[Prev][Next][Index]

CSLI Calendar, 26 Oct 1995, vol.11:05




        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
______________________________________________________________________________

26 October 1995                   Stanford                      Vol. 11, No. 5
______________________________________________________________________________

      A weekly publication of the Center for the Study of Language and
Information (CSLI), Stanford University, Ventura Hall, Stanford, CA 94305-4115
                               ____________

           CSLI ACTIVITIES DURING 25 OCTOBER -- 3 NOVEMBER 1995

  WEDNESDAY, 25 OCTOBER 
         3:15 - Philosophy of Computation Seminar
                Ventura Hall, Room 17
                Formal Symbol Manipulation (Part 2)
                Brian Cantwell Smith, Xerox PARC and Stanford Philosophy
                Abstract below

  THURSDAY, 26 OCTOBER
        10:00 - STASS Seminar
                Cordura Hall, Room 100
                The Reflexive/Incremental Approach to Information, Cognition,
                Language, and Action: Cognitive Content
                John Perry and David Israel
                Abstract below

        12:00 - Cognitive Science Lunch
                Cordura Hall, Room 100
                Making Robots Conscious of their Mental States
                John McCarthy, Stanford Computer Science
                Abstract below

         4:15 - SSP Forum
                Building 60, Room 61-F
                Harnessing Automated Inference to Enhance the
                Computer-Human Interface
                Eric Horvitz, Microsoft Research
                Abstract below

         7:30 - Phonology Workshop
                Building 460, Room 146
                Building Blocks for a Numerical Model of Intonation
                Arman Maghbouleh, Stanford University   
                Abstract below

  FRIDAY, 27 OCTOBER
        12:00 - Logic Lunch
                Building 380, Room 383-N
                Safe Stratified Datalog With Integer Order Does Not
                Have Syntax
                Alexei P. Stolboushkin and Michael A. Taitslin
                Abstract below

        12:30 - HCI Seminar
                Skilling Auditorium, SITN Channel E1
                The Wheel of Life, a Transformational Theater Piece
                Larry Friedlander, Stanford English
                Abstract below

         3:15 - Philosophy Colloquium
                Encina Hall, Room 423
                The Cunning of Trust
                Philip Petit, Australian National University
                
         3:30 - Linguistics Colloquium
                Building 460, Room 146
                A New Headed Morphological Structure in English (and the
                Demise of a Mythical Contraction Rule)
                Geoffrey K. Pullum, UC Santa Cruz Linguistics
                Abstract below

  MONDAY, 30 OCTOBER
         2:00 - Semantics Workshop
                Cordura Hall, Room 100
                What is a Possible Structural Argument?
                Dieter Wunderlich, Duesseldorf
                Abstract below

  TUESDAY, 31 OCTOBER
         4:00 - Seminar on Computational Learning and Adaptation
                Medical School Office Building, Room 303
                Conditional Mixture Models
                H. Scott Roy, Heuristicrats Research, Inc.
                Abstract below

         7:00 - SSP Film Series
                Cubberley Hall, Room 128
                Today's Innovators and Tomorrow's Technologies (1992)
                Abstract below

  WEDNESDAY, 1 NOVEMBER
         3:15 - Philosophy of Computation Seminar
                Ventura Hall, Room 17
                Effective Computability and Recursion Theory
                Brian Cantwell Smith, Xerox PARC and Stanford Philosophy
                Abstract below

  THURSDAY, 2 NOVEMBER
        10:00 - STASS Seminar
                Cordura Hall, Room 100
                The Reflexive/Incremental Approach to Information, Cognition,
                Language, and Action: Philosophy of Language
                John Perry and David Israel
                Abstract below

        12:00 - Cognitive Science Lunch
                Cordura Hall, Room 100
                Title to be announced
                Fred Dretske, Stanford Philosophy

         3:00 - CSLI Open House
                Cordura Hall
                Demos and Discussions of CSLI Interface Lab and
                Cognitive Science projects
                [http://www-csli.stanford.edu/csli/projects.html]

