[Prev][Next][Index][Thread]

CSLI Calendar, 11 March 1998, vol. 13:24



   
     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
______________________________________________________________________

11 March 1998                   Stanford               Vol. 13, No. 24
______________________________________________________________________

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

             ACTIVITIES DURING 11 MARCH TO 20 MARCH 1998

WEDNESDAY, 11 MARCH
         4:15pm EE380: Computer Systems Laboratory Colloquium
                Gates B03 (NEC Auditorium)
                AltaVista: A Look Under the Hood
                Paul Flaherty
                DEC
                Abstract below

         8:00pm Wesson Lecture
                Jordan Hall 420:040
                The Ironic Construction of Civic America
                Theda Skocpol 
                Harvard University
                
THURSDAY, 12 MARCH
        12 noon CSLI CogLunch
                Cordura Hall, Room 100
                The Creation of Virtual Pain
                Howard Fields
                UCSF, Neurology
                
        3:00pm  Shimizu Talk 1
                CSL Commons #2230, Xerox PARC
                The Dual-Centered Model of the Self: An overview of the
                Theory of Ba and Basho
                Hiroshi Shimizu      
                Emeritus Professor of Pharmacology, University of Tokyo
                Founder and Director of The "Ba" Research Institute,
                Kanazawa Institute of Technology
                (If you are interested in attending, please contact Kim  
                Paulon (kpaulon@parc.xerox.com) at least one day ahead to
                arrange for escort. Please plan to arrive not later than
                2:45.)
                Abstract below

         4:00pm Xerox PARC Forum
                George Pake Auditorium, Xerox PARC
                Nanotechnology
                Dr. Ralph Merkle
                Xerox Palo Alto Research Center
                Abstract below

         4:15pm Seminar on Computational Learning and Adaptation (SCLA)
                Gates 100
                Learning Situation-Dependent Planning Knowledge from
                Uncertain Robot Execution Data
                Karen Zita Haigh
                Carnegie Mellon University, Pittsburgh, PA
                Abstract below
                
         7:30pm Phonology Workshop
                Margaret Jacks Hall, 460:146
                From Bumps to Phonology: A Model for Recognizing
                Intonational "phonemes" from Duration and Fundamental
                Arman Maghbouleh
                Department of Linguistics, Stanford University
                Abstract below
                
         8:00pm Wesson Lecture
                Jordan Hall 420:040
                Democracy Unravelled: America in the Late 20th Century
                Theda Skocpol
                Harvard University
                
FRIDAY, 13 MARCH
        all day Wilbur Knorr Memorial Conference
                http://www.stanford.edu/dept/HPS/Memorial.html 
                
        12 noon Logic Lunch
                Room 380:383N 
                Frege Structures for Partial Applicative Theories
                Reinhard Kahle
                University of Tuebingen, visiting Stanford
                Abstract below
                
        12:30pm Seminar on People, Computers, and Design
                Gates B01 (HP classroom)
                3DDI: 3D Direct interaction
                John Canny
                UC Berkeley
                Abstract below
              
SATURDAY, 14 MARCH
        all day Wilbur Knorr Memorial Conference
                http://www.stanford.edu/dept/HPS/Memorial.html 

MONDAY, 16 MARCH
         3:30pm Psychology Social Lab
                Jordan Hall 420:100
                Old Sexism In New Guise
                Douglas Hofstadter
                Indiana University
                 
         4:15pm Shimizu Talk 2
                Cordura 100
                Fusing Eastern and Western Models of the Self: The "Logic 
                of Self" and the "Logic of Science"
                Hiroshi Shimizu      
                Emeritus Professor of Pharmacology, University of Tokyo
                Founder and Director of The "Ba" Research Institute,
                Kanazawa Institute of Technology
                Abstract below
   
         4:30pm Stanford Digital Libraries Seminar
                Gates B08
                Active Security Models: An Approach to
                Enterprise-oriented and Self-administering Security in
                Distributed Collaborative Environments
                Roshan Thomas
                Odyssey Research Associates
                Abstract below

         7:30pm George and Sandra Forsythe Memorial Lecture
                Hewlett-Packard classroom (Gates B01)
                Rationality and Intelligence
                Professor Stuart Russell
                University of California, Berkeley
                Abstract below

