; there are
many. There is not always a single answer to the question "is this
argument valid?"
This paper motivates logical pluralism by examining debates within
accounts of classical first-order logic, and between classical
first-order logic and two of its well-known rivals, relevant and
intuitionistic logic. We argue that debates between these "rival"
logics---and between different accounts of the one logic---motivates a
pluralist account of logical consequence.
We end the paper by defending logical pluralism against a number of
different objections to the view.
____________
SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION (SCLA)
on Thursday, 1 October 1998, 4:15pm to 5:30pm
Cordura 100
http://www-csli.stanford.edu/cll/scla.html
Predicting Lifetimes in Dynamically Allocated Memory
David A. Cohn
Adaptive Systems Group
Harlequin, Inc.
mailto:cohn@harlequin.com
I will describe an application of machine learning to computer memory
management. In most programming languages, a program requests blocks
of memory from the runtime system when needed and explicitly frees
them up after use. Typically, the runtime system handles all of these
requests uniformly, without any regard to how or how long the
requested block will be used. In memory-intensive applications,
programmers frequently use runtime profiles to analyze their programs'
behavior and write special purpose memory management routines tuned
for their purposes. Machine learning methods offer the opportunity to
automate this process.
Barrett and Zorn used a simple lifetime predictor to explore this
possibility on a variety of computer programs. I will describe an
approach which uses decision trees to do lifetime prediction on the
same programs and shows significantly better prediction. Our method
also has the advantage that during training we can use a large number
of features and let the decision tree automatically choose the
relevant subset.
(joint work with Satinder Singh)
____________
COMPUTER SCIENCE TALK
on Thursday, 1 October 1998, 4:15pm
Gates 104
Programming Shared-Memory Multiprocessors
Using the Cilk Multithreaded Language
Charles E. Leiserson
MIT Laboratory for Computer Science
Cambridge, Massachusetts
Cilk is a language being developed in the MIT Laboratory for Computer
Science with the goal of making parallel programming easy. Cilk
minimally extends the C programming language to allow interactions
among computational threads to be specified in a simple and high-level
fashion. Cilk's provably efficient runtime system dynamically maps a
user's program onto available physical resources using a
"work-stealing" scheduler, freeing the programmer from concerns of
communication protocols and load balancing. In addition, Cilk
provides an abstract performance model that a programmer can use to
predict the multiprocessor performance of his application from its
execution on a single processor. Not only do Cilk programs scale up
to run efficiently on multiple processors, but unlike existing
parallel-programming environments, such as MPI and HPF, Cilk programs
"scale down": the efficiency of a Cilk program on one processor rivals
that of a comparable C program.
In this talk, I will provide a brief tutorial on the Cilk language. I
will explain how to program multithreaded applications in Cilk and how
to analyze their performance. I will illustrate some of the ideas
behind Cilk using the example of MIT's championship computer-chess
programs, *Socrates and Cilkchess. I will also briefly sketch how the
software technology underlying Cilk works.
See <http://theory.lcs.mit.edu/~cilk> for more background on Cilk and
to download the Cilk-5.2 manual and software release.
____________
LOGIC LUNCH
on Friday, 2 October 1998, 12 noon
Math Corner 380:383N
http://www-logic.stanford.edu/seminars.html
Costing Non-Classical Approaches to the Paradoxes
Greg Restall
Macquarie University
The paradoxes of `self-reference' (featuring semantic notions such as
truth, reference, properties and predication, as well as non-semantic
ones such as set, class or number) are known to exploit a few critical
logical moves. Non-classical solutions to these paradoxes reject one
or other of these moves. For example, Paraconsistent/Dialetheist
solutions reject the inference from a given contradiction to an
arbitrary conclusion. Contraction-free and ``gappy'' solutions reject
excluded middle and the inference of contraction (from A->(A->B) to
infer A->B).
In this paper, I explore the cost of non-classical accounts of the
paradoxes. In particular, I show that the paradoxes strike at the
heart of our logic of propositions, and that a non-classical theory
need be non-classical in more than just the logic of negation and
implication. In particular, I show that given
1. A transitive relation of entailment.
2. Conjunction and disjunction as distributive lattice operations.
3. A technique of diagonalisation or self-reference and a means
to form paradoxical propositions.
4. Infinitary disjunction. (If X is a class of propositions, then
X has a least upper bound, the disjunction of the propositions in
X.)
every proposition is derivable. No assumptions about negation or
implication are necessary. The paper closes with a discussion of some
of the consequences of this formal result.
____________
COMPUTER SCIENCE INFOLAB SEMINAR
on Friday, 2 October 1998, 3:15pm - 4:30pm
Gates B12
http://www-db.stanford.edu/dbseminar/dbseminar.html
Building the InfoBus:
A Review of Technical Choices in the Stanford Digital Library Project
Andreas Paepcke
Stanford University
Our approach to digital library infrastructure is based on the premise
that DLs will not just be online catalogs and collections, but that
they will be made up of geographically wide-spread, interoperating
services that support users in their tasks. Our infrastructure, called
the 'InfoBus' is based on CORBA distributed object technology. We
build proxy objects to provide uniform interfaces to online services
that present different interaction models and access protocols. For
this talk, we will select several aspects of our digital library, and
discuss the tradeoffs we made. First, we explain how we distributed
user interface functionality among clients and services. Second, we
present our experience with an object-centric information retrieval
protocol we developed for the project. Third, we explain how metadata
plays an important role in DL interoperability, and why existing
metadata representations fall short when used for digital libraries.
Finally, we will explain how user traditions and expectations have
impacted our designs at a deeply technical level.
____________
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
mailto: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/csli/.
For maps to the Stanford University campus see
http://www.stanford.edu/home/visitors/maps.html.
____________