  FRIDAY, 3 NOVEMBER
        12:30 - HCI Seminar
                Skilling Auditorium, SITN Channel E1
                Interacting with the Digital Library
		Terry Winograd
		Abstract below

         3:15 - Philosophy Colloquium
                Encina Hall, Room 423
                The Ising Model: A Formal Example of Scientific Reduction
                Yair Guttmann, Stanford Philosophy

         3:30 - Linguistics Colloquium
                Building 460, Room 146
                Constructing Inflectional Paradigms
                Dieter Wunderlich, Heinrich-Heine-Universitaet, Duesseldorf

                               ____________

The CSLI Calendar appears on Wednesday of each week throughout the academic
year.  Announcements, abstracts, and other information to appear in the
Calendar can be submitted by e-mail to <incalendar@csli.stanford.edu>.

Further information about CSLI and past issues of the CSLI Calendar
are available on the Internet at URL <http://www-csli.stanford.edu/csli/>.
The Calendar is also posted each week to the <csli.bboard> newsgroup.

                               ____________
                                     
                    PHILOSOPHY OF COMPUTATION SEMINAR
                         on Wednesday, 25 October
                     3:15 p.m., Ventura Hall, Room 17
                   Formal Symbol Manipulation (Part 2)
                           Brian Cantwell Smith
                    Xerox PARC and Stanford Philosophy
                         [bcsmith@parc.xerox.com]

The reading for today's seminar is Volume I Chapter 4 of _The Middle
Distance_.  A course description can be found at [http://shr.stanford.edu/
BCSmith/phil395a.html]

                               ____________

                              STASS SEMINAR
                         on Thursday, 26 October
                    10:00 a.m., Cordura Hall, Room 100
      The Reflexive/Incremental Approach to Information, Cognition,
                 Language, and Action: Cognitive Content
                       John Perry and David Israel
                       Stanford Philosophy and SRI
                [john@csli.stanford.edu,israel@ai.sri.com]

As background, see Perry and Israel, "Fodor and Psychological Explanations."
In _Meaning in Mind_, edited by Barry Loewer and Georges Rey, 165--80 (Oxford:
Basil Blackwell, 1991).  Also relevant: Perry, "Self-Notions." _Logos_ 1990:
17--31.  Papers are available at [http://www-csli/users/john/phil.html].

                               ____________

                              CSLI COGLUNCH
                         on Thursday, 26 October
                    12:00 noon, Cordura Hall, Room 100
              Making Robots Conscious of their Mental States
                              John McCarthy
                        Stanford Computer Science
                          [jmc@cs.stanford.edu]

In AI, consciousness of self consists in a program having certain kinds of
facts about its own mental processes and state of mind.

We discuss what consciousness of its own mental structures a robot will need
in order to operate in the common sense world and accomplish the tasks humans
will give it.  It's quite a lot.

Many features of human consciousness will be wanted, some will not, and some
abilities not possessed by humans will be found feasible and useful.

We give preliminary fragments of a logical language a robot can use to
represent information about its own state of mind.

A robot will often have to conclude that it cannot decide a question on the
basis of the information in memory and therefore must seek information
externally.  Goedel's idea of relative consistency is used to formalize
non-knowledge.

Programs with the level of consciousness discussed in this article do not yet
exist.

Thinking about consciousness with a view to designing it provides a new
approach to some of the problems of consciousness studied by philosophers.
The advantage is that it focusses on the aspects of consciousness important
for intelligent behavior.

A description and schedule for the CogLunch series on consciousness can be
found at [http://www-csli.stanford.edu/csli/9495reps/coglunch.html].

