Question-Answering Systems
SPECIAL ISSUE OF THE TAL JOURNAL
Deadline for submission: 10 April 2006
GUEST EDITORS:
Brigitte Grau (LIMSI - Orsay) and Bernardo Magnini (ITC-irst, Trento)
SECOND CALL FOR PAPERS
Advanced Information Access Systems have to provide concise answers to
precise questions in order to satisfy user needs. When users are
interested in specific factual information, answers can be extracted
from a document collection only by using systems specifically designed
for the task. For example, answering a question such as "What is the
most expensive car in the world?" requires deep analysis of document
passages to extract the relevant piece of information. Interrogations
can cover either an open domain or a specialty domain. In both cases
the main characteristic of recent Question Answering (QA) systems is
that the answer is extracted from a document collection.
QA functionalities can be implemented only if IR systems can analyze
both queries and documents at a deeper level. As a consequence, QA is
at the crossroad of several research fields. It is grounded, of
course, on Information Retrieval (IR), but it also concerns Natural
Language Processing in an important way. To some extent, it also
comprises fields such as Machine Learning and Knowledge
Representation.
Moreover, QA scenarios look to new capabilities to address crucial
issues in Information Access. Some of these capabilities are still in
an embryonic state in current IR systems, including: evaluating
whether an answer to a question exists in a document collection,
achieving a synthesis from multiple or partial answers, and using
dialog for refining a query. Other capabilities, like the merging of
partial answers, needs the integration of techniques developed in
different research areas, such as anaphora resolution and textual
inferences.
The main focus of the Special Issue is on open-domain QA, either for
large collections of documents or for the Web. Papers presenting QA
systems for restricted domains or dedicated to QA on structured
information (e.g., knowledge bases, databases) will also be taken into
account provided that the problems and solutions they address can be
easily ported to textual open-domain QA.
Submissions are expected on all aspects related to Question-Answering.
More precisely, contributions can either present a Question-Answering
system as a whole or focus on one of its processes. If the submission
concerns a process, it should emphasize its role in the Question
Answering context.
Specific topics include, but are not limited to:
* Question analysis: question classification, extraction of the
expected answer type and focus, extraction of the question context
or, more generally, of semantic constraints which can help the
answer extraction.
* Passage extraction: matching between a question and document
passages, use of paraphrases and textual entailment for Question
Answering.
* Resources and tools for QA: recognition and extraction of terms and
their linguistic variants, question type classifications,
repositories of answer patterns, multilingual resources for QA.
* Answer extraction: Named Entities for QA, type checking for answer
extraction, methods for answers ranking.
* Answer justification: methods for enriching exact answers with
contextual information.
* Complex questions: temporally-restricted questions, context-related
questions, opinion questions.
* Evaluation: QA components evaluation, definition and construction of
benchmarks for QA, automatic or semi automatic methods for
evaluating QA systems, methods for estimating question difficulty.
Submissions can also consider cross-domain topics in relation to
application scenarios in Question Answering, such as:
* QA and machine learning: use of machine learning for selecting and
extracting answers to a question, datasets for training machine
learning algorithms for Question Answering applications.
* Interactive Question Answering.
* Multilingual QA: portability of QA systems to another language;
cross-language QA (i.e. questions in a source language and search in
a document collection in a target language); machine translation
applied to QA.
* QA applications on the Web: using the Web as a source of knowledge
or a source of answers; specific aspects of searching answers on the
Web.
* Multi-document QA: fusion and coherence of multiple answers, partial
answers, answer generation.
THE JOURNAL
(see http://www.atala.org/ or http://tal.revuesonline.com/)
TAL (Traitement Automatique des Langues/Natural Language Processing)
is a forty year old international journal edited by ATALA (French
Association for Natural Language Processing) with the support of CNRS
(National Center for Scientific Research). It is published and
distributed by Hermès Lavoisier.
FORMAT
Papers will be submitted in PDF format. Style sheets (Word and LateX)
are available for download at http://tal.e-revues.com/appel.jsp
LANGUAGE
Manuscripts may be submitted in English or French. French
language is mandatory for French-speaking authors.
IMPORTANT DATES
10/04/2006 Deadline for submission
31/05/2006 Notification to authors
10/07/2006 Revised version
29/08/2006 Final version of papers
PAPER SUBMISSION
Contributions (25 pages maximum) have to be sent by e-mail to
both addresses below:
Brigitte Grau <brigitte.grau@iie.cnam.fr>
Bernardo Magnini <magnini@itc.it>
SCIENTIFIC COMMITTEE
Patrice Bellot (LIA, Avignon)
Moh Boughanem (IRIT, Toulouse)
Claude de Loupy (Sinequa, Paris)
Olivier Ferret (CEA-LIST, Fontenay aux roses)
Patrick Gallinari (LIP6, Paris)
Claire Gardent (Loria, Nancy)
Gabriel Illouz (LIMSI, Orsay)
Guy Lapalme (RALI, Montreal)
Dominique Laurent (Synapse, Toulouse)
Jimmy Lin (University of Maryland)
Diego Mollá (Macquarie University)
Laura Monceaux (LINA, Nantes)
Thierry Poibeau (LIPN, Villetaneuse)
Isabelle Robba (LIMSI, Orsay)
Sophie Rosset (LIMSI, Orsay)
Horacio Saggion (University of Sheffield)
Richard Sutcliffe (University of Essex)
Isabelle Tellier (GRAPPA, Lille)
Anne Vilnat (LIMSI, Orsay)
This archive was generated by hypermail 2b29 : Tue Mar 14 2006 - 15:11:19 MET