TUESDAY, 17 MARCH
         2:00pm Shimizu Talk 3
                FX Palo Alto Laboratories
                Design Principles for Knowledge Systems from the
                Dual-Centered Perspective
                Hiroshi Shimizu      
                Emeritus Professor of Pharmacology, University of Tokyo
                Founder and Director of The "Ba" Research Institute,
                Kanazawa Institute of Technology
                (If you are interested in attending, please contact Livia
                Polanyi (polanyi@pal.xerox.com) at least one day in
                advance to get directions and arrange for escort.)
                Abstract below
                
         3:15pm Philosophy Department Colloquium
                Bldg. 200:305
                Laws and Induction
                Barry Loewer
                Professor, Rutgers University

         4:15pm Logic Seminar
                Room 380:381T
                Recent Results in Metapredicative Proof Theory
                Thomas Strahm
                University of Bern, Switzerland (visiting Stanford)
                Abstract below
   
         4:15pm George and Sandra Forsythe Memorial Lecture  
                Hewlett-Packard classroom (Gates B01)  
                Object Identification in a Bayesian Context
                Professor Stuart Russell
                University of California, Berkeley
                Abstract below
   
WEDNESDAY, 18 MARCH
         3:15pm ME297: Design Theory and Methodology Seminar
                Grammatical Generation and Optimizing Search in Early
                Design
                Professor Jonathan Cagan
                George Tallman and Florence Barrett Ladd Development
                Professor in Engineering
                Department of Mechanical Engineering,
                Carnegie Mellon University  
                (this may be a closed seminar, check the web page)
                http://cdr.stanford.edu/DD/Courses/me297/
                Abstract below
     
THURSDAY, 19 MARCH
        12 noon CSLI Talk
                Cordura Hall, Room 100
                Global Intelligence Project
                Prof. Masa Numao
                Tokyo Institute of Technology
                Abstract below
            
         2:15pm Semantics Workshop
                Margaret Jacks Hall 460:146
                Frank Veltman
                University of Amsterdam and the Department of Philosophy,
                Stanford University
                http://www-linguistics.stanford.edu/Linguistics/semgroup/
                
         4:15pm Seminar on Computational Learning and Adaptation (SCLA)
                Gates 100
                Learning Evaluation Functions for Optimization
                Justin Boyan 
                Carnegie Mellon University, Pittsburgh, PA
                Abstract below

FRIDAY, 20 MARCH
        12 noon Logic Lunch
                Room 380:383N
                Reductions of Theories of Countable Tree Ordinals to ID_1
                Sol Feferman
                Abstract below
                             ____________
                                   
                  EE380: COMPUTER SYSTEMS COLLOQUIUM
            on Wednesday, 11 March 1998, 4:15pm to 5:30pm
        NEC Auditorium (B03), Gates Computer Science Building
             http://www.stanford.edu/class/ee380/

                   AltaVista: A Look Under the Hood
                            Paul Flaherty
                                 DEC

Three years ago, a team of three researchers at Digital Equipment
Corporation combined lots of memory, lots of compute power, and some
innovative ideas about databases and the Internet to create (in four
months) a search engine called Alta Vista.  The primary goal was to
create a public showpiece for DEC's database and Internet performance
prowess.  However, the Law of Unintended Consequences seems to have
guaranteed an interesting life for what was originally a demo.  People
around the world have found all sorts of interesting uses for the
search engine, the company managed to put together a business model
for perpetuating the demonstration, and the spammers have made life
very interesting for the production team.

This talk will have both technical and nontechnical portions.  The
technical portion will cover some of the details of the database, the
crawler, and the front end.  The nontechnical portion will cover
marketing(!), and some of the unexpected social issues that the
production team has encountered in operating a public service, which
is used more than 90 million times per day.

Biography: Paul Flaherty was the inventor of AltaVista (for some value
of invention).  His current work includes projects in infrared LANs
and signal processing (spam filtering).  For fun, he runs the domain
viking.org, dabbles in ham radio, and is trying to brush up on his
Danish.

Dr. Flaherty received his Masters and Ph.D. in Electrical Engineering
from Stanford.  During his eight year tenure, he worked on a diverse
collection of projects, from radioscience to error coding to computer
architecture and networking.
                             ____________

                            SHIMIZU TALKS
                           Hiroshi Shimizu
       Emeritus Professor of Pharmacology, University of Tokyo
         Founder and Director of The "Ba" Research Institute,
                   Kanazawa Institute of Technology
              http://www.kanazawa-it.ac.jp/index_e.html

                                Talk 1
                 The Dual-Centered Model of the Self:
              An Overview of the Theory of Ba and Basho
              On Thursday, 12 March 1998, 3:00 - 4:00pm
                 Xerox PARC (CSL Commons #2230, PARC)
(If you are interested in attending, please contact Kim Paulon
(kpaulon@parc.xerox.com) at least one day ahead to arrange for escort.
Please plan to arrive not later than 2:45.)

