alexis nasr alexis.nasr at
Mon Jun 18 16:02:51 UTC 2001

                           CALL FOR PAPERS

       Embedded MT Systems: Leveraging for Real World Applications

Carol Van Ess-Dykema, U.S.Department of Defense,cjvanes at
Clare R. Voss, Army Research Laboratory, voss at

An "embedded machine translation (MT) system" is a computational
system with one or more MT engines embedded among its components.  These
systems accept various well-formed and degraded types of multilingual
and multi-modal input, including
     * hard-copy pages (original and OCR-ed image),
     * online files (web pages, word processing files, email),
     * video (image and text),
     * speech (natural signal and transcribed).
>>From this range of input, such systems enable users to access the
original, foreign language information in their own language.

Traditionally, the term "MT" has been associated with the task of
single document translation.  More recently, MT engines within embedded
larger systems have been used to facilitate tasks that require processing
multilingual information both within and across documents. Several
real-world applications have led to the widespread use of embedded MT
systems for cross-language tasks, such as:
     * content extraction           * document filtering
     * information retrieval        * question-answering
     * summarization

This call is for original or updated research articles on embedded MT
systems. Authors should address two or more of the following foci in
their papers:
1. Input Focus, such as
- Processing methods for range of system input
- Real-time analysis of input
- Channel-specific feedback processes from system components to correct
  noisy input
- Multiple document or channel input integration for summarization or
  other tasks

2. MT Methods Focus, such as
- Symbolic, statistical, or hybrid techniques for translation
- Machine learning applied to construction of MT engine components
- Weighting of candidates for selection of high quality MT output
- Quick ramp-up MT engines

3. System Design Focus, such as
- Architectures to support novel cross-language tasks
- Systems with multiple MT engines
- Pre-MT compensation components or strategies that adjust speech and
  OCR degraded input
- Post-MT processing of output for display to user
- Integration of system output within application software

4. Evaluation Focus, such as
- Linguistic measures of MT component performance
- Task-based measures of system effectiveness
- Experimental designs for assessing or comparing system effectiveness
- Software tools for assisting or automating evaluation process
- Algorithms

The idea of this special issue originates from workshops at ANLP/NAACL-2000
and AMTA-98. Authors from these workshops are encouraged to submit papers
describing progress on their systems and incorporating feedback from
workshop participants. We also encourage other researchers developing
embedded MT systems to submit papers.

Note that, for this special issue, two types of papers will be considered
for review: long papers (20 pages or more) and short papers (less than 20
pages). Longer papers are advised for research projects that have been
implemented and evaluated. Shorter papers are appropriate for recent
research initiatives. We would ideally like to have papers of both types
in the journal.


Lynn Carlson, U.S. Department of Defense
Gary Coen, Boeing Mathematical and Computing Technologies
Mike Dillinger, Logos Corporation
Robert Frederking, Carnegie Mellon University
Laurie Gerber, Language Technology Broker
Ulf Hermjakob, ISI/University of Southern California
Pierre Isabelle, Xerox Research Centre Europe
Lori Levin, Carnegie Mellon University
Kathryn Taylor, Georgetown University
Takehito Utsuro, Toyohashi University of Technology
Remi Zajac, New Mexico State University
Joe F. Zhou, Intel China Research Center

Call for papers issued:    1 June 2001
Papers due:                19 October 2001
Notification to authors:   22 February 2002

Initial submissions should be sent to:
1. Guest Editors      cjvanes at, voss at
2. Publishing Editor  tamara.welschot at

For initial submissions only, authors should send electronic copies
(postscript, pdf, rtf, or doc)  to the Guest Editors AND the Publishing
Editor. Please indicate that the submission is for the Special Issue of
Machine Translation on Embedded MT Systems: Leveraging for Real World
Applications. Details concerning the format for initial submissions
forthcoming and will be posted on the homepage:

Formatting instructions for final submissions will be made available with
acceptance. The Guest Editors will notify the submitters of the acceptance
status of their papers. All other information concerning the Special Issue
will be posted on the homepage.

Questions about submissions should be directed to the two Guest Editors by
email, rather than the Journal or Publishing Editors.
Message diffusé par la liste Langage Naturel <LN at>
Informations, abonnement :
English version          :
Archives                 :

La liste LN est parrainée par l'ATALA (Association pour le Traitement
Automatique des Langues)
Information et adhésion  :

More information about the Ln mailing list