LINGUIST List 15.637
Thu Feb 19 2004
Calls: Computational Ling; Computational Ling/Portugal
Editor for this issue: Andrea Berez <andrealinguistlist.org>
As a matter of policy, LINGUIST discourages the use of abbreviations
or acronyms in conference announcements unless they are explained in
the text.
To post to LINGUIST, use our convenient web form at
http://linguistlist.org/LL/posttolinguist.html.
Directory
rschneid, Lessons Learned from Evaluation: Towards Transparency and Integration in CLIR
Vincenzo.Pallotta, Robus Methods in Analysis of Natural Language Data
Message 1: Lessons Learned from Evaluation: Towards Transparency and Integration in CLIR
Date: Thu, 19 Feb 2004 04:13:42 -0500 (EST)
From: rschneid <rschneiduni-hildesheim.de>
Subject: Lessons Learned from Evaluation: Towards Transparency and Integration in CLIR
Lessons Learned from Evaluation: Towards Transparency and Integration
in CLIR
Short Title: LECLIQ
Date: 29-May-2004 - 29-May-2004
Location: Lisbon, Portugal
Contact: René Schneider
Contact Email: lecliquni-hildesheim.de
Meeting URL: http://www.uni-hildesheim.de/~lecliq
Linguistic Sub-field: Computational Linguistics
Call Deadline: 29-Feb-2004
This is a session of the following conference: 4th International
Conference on Language Resources and Evaluation
Meeting Description:
A half-day workshop will be held in conjunction with the LREC
Conference in Lisbon on the morning of May 29, 2004. The topic of the
workshop is ''Lessons Learned from Evaluation: Towards Transparency
and Integration in Cross-Lingual Information Retrieval (LECLIQ)'' with
special focus on the use and benefits of quality gates.
Extended Call for papers
Workshop on
LESSONS LEARNED FROM EVALUATION:
TOWARDS TRANSPARENCY AND INTEGRATION
IN CROSS-LINGUAL INFORMATION RETRIEVAL
(LECLIQ)
to be held in
Lisbon, Portugal, on May 29, 2004
in conjunction with the
4th Conference on Language Resources and Evaluation (LREC-2004)
TECHNICAL DESCRIPTION
Recent work in Cross-Lingual Information Retrieval (CLIR) has shown
that systems perform differently with respect to queries, topics, data
sets and the corresponding query or result language. Since the
state-of-the-art in CLIR is far from finding an ideal method for this
variety of parameters, a better performance might be possible through
continuous analysis and integration of evaluation results achieved
from several retrieval devices. The process of integrating these
different methodologies might be controlled by quality gates. Quality
gates - having their origin in car manufacturing and being used in
IT-project management as well - are characterized as preventive and
process accompanying control mechanisms to check critical parameters
and ensure quality standards during the design, development, and
deployment of software. They generally consist of checklists combined
with appropriate rules to guarantee that work procedure fail is
recognized in time to prevent repetition.
The objective of this workshop is to bring together researchers from
the field of Information Retrieval and Language Engineering with three
primary goals:
(1) to discuss, evaluate and judge the different methodologies used in
CLIR and to relate them to the languages being retrieved,
(2) to design the requirements for a dynamic network of quality gates
for the tuning of retrieval results,
(3) to define desired learn effects caused by the interaction of
quality gates.
Topics of interest include, but are not restricted to:
- Lessons learned in evaluation initiatives such as CLEF, TREC, INEX,
NTCIR, etc.
- Criteria, checklists, parameters and metrics for the evaluation of
search and retrieval methods
- Quality criteria and evaluation for resources and tools in CLIR
(thesauri stemmer, translation tools and services, etc.)
- Explanation and verification of success stories and failure analysis
in CLIR
- Data and Text Mining on CLIR evaluation results
- Current evaluation issues: shortcomings, gaps, reliability of
results, etc.
- Transfer of evaluation results to system design and product
development
- Design for (coupled or networked) quality gates in CLIR (for systems
and building blocks)
- Semaphore logic for the integration of distributed retrieval systems
- Crossing-over techniques for quality assurance
- Hybrid systems for CLIR
The results of this workshop will be a first step toward outlining a
general framework for quality gates in (Cross-Language) Information
Retrieval with further elaboration in future workshops.
We are inviting contributions discussing the themes above from
participants in evaluation initiatives as well as from experienced
researches in the areas of (Cross Language) Information Retrieval,
quality assurance, Human Language Technology (HLT) or related fields.