                               ____________

                          SYMBOLIC SYSTEMS FORUM
                         on Thursday, 26 October
                    4:15 p.m., Building 60, Room 61-F
  Harnessing Automated Inference to Enhance the Computer-Human Interface
                               Eric Horvitz
                            Microsoft Research
        [http://www.research.microsoft.com/research/dtg/horvitz/]

I will first provide a brief overview of Microsoft Research with a focus on
projects of the Decision Theory Group.  Then, I will present work on the use
of decision-theoretic reasoning techniques to enhance user interfaces.  After
providing some background on the use of probability and utility in intelligent
reasoning systems, I will describe the Vista project, an effort I directed
before coming to Microsoft Research.  Vista is centered on building an
intelligent interface to assist NASA Mission Control flight engineers with
managing the complexity of information about Space Shuttle propulsion systems
in time-critical contexts.  The Vista system went online at the Mission
Control Center last year and is now used to support flight.  Moving from
special-purpose applications to personal computing, I will provide an overview
of related efforts to develop flexible and responsive interfaces for users of
personal computers.

ERIC HORVITZ is a Senior Researcher in the Decision Theory Group at Microsoft
Research and an Affiliate Associate Professor at the University of Washington.
He received his PhD and MD at Stanford University.  In his doctoral work, he
explored decision-theoretic methods for computing rational decisions under
constraints in computational resources.  His current work centers on the use
of decision-theoretic principles in the construction and operation of
automated reasoning systems.  His interests include the use of probability and
utility to solve problems in time-critical decision making, operating systems,
user interfaces, information retrieval, and medicine.

                               ____________
 
                    THE PHONOLOGY WORKSHOP AT STANFORD
                         on Thursday, 26 October
                    7:30 p.m., Building 460, Room 146
           Building Blocks for a Numerical Model of Intonation
                             Arman Maghbouleh
                           Stanford University
                        [arman@csli.stanford.edu]

Even though duration is one of the prime acoustic correlates of intonation, it
is not trivial to assign intonational function (e.g., emphasis) to observed
duration values because duration is crucially affected by many factors other
than intonation. For example, even though greater length is associated with
greater emphasis, one can not compare the segment durations in the words
"bead" and "bet" and assign greater emphasis to the longer word. Such simple
comparisons are not possible because it can be shown that given the same
emphasis, the two words will have differing durations. For example, it can be
shown that the /i/ in "bead" is inherently longer than the /e/ in "bet". That
is, the /i/ is pronounced with greater length than the /e/ in the same context
with same perceived emphasis. Furthermore, vowels before voiced stops such as
/d/ are pronounced with greater length than vowels before unvoiced stops such
as /t/. Therefore, the greater length in "bead" may be due to segmental
factors and not a reflection of greater emphasis. In order to isolate
intonation-related contributions to duration, one needs to know precisely how
the various factors interact.

The interaction of the various durational factors is of interest not only to
linguists but also to other speech scientists who have devised duration models
for use in speech synthesis and speech recognition applications. In this talk,
I will provide an overview of the various duration models, before describing
the implementation of a variant of the model described in van Santen 1992. The
new model was trained and tested on the Boston University Radio News Corpus
and on the average, the predicted duration values were 31 milliseconds off
>From the actual durations. The model was then modified for the task of
identifying points of emphasis given duration values, energy values, and
segmental context. The new model achieves 87% accuracy in locating pitch
accents. This model improves on previous work (Wightman & Ostendorf 1994) by
achieving slightly higher accuracy rates with many fewer training
requirements.

This work is part of a larger attempt to devise a numerical representation of
intonation which unlike current phonological representations can be easily
extracted from speech data and turned back into speech data without loss of
any of the nuances present in the original speech.

                              _____________

                               LOGIC LUNCH
                          on Friday, 27 October
                   12:00 Noon, Building 380, Room 383-N
   Safe Stratified Datalog With Integer Order Does not Have Syntax    
                    Alexei P. Stolboushkin (UCLA) and
           Michael A. Taitslin (Tver State University, Russia)

Stratified Datalog (Dtl) with integer (gap)-order is considered.  A
Dtl-program is said to be safe iff its bottom-up processing terminates on all
valid inputs.  We prove that safe Dtl-programs do not have effective syntax in
the sense that there is no recursively enumerable set S of safe Dtl-programs
such that any safe Dtl-programs is equivalent to a program in S.