                                Talk 2
            Fusing Eastern and Western Models of the Self:
              The Logic of Self and the Logic of Science
                   On Monday, 16 March 1998, 4:15pm
                             Cordura 100

                                Talk 3
             Design Principles for Knowledge Systems from
                 the Dual-Centered Perspective (Kumo)
               On Tuesday, 17 March 1998, 2:00 - 4:00pm
                      FX Palo Alto Laboratories 
(If you are interested in attending, please contact Livia Polanyi
(polanyi@pal.xerox.com) at least one day in advance to get directions
and arrange for escort.)

In recent years, scientists have recognized that information is
generated only with life and that humans are living members of the
earth's ecological system. As a living system, we live in the
non-deterministic complex world, with an indefinite and vague future.
It is important, then, for the sciences to embrace complexity,
subjectivity and vagueness if we are to answer the perplexing
questions which face us all. The concepts of Ba and Basho, strongly
related with Eastern (especially Japanese) culture, can be useful to
our thinking about these issues.  As the philosopher Kitaro Nishida
pointed out, Ba and Basho, are based on non-separated self, and must
be expressed by a predicative logic in contrast to the logic of
Western rational thinking which is subjective, based on separated
self, where the object is observed as definitely separate by the
subject which occupies the position of observer).

A living system lives, however, by fusing both selves, the separated
self and the non-separated self. How do these systems achieve
integration, maintaining self-consistency while avoiding contradiction?
A number of aspects of these issues will be addressed in this series of
public talks.

Biography: Professor Hiroshi Shimizu, a leading Japanese biophysicist,
has been working for many years on philosophical issues about the
nature of self, representation and the environment. His work combines
Eastern and Western views of the Self and Consciousness in order to
answer the fundamental question of how an organism can function in a
world in which it, itself, both forms part of the environment and is
determined by it.  His "dual centered theory of self" which builds on
and extends the Japanese concepts of "ba" and "basho" has emerged
recently as a very powerful unifying concept in Japanese thought not
only within theoretical philosophical circles but also in both the
business and national governmental sectors. In the corporate realm,
many major Japanese corporations are currently exploring how to build
upon his model in the design of products ranging from automobiles to
robots to assist people disabled by stroke or other brain injury to
learn to walk again. On a broader social level, the Japanese
government is interested in the application of Shimizu's theoretical
perspective to broader social issues including the resolution of
societal conflicts and the management of cultural diversity in a
global economy.

Since 1993. Professor Shimizu has been a Professor in the Faculty of
Information Technology, Kanazawa Institute of Technology and Director
of the "Ba" Institute. Before joining Kanazawa Institute, Professor
Shimizu was Professor at the University of Tokyo from 1977-1993.
                             ____________

                           XEROX PARC FORUM
             on Thursday, 12 March 1998, 4:00pm - 5:00pm
                    George Pake Auditorium, Xerox
            http://www.parc.xerox.com/ops/projects/forum/

                            Nanotechnology
                           Dr. Ralph Merkle
                   Xerox Palo Alto Research Center

Manufactured products are made from atoms.  The properties of those
products depend on the arrangement of those atoms.  Today's
manufacturing methods arrange atoms statistically, without control
over the placement of individual atoms.  Casting, milling,
lithography, and other traditional "bulk" manufacturing technologies
provide only approximate control over the molecular structure of the
material being manufactured.  In the future, we will be able to
inexpensively manufacture structures in which each individual atom is
in its proper place. This will be essential for the manufacture of
molecular computers, and will also let us cleanly make diamondoid
products of remarkable strength and lightness.

This talk will provide a general introduction to the field: where we
are, where we're going, and some of the possibilities once we get
there.