PROGRAMME COMMITTEE
Norbert Fuhr, University of Duisburg, Germany
Marcello Federico, I.R.S.T, Italy
Gregory Grefenstette, Clairvoyance Corp., Pittsburgh, USA
Donna Harman, NIST, USA
Paul Heisterkamp, DaimlerChrysler AG, Germany
Thomas Mandl, University of Hildesheim, Germany
Dagobert Soergel, University of Maryland, USA
Christa Womser-Hacker, University of Hildesheim, Germany
[Final list forthcoming]
SUBMISSION OF CONTRIBUTIONS
Presentations and active participation in the workshop will be based
on a 2 - 6 page position statement or research paper. Submissions must
be in English and should consist of an abstract, plus a title page
providing the following information: official title of the paper;
names and affiliations of the authors; full address of the first
author including phone, fax, e-mail, URL; presentation facilities
required.
Only electronic submissions in Word, PDF or RTF will be accepted.
Demonstrations of cross-lingual retrieval tools will be considered as
well if the demonstration outlines the rationale of the system and
sheds light on the success or failure of the methods applied.
Contributions should be submitted to: lecliquni-hildesheim.de
Reception will be acknowledged within three days. If you do not get an
acknowledgement, please check the e-mail address of your submission.
PROCEEDINGS
The workshop organizers will publish proceedings of the
workshop. Ready-to-print versions of the papers have to be submitted
according to LREC Guidelines for Authors, which can be found at the
conference homepage (to be announced in February 2004). They should
not exceed six pages. These final versions should be submitted
electronically to the following e-mail address:
lecliquni-hildesheim.de
IMPORTANT DATES
Deadline for submitting papers: February 29
Notification of acceptance: March 15
Final versions of papers for proceedings: April 4
Workshop: May 29
FURTHER INFORMATION
Information about the workshop such as call, outline, and program can
be found on the LREC web-page: http://www.lrec-conf.org/lrec2004/
where you can also find information about the LREC conference itself.
Information about the general purpose of quality gates in
manufacturing and software-engineering and their transduction to the
field of (Cross Lingual) Information Retrieval will be available under
http://www.uni-hildesheim.de/~lecliq
REGISTRATION
The registration fee for the workshop is:
- 50 EURO for conference participants
- 85 EURO for those not attending LREC-2004
Registration and payment procedures are explained on the LREC-2004
web-page. The registration fee includes one copy of the proceedings
and a coffee break.
QUESTIONS
For all questions on the focus of the workshop, please contact the
workshop co-ordinator under lecliquni-hildesheim.de.
ORGANIZING COMMITTEE
Christa Womser-Hacker, University of Hildesheim, Center for
Information Science (Chair)
Norbert Fuhr, University of Duisburg
Paul Heisterkamp, DaimlerChrysler AG
Thomas Mandl, University of Hildesheim
René Schneider, University of Hildesheim (Co-ordinator)
René Schneider
University of Hildesheim, Marienburger Platz 22, D-31141 Hildesheim,
Germany
Phone: +49(0)5121/883-834 Fax: +49(0)5121/883-802
e-mail: rschneid uni-hildesheim.de
Message 2: Robus Methods in Analysis of Natural Language Data
Date: Wed, 18 Feb 2004 20:37:23 -0500 (EST)
From: Vincenzo.Pallotta <Vincenzo.Pallottaepfl.ch>
Subject: Robus Methods in Analysis of Natural Language Data
Robus Methods in Analysis of Natural Language Data
Short Title: ROMAND 2004
Date: 29-Aug-2004 - 29-Aug-2004
Location: Geneva, Switzerland
Contact: Vincenzo Pallotta
Contact Email: romandepfl.ch
Meeting URL: http://lithwww.epfl.ch/romand2004
Linguistic Sub-field: Applied Linguistics ,Computational Linguistics
,Discourse Analysis ,General Linguistics ,Morphology ,Pragmatics
,Semantics ,Syntax ,Text/Corpus Linguistics ,Translation
,Neurolinguistics ,Cognitive Science
Call Deadline: 31-Mar-2004
Meeting Description:
ROMAND 2004 is the third of a series of workshops aimed at bringing
together researchers and students that work in fields like artificial
intelligence, computational linguistics, human-computer interaction,
cognitive science and are interested in robust methods in natural
language processing and understanding. ROMAND 2004
3rd workshop on RObust Methods in Analysis of Natural language Data
A satellite event of COLING 2004
Geneva - August 29th 2004
http://lithwww.epfl.ch/romand2004/
Robustness in Computational Linguistics has been recently recognized
as a central issue for the design of interactive applications based on
natural language communication. If a failure of the system can be
acceptable in batch applications requiring a human intervention, an
on-line system should be capable of dealing with unforeseen situations
in a more flexible way. When we talk about system failure we do not
think at inherent program failures like infinite loops or system
exception, we consider, rather, failures related to the processing of
the input and its assimilation in the system's knowledge base. A
failure of this kind means simply that the system does not
''understand'' the input. The automated analysis of natural language
data has become a central issue in the design of Intelligent
Information Systems. Processing unconstrained natural language data is
still considered an AI-hard task. However, various analysis techniques
have been proposed in order to address specific aspects of natural
language. In particular, recent interest has been focused on providing
approximate analysis techniques, assuming that when perfect analysis
is not possible, partial results may be still very useful.