                              _____________

                  SEMINAR ON HUMAN-COMPUTER INTERACTION
                          on Friday, 27 October
                     12:30 p.m., Skilling Auditorium
           The Wheel of Life, a Transformational Theater Piece
                            Larry Friedlander
                 Stanford English Literature and Theater
                    [friedland.l@applelink.apple.com]

In the Fall of 1993 Larry Friedlander, along with Professor Glorianna
Davenport and students at the MIT Media Lab, created an experimental
installation in the Villers theater at MIT called Wheel of Life: An
Interactive Transformational Environment.  This installation explored what
happens when interactive technology expands beyond the computer box and moves
into our daily environments.  Wheel of Life drew its techniques from the worlds
of theater, architectural design, cinema, and interactive computing to to
create a new kind of space, one that was magically responsive to the visitors'
presence and actions.  As visitors moved through the various environments, they
worked with partners who were seated outside of the installation at computer
workstations to transform the spaces physically and to create stories and
changing multimedia events.  Using technology to create complex narrative
spaces raises fascinating issues in the psychology of collaborative invention
and in the design of spaces shared by people and machines.  Friedlander will
discuss the implication of this move 'out of the box' for the future of
interface design and of interactive applications.

LARRY FRIEDLANDER, a professor of English Literature and Theater at Stanford
University, began working in multimedia design and applications in 1983 with
the Shakespeare Project.  In 1990 Friedlander formed the Interactive
Shakespeare Group with Professors Donaldson and Murray from MIT to further
explore this area.  He has developed many applications in theater and
education, and has worked in major research laboratories: at the Apple
Multimedia Lab, the MIT Media Lab, and the Mitsubishi Electronic Research
Laboratory among others.  He is also heavily involved in museum design and
planning, and is now advising the Museum of Scotland, a new national museum
due to open in Edinburgh in 1999.  He has done work with numerous other
institutions such as the Musee d'Orsay in Paris, and is currently Osher Fellow
at the San Francisco Exploratorium.

                               ____________

                          PHILOSOPHY COLLOQUIUM
                          on Friday, 27 October
                     3:15 p.m., Encina Hall, Room 423
                           The Cunning of Trust
                               Philip Petit
                      Australian National University

                               ____________

                          LINGUISTICS COLLOQUIUM
                          on Friday, 27 October
                    3:30 p.m., Building 460, Room 146
             A New Headed Morphological Structure in English
             (and the Demise of a Mythical Contraction Rule)
                            Geoffrey K. Pullum
                        UC Santa Cruz Linguistics
                          [pullum@cats.ucsc.edu]

For a quarter of a century there has been controversy in the literature of
generative grammar about alleged syntactic constraints on the contraction rule
that realizes certain sequences of verb plus infinitival "to" as portmanteau
forms like "wanna", "hafta", and "gonna".  Most notably, it has been claimed
that the distribution of such forms confirms claims that movement operations
in syntax leave traces.  It has occasionally been suggested that a better
account might be given in lexical or morphological terms, but no such
treatment has been fully elaborated, and the simple suggestion that the
contracted forms in question have been "lexicalized" has morphological
consequences that are thoroughly unacceptable.  In this talk, based on joint
research with Arnold Zwicky, I show how an analysis more seriously grounded in
morphological theory can provide a fully satisfactory theory of the syntax,
morphology, and phonology of the items involved.  All the syntactic evidence
that has ever been brought to bear on the issues becomes completely tractable
under this account, giving it a coverage that none of the syntactic proposals,
pro-trace or anti-trace, ever achieved.  And in addition, the morphological
and phonological puzzles about the forms in question also submit to satisfying
solutions.

                               ____________

                            SEMANTICS SEMINAR
                          on Monday, 30 October
                    2:00 p.m., Cordura Hall, Room 100
                 What is a Possible Structural Argument?
                            Dieter Wunderlich
                 Heinrich-Heine-Universitaet, Duesseldorf
                        [dieter@csli.stanford.edu]

Lexical Decomposition Grammar (LDG) tries to account for argument structure on
the basis of semantic decomposition, by which the meaning of a verb may be
decomposed into several more primitive predicates. A distinction is made
between the level of Semantic Form (SF) on which the logical-categorial
structure is minimally specified and the level of Conceptual Interpretation on
which all kinds of inferences may be added.

