Workshop on Operational Text Classification Systems 2001
New Orleans, Louisiana, USA
September 13, 2001
http://www.DavidDLewis.com/events/otc2001
otc2001info@DavidDLewis.com

in conjunction with
ACM SIGIR 2001
September 9-13, 2001
http://www.sigir2001.org

Text classification research and practice has exploded in the past decade. This work has been pursued under a variety of headings (text categorization, automated indexing, text mining, topic detection and tracking, etc.).   Both the automated assignment of textual data to classes, and the automated discovery of such classes (by techniques such as clustering) have been of intense interest.   A variety of practical applications have been fielded, in areas such as indexing of documents for retrieval, hierarchical organization of Web sites, alerting and routing of news, creation of specialized information products, enforcement of information security, content filtering (spam, porn, etc.), help desk automation, knowledge discovery in textual and partially textual databases, and many others.

Experiments on text classification data sets have been widely presented in a variety of forums.  The technical details of operational text classification, however, have rarely been discussed.

The goal of this workshop is to expose researchers and practitioners to the challenges encountered in building and fielding operational text classification systems. We hope to begin the systematizing of engineering principles in this area, and spark new directions for research as well.

TOPICS

Workshop topics will include (but are not limited to):

* Cost effectiveness of automating text classification tasks

* Understanding what users want from classification systems

* Technical and personnel issues in using training data and prior knowledge

* Trading off space, time, and other resources in the training, adaptation, and execution phases of classification

* Integrating automated classification systems with pre-existing software, organizational procedures, relevant laws, and cultural expectations

* Maintaining and monitoring effectiveness as text sources and classes change over time

* Discovering, defining, updating, and explaining classes and classifiers

* The roles of classification and related technologies (information extraction, terminology discovery, etc.)
 

PARTICIPATION

To facilitate discussion, workshop attendance will be limited to a maximum of 70 participants. Anyone interested in attending should apply in one of these two ways:

1. Researchers, practitioners, and users with an interest in text classification:
     **Please submit a paragraph describing your background, organizational affiliation (if any), and interest in text classification.

2. Prospective speakers with substantial knowledge of one or more operational text classification systems and an interest in presenting a talk based on their experience:
     **Please submit both a paragraph of interest (as described above) and an abstract (maximum 750 words) outlining the major points you would speak on.  Talks whose focus is experimental results on standard test collections are discouraged.  Conversely, operational text classification at any scale from the tiny (e.g. an evaluation of content filtering software for a small organization) to the huge (e.g. categorizing hundreds of newswires each day) is of interest.  Selection of talks will be largely based on the speaker's ability and willingness to discuss technical details of operational systems, as reflected in their abstract.

Submissions should be sent in ASCII or PDF form to:
                      otc2001submit@DavidDLewis.com
All submissions will be reviewed by the organizers and program committee. The interest paragraphs and talk proposals of invited participants will be reproduced and distributed as an informal notebook at the workshop.

IMPORTANT DATES

Interest paragraphs must be received: June 15, 2001
Talk abstracts must be received: June 15, 2001
Notification of acceptance: July 15, 2001
Workshop: September 13, 2001

Please visit http://www.sigir2001.org  for hotel and registration deadlines.

ORGANIZERS
        David D. Lewis, independent consultant (Chair)
        Susan Dumais, Microsoft
        Ronen Feldman, Clearforest
        Fabrizio Sebastiani, Italian National Council of Research

PROGRAM COMMITTEE
        James Allan, University of Massachusetts
        David Evans, Clairvoyance
        Sue Feldman, IDC
        Norbert Fuhr, University of Dortmund
        Thorsten Joachims, GMD
        Andras Kornai, Northern Light
        Wai Lam, Chinese University of Hong Kong
        Dunja Mladenic, J. Stefan Institute and Carnegie Mellon Univ.
        Isabelle Moulinier, Thomson
        Christopher Porter, Factiva
        Prabhakar Raghavan, Verity
        Mehran Sahami, E.piphany
        Robert Schapire, AT&T
        Frank Smadja, Elron Software
        Richard Tong, Tarragon Consulting
        Mark Wasson, LexisNexis
        Scott Waterman, Kanisa Inc.
        Yiming Yang, Carnegie Mellon University