ROMAND 2004 is the third of a series of workshops aimed at bringing
together researchers and students that work in fields like artificial
intelligence, computational linguistics, human-computer interaction,
cognitive science and are interested in robust methods in natural
language processing and understanding. The term ''natural language''
is intended as all the possible modalities of human communication and
it is not restricted to written or spoken language. Theoretical
aspects of robustness in Natural Language Processing (NLP) and
Understanding (NLU) are welcome as well as engineering and industrial
experiences. We are also interested in research works that investigate
different types of linguistic modules integration, in order to
characterise their advantages and disadvantages with respect to
different types of applications, and works that propose different
types of NLP/NLU systems, ranging from sequential/pipelined
architectures to more loosely-coupled multi-level distributed
architectures.
We invite papers on all topics related to robustness in natural
language processing, including, but not limited to:
* Natural Language Architectures
* Robust Morpho-Syntactic Parsing
* Robust Semantics
* Robust Discourse Analysis
* Robust Computational Pragmatics
* Complexity of linguistic analysis
* Spell checking and automated spell correction
* NLP and Soft Computing
* Hybrid methods in computational linguistics
* Information Extraction
* Multimedia Document Analysis
* Spoken Language Understanding
* Multimodal Human-Computer Interfaces
* Cognitive Linguistics
INVITED TALKS
Dr. Frank Keller - University of Edinburgh: ''Gradient Grammaticality:
Applications in Robust Parsing''
SUBMISSION PROCEDURE:
Authors should submit an anonymous paper (avoiding, as much as
possible, hints for the identification of authors) of at most 10 pages
(including pictures and references), intended for talks with a
duration of 20 minutes plus 10 minutes for discussions. The paper
should follow the COLING formatting style available on the main
conference web site (http://www.issco.unige.ch/coling2004/), and the
authors must also send us a wrapper e-mail message indicating:
1. the AUTHOR'S NAMES, AFFILIATION, ADDRESS, and E-MAIL address of
(at least) the contacting author,
2. an ABSTRACT of the submitted paper (less than 500 words),
3. a list of KEYWORDS (no more than 5).
The papers should be submitted electronically in PDF (preferred) or
postscript format to:
mailto:romandepfl.ch.
IMPORTANT Papers due: March 31st
Acceptance notification: May 7th
Final version due: June 4th
Workshop: August 29th
PROGRAM CHAIRS
Vincenzo Pallotta (Swiss Federal Institute of Technology - Lausanne,
International Computer Science Institute - Berkeley)
Amalia Todirascu (University of Troyes and University of Iasi)
PROGRAM COMMITTEE
Afzal Ballim
Alberto Lavelli
Alexander Clark
Atro Voutilainen
Bangalore Srinivas
C.J. Rupp
Dan Cristea
Diego Mollá-Aliod
Eric Wherli
Josè Iria
Fabio Massimo Zanzotto
Fabio Rinaldi
Frank Keller
Giovanni Coray
Günther Görz
Hatem Ghorbel
Jean-Cédric Chappelier
Jean-Pierre Chanod
Joachim Niehren
John Carroll
Kay-Uwe Carstensen
Manuela Boros
Maria Teresa Pazienza
Martin Rajman
Michael Hess
Patrick Ruch
Roberto Basili
Rodolfo Delmonte
Salah Ait-Mokhtar
Susan Armstrong
Wolfgang Menzel
Yuji Matsumoto
LOCAL ORGANIZATION
Violeta Seretan (University of Geneva)
Hatem Ghorbel (Swiss Federal Institute of Technology - Lausanne)
News about the conference will be announced on the workshop's Web page
at http://lithwww.epfl.ch/romand2004/
CONTACT INFORMATION:
All inquiries should be sent to [romandepfl.ch]. Please note that
general organizational details (registration, accommodation, etc.) are
taken care of by COLING 2004, not by the workshop organizers.