SF is regarded as a binarily branched structure in terms of logical types.
Conjunction is therefore asymmetric in SF, and, by convention, A & B is
structured as [A [& B]], with A being called the head predicate. The arguments
of A thus rank over those of B. In general, SF determines the ranking of
arguments, but in certain cases some segments of the argument hierarchy may be
reversed by lexical marking (which might be motivated by additional conceptual
reasons).

The crucial idea for argument linking (adopted from Kiparsky) is that the
argument positions are encoded by purely structural features, and that the
same set of features characterizes the structural cases.  As a consequence,
the medial argument of an unmarked 3-place verb like German geben 'give' is
linked to Dative and the lowest one to Accusative, while it is only for marked
verbs such as unterziehen 'subject to' that the lowest argument is linked to
Dative and the medial one to Accusative.

The question now is why in certain cases the medial argument is blocked from
being projected into syntax. Compare the resultative extension in (3) and the
locative alternation in (4) with the regular Dative verbs in (1) and (2).

        (1)  Sie entnahm der Dose einen Keks.
             'She took a cookie out of the box'
        (2)  Sie stahl dem Koch ein Ei.
             'She stole an egg from the cook'
        (3)  Sie trank die Flasche leer (*dem Bier/*von Bier).
             'She drank the bottle empty (*from beer)'
        (4)  Sie bespruehte die Wand (*der Farbe/mit Farbe).
             'She sprayed the wall (with paint)

Note that there is no obvious semantic reason why the liquid she is drinking
in (3) and the substance she is spraying in (4) cannot be expressed by a
structural case.

I will present a purely structural account by proposing the following
Restriction on Structural Arguments:

        (R)  An argument is structural only if it is either the lowest
             argument or (each of its occurrences) L-commands the
             lowest argument (with L for 'Lexical').
 
The working of (R) is shown by the minimal semantic representations for (1) to
(4) given in (5) to (8).

        (5)  TAKE(x,z) & BECOME NOT CONTAIN(y,z)
        (6)  TAKE(x,z) & BECOME NOT POSS(y,z)
             Both x and y L-command the lowest argument, that is z.
        (7)  DRINK(x,y) & BECOME P(z)
             y does not L-command z.
        (8)  SPRAY(x,y) & BECOME LOC(y, AT z)
             The first occurrence of y does not L-command z.

A corollary of (R) is that for every non-last member of an SF-conjunction no
argument except the highest one can be expressed as a structural argument in
syntax.

Furthermore, this account rules out all more elaborated representations such
as that in (9) for the resultative extension.

        (9)   CAUSE(DRINK(x,y), BECOME EMPTY(z))

The causal interpretation is ensured by means of a general Coherence
Principle rather than being part of the lexical entry.

                               ____________

             SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION
                         on Tuesday, 31 October
           4:00 p.m., Medical School Office Building, Room 303
                        Conditional Mixture Models
                               H. Scott Roy
                       Heuristicrats Research, Inc.

This talk describes how to use mixture models to predict discrete,
categorical, and continuous quantities.  These models are identical to those
that have been used with great success in unsupervised clustering, but with
one crucial difference: they are evaluated for the conditional probability
they assign to a single feature rather than the joint probability they assign
to all features.  Good models are therefore forced to classify the data in a
way that is useful for a single, desired prediction, rather than simply
identifying the strongest overall patterns in the data.  The approach gives
outstanding performance with clear probabilistic semantics and produces models
that are easy to understand.

                               ____________

                             SSP FILM SERIES
                          on Tuesday, 31 October
                   7:00 p.m., Cubberley Hall, Room 128
          Today's Innovators and Tomorrow's Technologies (1992)

This 56 minute video shows what students think is in the future for computer
technology and what they discover is the value of interdisciplinary learning.
The faculty and administration of various universities also talk about the
possible technologies in the near future and how they can affect teaching and
learning.

