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

CSLI Calendar, Correction



First a correction on a time and place.  The special
AI-Vision-Robotics Division Colloquium on Friday will take place at
11:00am not 4:15pm and in Gates B12 not Gates 104.

FRIDAY, 8 MAY                
        11:00am AI-Vision-Robotics Division Colloquium
                Gates B12
                Exploiting Structure in Solving Planning Problems
                or
                Coping with the Blooming and Buzzing World that Surrounds
                Us
                Tom Dean
                Brown University
                special colloquium  

Since I had to send this message out anyway, here are a few late
announcements

MONDAY, 11 May 1998
         4:30pm Stanford Digital Libraries Seminar
                Gates B08
                Untangling Text Data Mining
                Marti Hearst
                Berkeley
                Abstract below

TUESDAY, 11 MAY
         4:15pm Logic Seminar
                Room 380:381T
                Reductions of Finite and Infinite Proofs
                G. Mints
                Stanford
                Abstract below

FRIDAY, 15 MAY
        12 noon Logic Lunch
                Room 380:383N
                Deconstructing Tarski's Semantics for Predicate Logic
                or
                Secret World of Decidable First-Order Logics
                Johan van Benthem
                Stanford and Amsterdam
                Abstract below
                             ____________

                  STANFORD DIGITAL LIBRARIES SEMINAR
                    on Monday, 11 May 1998, 4:30pm
                         Gates Building, B08
       http://diglib.stanford.edu/diglib/seminars/seminars.html

                     Untangling Text Data Mining
                             Marti Hearst
                             UC Berkeley

The possibilities for data mining from large text collections are
virtually untapped.  Text expresses a vast, rich range of information,
but encodes this information in a form that is difficult to decipher
automatically.  For this reason, there has been little work on text
data mining to date.  Furthermore, most people who have talked about
text data mining have either conflated it with information access or
have not made use of text directly.

In this talk I will define data mining, information access, and
corpus-based computational linguistics, and talk about the
relationship of these to text data mining.  I will then provide
examples of what I think text data mining will really be about once
people actually start working on it.  Caveat emptor: I have not done
any text data mining work myself to date, but may do so in future.

About the Speaker: Marti A. Hearst (hearst@sims.berkeley.edu) is an
assistant professor at SIMS (School of Information Management and
Systems) at the University of California, Berkeley.  Her research
interests focus on the design and development of user interfaces and
robust language analysis for information access systems, and on
furthering our understanding of how people use and understand such
systems.  She is an associate editor for ACM Transactions on Office
Information Systems and edits the Trends and Controversies feature for
IEEE Intelligent Systems.  She received her PhD in computer science
from UC Berkeley in 1994.

                             ____________
                                     
                            LOGIC SEMINAR
                   on Tuesday, 12 May 1998, 4:15pm
                    Math Corner 380:381T (or 383N)
             http://www-logic.stanford.edu/seminars.html

               Reductions of Finite and Infinite Proofs
                               G. Mints
                               Stanford

Standard expansion of a finite proof d into an infinite proof h(d) was
a basis of many "extensional" proof-theoretic results, for example
ordinal bounds.  A slight change of this expansion allows to obtain
also intensional results like normalization theorems or connections
between finite and infinite reductions.
                             ____________

                             LOGIC LUNCH
                   on Friday, 15 May 1998, 12 noon
                         Math Corner 380:383N
             http://www-logic.stanford.edu/seminars.html

        Deconstructing Tarski's Semantics for Predicate Logic
                                  or
             Secret World of Decidable First-Order Logics
                          Johan van Benthem
                        Stanford and Amsterdam

First-order predicate logic is the general-duty tool of modern
logic. Its syntax, semantics, and metatheory are well-known, and form
the backbone of standard textbooks.  But a price has to be paid for
this broad applicability.  This basic tool of the logician is
undecidable, contrary to what many founding fathers may have
expected. But is the price inevitable? Could things have gone
differently?

Returning to an earlier lunch presentation, we'll show how to analyze
'standard Tarski semantics' into a number of independent decisions,
the conjunction of which leads to undecidability. Other choices turn
out to generate a landscape of decidable 'predicate logics'. This
insight needs no more, essentially, than a fresh closer look at what
students naturally do when they show that certain predicate-logical
principles are 'valid'. We'll survey some of the current technical
theory in this area. We'll also discuss briefly what this
deconstruction means for the usual monolithic view of 'standard
logical systems'.

Literature

J. van Benthem, 1996, "Exploring Logical Dynamics", Chapter 9, CSLI
Publications, Stanford & Cambridge University Press.

M. Marx & Y. Venema, 1996, "Multi-Dimensional Modal Logic", Kluwer
Logic Library, Kluwer Academic Publishers, Dordrecht.