Biography: Dr. Merkle got his Ph.D. from Stanford University in 1979
where he co-invented public key cryptography.  He joined Xerox PARC in
1988, where he is pursuing research in computational nanotechnology.
He chaired the Fourth and Fifth Foresight Conferences on
Nanotechnology and is on the Executive Editorial Board of the journal
Nanotechnology.  Dr. Merkle has seven patents and has published
extensively.
                             ____________
   
       SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION (SCLA)
             on Thursday, 12 March 1998, 4:15pm to 5:30pm
                              Gates 100
              http://www-csli.stanford.edu/cll/scla.html

           Learning Situation-Dependent Planning Knowledge
                 from Uncertain Robot Execution Data
                           Karen Zita Haigh
                      Carnegie Mellon University
          mailto:Karen_Zita_Haigh@seldon.prodigy.cs.cmu.edu
     
Real-world domains are notoriously hard to model completely and
correctly. Robotics researchers have developed some learning
capabilities for their systems, but generally these have been limited
to learning operational parameters or other low-level information.
Machine Learning techniques, meanwhile, have generally not been
applied to real-world domains.

Real-world systems should adapt to changing situations and absorb
information that will improve their performance. My approach
automatically extracts information from execution experience, which is
then used by planning systems to improve performance.

I will be presenting the complete integrated planning, executing and
learning robot ROGUE. ROGUE analyzes execution experience to detect
patterns in the environment that affect plan quality. ROGUE extracts
learning opportunities from massive, continuous, probabilistic
execution traces. These learning opportunities are then correlated
with environmental features, thus detecting patterns in the form of
situation-dependent rules. I will describe the development and use of
these rules for two planners: the path planner and the task planner,
and present empirical data to show the effectiveness of ROGUE's novel
learning approach.

This learning approach is applicable for any planner operating in any
real-world domain. Situation-dependent rules effectively improve the
planner's model of the environment, thus allowing the planner to
predict and avoid failures, to respond to a changing environment, and
to create plans that are tailored to the real world.
                             ____________

                     STANFORD PHONOLOGY WORKSHOP
                  on Thursday, 12 March 1998, 7:30pm
                     Margaret Jacks Hall 460:146
      http://www-linguistics.stanford.edu/Linguistics/pinterest/

           From Bumps to Phonology: a model for recognizing
   intonational "phonemes" from duration and fundamental frequency
                           Arman Maghbouleh
                      Department of Linguistics
                         Stanford University

The mapping from low-level phonetic features of intonation, such as
fundamental frequency and duration, to high-level phonological
descriptions, such as ToBI labels, is complex because of linguistic
variability, interactions with phone segments, and errors in physical
measurements. In this talk, I will present a model for performing this
mapping. In particular, I will motivate, explain and analyze a set of
statistical models for recognizing ToBI accents from duration and
fundamental frequency (F0).

The models I have developed are theoretically relevant because their
structure reflects the underlying structures of intonation in
language.  They also have practical uses in human-computer dialog
systems or for labeling corpora in language research. I will explore
both the practical and theoretical implications of the models.

This talk will not presume any knowledge of intonation or statistics.

Technical Abstract
------------------
The difficulties in mapping duration and F0 to intonational labels arise
for three reasons:

   1. The realization of each intonational gesture is affected by its
      intonational context (e.g., the neighboring gestures, rate of
      speech, stress pattern). For example, the fundamental frequency
      (F0) contour associated with an H* label at the end of a phrase
      looks very different from the F0 of an H* in relative isolation.

   2. The duration and F0 of an utterance is affected by the segmental
      makeup as well as the intonational content of the utterance. For
      example, F0 gaps due to unvoiced segments and F0 perturbations
      due to consonants can be even more visually striking than F0
      variations due to intonation.

   3. There is uncertainty and error in measurements of duration and
      F0. For example, pitch trackers often report erroneous values
      for "creaky" sections of speech and report F0 values twice or
      half of the actual values.

In a previous Phonology Workshop, I outlined a method for recognizing
prominences in speech using duration values of syllable nuclei. In
that work, I explicitly factored out the segmental contributions to
duration (problem 2 above) and relied on the observation that
prominences lengthen syllable nuclei differentially more than
boundaries do (Campbell and Sagisaka, 1992) to distinguish prominences
from boundaries (problem 1 above).

In this workshop, I will concentrate on F0. I have followed other
researchers (Pierrehumbert, 1983; Hirst and Espesser, 1993; Taylor,
1994) in smoothing and stylizing F0 contours in order to tackle F0
detection errors and segmental perturbations (problems 2 and 3 above).