The Symbolic Systems Film Series showcases videos and films of general
cognitive science interest. Attendance at the films can substitute for
attendance at the Symbolic Systems Forum for students enrolled in SSP 10 for
one unit. All are welcome at these events. The showing of the videos is
followed by a discussion, and researchers who are knowledgeable about the
program's topic are urged join us in evaluating the program.

                               ____________

                    PHILOSOPHY OF COMPUTATION SEMINAR
                         on Wednesday, 1 November
                     3:15 p.m., Ventura Hall, Room 17
               Effective Computability and Recursion Theory
                           Brian Cantwell Smith
                    Xerox PARC and Stanford Philosophy
                         [bcsmith@parc.xerox.com]

                               ____________

                              STASS SEMINAR
                         on Thursday, 2 November
                    10:00 a.m., Cordura Hall, Room 100
      The Reflexive/Incremental Approach to Information, Cognition,
                 Language, and Action: Philosophy of Language
                       John Perry and David Israel
                       Stanford Philosophy and SRI
                [john@csli.stanford.edu,israel@ai.sri.com]

As background, see Perry, "Indexicals and Demonstratives."  In _Companion to
the Philosophy of Language_, edited by Robert Hale and Crispin Wright (Oxford:
Basil Blackwell, forthcoming).  Also relevant: Perry, "Reflexivity,
Indexicality and Names" (forthcoming).  Papers are available at
[http://www-csli/users/john/phil.html].

                               ____________

                              CSLI COGLUNCH
                         on Thursday, 2 November
                    12:00 noon, Cordura Hall, Room 100
                          Title to be announced
                               Fred Dretske
                           Stanford Philosophy
                       [dretske@csli.stanford.edu]

                               ____________

                             CSLI OPEN HOUSE
                         on Thursday, 2 November
                    3:00 p.m.--6:00 p.m., Cordura Hall

    Featuring demonstrations from CSLI's Interface Laboratory projects
        and poster talks about CSLI's Cognitive Science projects.

        Elegant automated parsing performed before your very eyes!
                         See no hands computing!
              Voice control of Macs, Suns, and SGI machines!
                 A computer controlled by eye movements!
  A computer controlled by brain activity: no hands, no voice, no eyes!

                   See hyper logicians do hyper proofs!
         Be conscious, and talk about it with real philosophers!
         Talk about it some more, and then talk about talk itself
                  with real linguists and psychologists!

                   Tell your friends! Tell your family!
                  Bring your friends! Bring your family!

                 Stretch your mind and fill your stomach!
                      Learn what CSLI is all about!
   See what your friends and colleagues do when they disappear to CSLI!

   And while you're at it, be part of the drawing for a free CSLI book!

                              _____________

		  SEMINAR ON HUMAN-COMPUTER INTERACTION
			  on Friday, 3 November
		     12:30 p.m., Skilling Auditorium
		   Interacting with the Digital Library
			      Terry Winograd
			Stanford Computer Science
			[winograd@cs.stanford.edu]

As information of widely diverse types and origins becomes a part of the
world's digital library, a key problem is the integration of heterogeneous
materials into a common platform.  This includes work on widely-applicable
formats, protocols, and representations, which can support services for
access, search, information transformation, display, archiving, and the like.
It also requires new tools for the user, which can bring a uniform conceptual
model to the full range of information tasks, for providing, organizing,
manipulating and viewing information.

This talk describes work in progress in the digital libraries project at
Stanford, in the area of interface and interaction.  I will describe the
general context of the research and present some examples of specific projects
and the issues that they are addressing.

TERRY WINOGRAD is Professor of Computer Science at Stanford.  He began his
career in computing in the field of Artificial Intelligence, specializing in
natural language understanding.  After rethinking the problems and potentials
of computational intelligence (see Winograd and Flores, Understanding
Computers and Cognition, Addison-Wesley, 1987), he left AI research and for
the last decade has worked in the area of human-computer interaction.  He
initiated the Project on People, Computers and Design and the program in
Human-Computer Interaction at Stanford.  His previous books include Usability:
Turning Technologies Into Tools (edited with Paul Adler, Oxford, 1992) and he
is the editor of a forthcoming book, Bringing Design to Software,
Addison-Wesley, 1996.