For recognizing accents, I have chosen to view the problem in the
framework of statistical pattern recognition (Devijver, and Kittler,
1980; see Wightman et al., 1994 and 1995 for a contrasting
application). Assuming that prominent syllables can be located in a
stream of speech by duration factors, the problem of recognizing
prominence types can be reduced to classifying the pitch movements on
prominent syllables. Surprisingly, the classification methods I am
using (Hastie et al., 1995) have been able to recognize distinctive
features of various intonational gestures even in the presence of
contextual factors (problem one above). In addition to describing the
results of the models, I will analyze their structure to shed light on
the distinctive features in intonational prominence.
                             ____________

                             LOGIC LUNCH
                  on Friday, 13 March 1998, 12 noon
                         Math Corner 380:383N
              http://www-philosophy.stanford.edu/Logic/

          Frege Structures for Partial Applicative Theories
                            Reinhard Kahle
              University of Tuebingen, visiting Stanford

Beeson has discovered that Frege structures can be considered as a
truth theory over applicative theories. Due to problems concerning
strictness the straightforward approach works in the total framework
only. Here we define a truth theory for the partial setting, by use of
a certain notion of pointer. This concept is closely related to the
treatment of streams in strict functional programming languages.
                             ____________

               SEMINAR ON PEOPLE, COMPUTER, AND DESIGN
                on Friday, 14 October 1998, 12:30-2:00pm
                      Gates B01 (HP Classroom)
                 http://www-pcd.stanford.edu/seminar/
                          (SITN Channel E2)

                     3DDI: 3D Direct interaction
                              John Canny
                Computer Science Division. UC Berkeley
                      mailto:jfc@cs.berkeley.edu

The 3DDI project is about direct interaction with simulated or remote
3D worlds. Users interact with the world without gloves or motion
capture sensors and view the world stereoscopically without glasses.
3D interaction preserves the spatial relationships and body language
cues among a group of people in ways that 2D video cannot. It also
supports a strong form of direct interaction where humans see no
interface at all, only 3d objects in the world.

The project involves 3 campuses and covers the technologies from
real-time 3d capture through physical modeling, to rendering on
autostereoscopic and volumetric displays. This talk will summarize the
project, and give details of 3 of its components: real-time depth
capture, physical behavior prototyping, and volumetric display.  3DDI
is one of the seeds of a much larger effort at Berkeley toward
Human-Centered Computing (HCC). I will talk about one other seed,
which is strong telepresence through "PRoPs" or robot avatars, and
then say something about what the rest of the HCC effort will
comprise.

Biography: John Canny is a professor in Computer Science at UC
Berkeley. He came from MIT in 1987 after his thesis on robot motion
planning, which won the ACM dissertation award. He received a Packard
Foundation Fellowship and a PYI while at Berkeley. His robotics work
was on path planning, grasping and the co-creation (with Ken Goldberg)
of RISC robotics, which is a fusion of algorithmic intelligence and
traditional manufacturing hardware. He has worked in applied
computational geometry and with Brian Mirtich on the development of a
physically-based simulator called IMPULSE. He developed inexpensive,
ubiquitous telepresence robots called "PRoPs", which evolved from
airborne to terrestrial locomotion. Two years ago, he started the 3DDI
project on direct 3D interaction with researchers from Berkeley, MIT
and UCSF. 3DDI includes the balanced co-development of simulation and
rendering algorithms with radically new hardware for acquiring and
displaying "into the world".

This is the last PCD talk of the Winter Quarter. Spring Quarter
sessions will begin on April 3.
                             ____________

                  STANFORD DIGITAL LIBRARIES SEMINAR
                   on Monday, 16 March 1998, 4:15pm
                         Gates Building, B08
       http://diglib.stanford.edu/diglib/seminars/seminars.html

                       Active Security Models:
      An Approach to Enterprise-oriented and Self-administering
          Security in Distributed Collaborative Environments
                            Roshan Thomas
                     Odyssey Research Associates
   
We argue for the need for a new paradigm for access control and
security to handle the demands of emerging applications and
environments. These include workflow applications involving multiple
points of decision making and distributed control, distributed
collaboration involving dynamically formed teams, as well as
agent-based systems. We introduce the notion of "active" security
models as an approach to overcoming the limitations of existing
security models. By active security models, we mean models that
approach security modeling and enforcement from the perspective of
activities or tasks, and as such, provide the abstractions and
mechanisms for the active runtime management of security as tasks
progress to completion. This task-based, active view of access control
and security is in sharp contrast to the existing passive,
resource-centric subject-object view of security typically found in
access control list based systems, where a subject is given access to
objects in a system based on some permissions (rights) the subject
possesses. The subject-object view of security typically divorces
access mediation from the larger context (such as the current state of
tasks) in which a subject performs an operation on an object and is
really an artifact of the early development of computer security.  We
show how active access controls can be used to overcome the
limitations of passive models by providing context-based security
management and enforcement, strict controls on usage, and an approach
to security administration that is scalable and self administering.
Through out the talk, we will illustrate the usefulness of active
controls with examples from workflow-based office automation and
healthcare applications.
   
Biography: Roshan Thomas is a researcher and Principal Scientist at
Odyssey Research Associates in Ithaca, NY. He is currently a principal
investigator on two DARPA-funded research projects exploring various
aspects of next generation security models and also the technical lead
on a consulting project with General Electric that is investigating
security in clinical workflows. He was previously involved as a
security consultant to Microsoft Corporation and worked on security
for Microsoft's OLE DB component database framework.
                             ____________

           THE GEORGE AND SANDRA FORSYTHE MEMORIAL LECTURES
                        Two lectures given by
                       Professor Stuart Russell
                  University of California, Berkeley

               Lecture 1: Rationality and Intelligence
                   on Monday, 16 March 1998, 7:30pm
                Gates, Hewlett-Packard classroom (B01)

        Lecture 2: Object Identification in a Bayesian Context
                  on Tuesday, 17 March 1998, 4:15pm
                Gates, Hewlett-Packard classroom (B01)

Stuart Russell was born in 1962 in Portsmouth, England. He received
his B.A.  with first class honours in Physics from Oxford University
in 1982, and his Ph.D. in Computer Science from Stanford in 1986
(working with Prof. Michael Genesereth).  He then joined the faculty
of the University of California at Berkeley, where he is currently a
Professor of Computer Science.  He is a recipient of the Presidential
Young Investigator Award of the National Science Foundation, the 1995
Computers and Thought Award, and a 1996 Miller Professorship of the
University of California. He was elected Fellow of the American
Association of Artificial Intelligence in 1997.

His research interests include machine learning, limited rationality,
real-time decision-making, intelligent agent architectures, autonomous
vehicles, search, game-playing, reasoning under uncertainty, and
commonsense knowledge representation.  He has published over ninety
papers and three books, "The Use of Knowledge in Analogy and
Induction" (Pitman, 1989), "Do the Right Thing: Studies in Limited
Rationality" (MIT Press, 1991, with Eric Wefald), and most recently
"Artificial Intelligence: A Modern Approach" (Prentice Hall, 1995,
with Peter Norvig).

The first lecture, entitled "Rationality and Intelligence" will be
given in the Hewlett-Packard classroom (B01) in the Gates Computer
Science Building at 7:30 in the evening on March 16.  The lecture will
be of general interest to people in the computing, engineering, and
cognitive community.  It will outline a gradual evolution in our
formal conception of intelligence that brings it closer to our
informal conception and simultaneously reduces the gap between theory
and practice.  There will be a reception after the lecture in the
lobby.

The second lecture, entitled "Object Identification in a Bayesian
Context" will be given in the Hewlett-Packard classroom (B01) in Gates
at 4:15 on March 17.  It will discuss the fundamental problem of
object identification, by computing the probability that any two
perceived objects are the same, given a stream of noisy observations.
                             ____________
                                     
                            LOGIC SEMINAR
                  on Tuesday, 17 March 1998, 4:15pm
                         Math Corner 380:381T
              http://www-philosophy.stanford.edu/Logic/

            Recent Results in Metapredicative Proof Theory
                            Thomas Strahm
         University of Bern, Switzerland (visiting Stanford)

Metapredicativity is a new general term in proof theory which
describes the analysis and study of formal systems whose
proof-theoretic strength is beyond the Feferman-Schuette ordinal
Gamma-0 but which are nevertheless amenable to predicative methods.

In this talk we give a general survey and introduction to
metapredicativity. In particular, we discuss various examples of
metapredicative systems, including (i) subsystems of second order
arithmetic, (ii) first and second order fixed point theories, (iii)
extensions of Kripke-Platek set theory without foundation, and (iv)
systems of explicit mathematics with universes.