Winograd was a founder and past-president of Computer Professionals for Social
Responsibility and is on the USACM public policy committee, and the advisory
Board of the Association for Software Design , the Universal Access Project of
the World Institute on Disability.  He was a founder of Action Technologies,
and consults at Interval Research, which is one of the sponsors of the work
being presented in this talk.

                               ____________

			  PHILOSOPHY COLLOQUIUM
			  on Friday, 3 November
		     3:15 p.m., Encina Hall, Room 423
        The Ising Model: A Formal Example of Scientific Reduction
                              Yair Guttmann
                           Stanford Philosophy
                       [guttman@csli.stanford.edu]

                               ____________

                          LINGUISTICS COLLOQUIUM
                          on Friday, 3 November
                    3:30 p.m., Building 460, Room 146
                   Constructing Inflectional Paradigms
                            Dieter Wunderlich
                 Heinrich-Heine-Universitaet, Duesseldorf
                        [dieter@csli.stanford.edu]

                               ____________

                            1995 IAP TUTORIALS
                 Monday through Wednesday, 6--8 November
                      Cordura Hall, Stanford Campus
               KNOWLEDGE MILLENNIUM: NETWORKED LEARNING AND
                     COMMUNICATION THROUGH TECHNOLOGY
            [http://www-csli.stanford.edu/csli/tutorials.html]

Three days of tutorials on the Internet, Social Responses to Communication
Technology, and Automatic Speech Recognition.  Open to registered participants
only.  Registration information -- by phone: call 415-723-1224; by e-mail:
contact [mking@csli.stanford.edu]

MONDAY, 6 NOVEMBER:
The Internet: Agents, Just-in-Time Learning, and Distance Education

 The first day's tutorial covers a range of topics centered on the design
 and development of effective internet technologies:
  * Internet Agents -- software tools for finding and accessing
    information on the internet;
  * Distance Learning -- delivery of traditional lectures and educational
    material over the internet; new ways of participating in interactive
    study groups; organizing training material so that it can be
    found "on demand";
  * Professional Support Systems -- providing doctors with diagnostic
    information over the internet; professional teamwork on the internet.

 Speakers: Yoav Shoham, Marko Balabanovic, J. Marty Tenenbaum, Terry
 Winograd, William M. Detmar, Larry Leifer, George White

TUESDAY, 7 NOVEMBER:
Social Responses to Communication Technology

 The second day's tutorial is on "Social Responses to Communication
 Technology."  This theory states that individuals' interactions with
 computers, television, and other communication technologies are fundamentally
 social and natural.  That is, all people expect computers, televisions, and
 other technologies to obey a wide range of social rules, and all people use
 these rules in responding to these media.  Similarly, all individuals respond
 to pictures on the screen as if the objects they represent were actually
 present.  We will discuss a series of over 25 experiments that demonstrate
 that the real and mediated world are essentially the same.  We will discuss
 how our research on the following concepts inform the design of interfaces,
 multimedia, television, and all other media: politeness, praise and criticism,
 interpersonal distance, personality, agents, emotion, arousal, gender
 stereotyping, size and shape, motion, use of voice, fidelity, and synchrony.

 Speakers: Clifford Nass and Byron Reeves

WEDNESDAY, 8 NOVEMBER:
Automatic Speech Recognition

 The third day covers theoretical and practical aspects of Automatic Speech
 Recognition.  After nearly 30 years of gradual progress, ASR is now showing up
 everywhere, over the phone, on PCs, and soon on the internet.  The history,
 theory, current practice, and future of the ASR field is treated in this
 tutorial by several acknowledged authorities in the area who have both
 industrial and academic experience.

 Speakers: Bill Meisel, Elizabeth Macken, Neil Scott, George White, Brian Scott

REGISTRATION INFORMATION:
 By phone: call 415-723-1224
 By e-mail: contact mking@csli.stanford.edu

                               ____________