Relevant keywords for our talk are: arithmetical transfinite recursion
and dependent choice; restricted bar induction; transfinite
hierarchies of fixed points; transfinite fixed point recursion; hyper
inaccessibility, Mahloness and Pi-3 reflection without foundation;
universe operators.
                             ____________

             ME297: DESIGN THEORY AND METHODOLOGY SEMINAR
                 on Wednesday, 18 March 1998, 3:15pm
              http://cdr.stanford.edu/DD/Courses/me297/
               (seminar may be closed, check web page)

     Grammatical Generation and Optimizing Search in Early Design
                            Jonathan Cagan
               George Tallman and Florence Barrett Ladd
                 Development Professor in Engineering
                 Department of Mechanical Engineering
                      Carnegie Mellon University
                                  
The early stages of design require search through large spaces of
design concepts. To keep the process tractable and productive, the
description of the design space must be concise, yet the scope of the
space still vast. Further, once a design space is defined it must be
searched in an efficient manner to determine good design solutions. In
this presentation we will explore the benefits of shape grammars to
describe engineering design spaces. These grammars either model
knowledge intensive products such as coffee makers, or
near-knowledge-free configurations such as truss structures. To search
design spaces such as those generated by knowledge-free grammars,
techniques must be able to move through discontinuous, non-smooth
topologies and models and determine not only feasible solutions but
optimally directed ones. We will explore how stochastic optimization
methods find optimal configurations in these spaces. One such
application will be in the area of structural topology design and
another in the area of product layout. We will also briefly examine
the relationship between these search methods and cognitive models of
human problem solving.
                             ____________

                              CSLI TALK
                 on Thursday, 19 March 1998, 12 noon
                        Cordura Hall, Room 100
             http://www-csli.stanford.edu/csli/Coglunch/

                     Global Intelligence Project
                              Masa Numao
               Professor, Tokyo Institute of Technology
                  http://www.titech.ac.jp/home.html

We propose a cooperative system of human beings and computers, called
"Global Intelligence" (GI). Human workers in GI are supported by a
computer system called GIANT (GI Associating NeTwork), which consists
of an inference mechanism by using a dynamically transforming network,
and has a learning mechanism by adjusting a weight on each link in the
network. After showing their structures and experiments, we point out
that we can construct an interesting model of intelligence not only by
analyzing a "Society of Mind", but also by synthesizing "Mind of a
Society".
                             ____________
   
       SEMINAR ON COMPUTATIONAL LEARNING AND ADAPTATION (SCLA)
             on Thursday, 19 March 1998, 4:15pm to 5:30pm
                              Gates 100
              http://www-csli.stanford.edu/cll/scla.html

            Learning Evaluation Functions for Optimization
                             Justin Boyan
                    mailto:jab@lefty.sp.cs.cmu.edu
       
STAGE is a new search technique which learns a problem-specific
heuristic evaluation function as it searches. The heuristic is trained
by least-squares TD(lambda) to predict, from features of states along
the search trajectory, how well a fast Markovian search method such as
hill-climbing will perform starting from each state. Search proceeds
by alternating between two stages: performing the fast search to
gather new training data, and following the learned heuristic to reach
a promising new start state.

STAGE has produced good results on a variety of combinatorial
optimization domains, including VLSI channel routing, Bayes net
structure-finding, bin-packing, Boolean satisfiability, radiotherapy
treatment planning, and geographic cartogram design. I'll discuss as
many of these successes as time permits, and also explain a STAGE
failure on the domain of inverse Boggle.
                             ____________

                             LOGIC LUNCH
                  on Friday, 20 March 1998, 12 noon
                         Math Corner 380:383N
              http://www-philosophy.stanford.edu/Logic/

      Reductions of Theories of Countable Tree Ordinals to ID_1.
                             Sol Feferman

This is the last logic lunch of the quarter.  We plan an informal
lunch at the coffee house afterward.

This is a report of a section from the forthcoming chapter by Jeremy
Avigad and myself on Goedel's functional ("Dialectica"
interpretation.* I will sketch two applications of this
interpretation, one for a classical and one for an intuitionistic
theory of countable tree ordinals, to reduce them to corresponding
systems of one arithmetical inductive definition.  The questions
following this work are: (i) how these are related, and (ii) can
anything similar be done for theories of higher ordinal number
classes?

*To appear in the Handbook of Proof Theory, S. Buss ed.
                             ____________

                             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.  They can also type 'help quip' to see the
Linguistics Department newsletter, the Sesquipedalian,
http://www-linguistics.stanford.edu/Linguistics/.

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.
                             ____________