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

[eu_members at aclweb dot org] Singapore: ACL/IJCNLP 2009: 2nd Workshop on Building and Using Comparable Corpora -- Call for Participation and List of Panelists

       (NEW: List of Panelists)

2nd Workshop on Building and Using Comparable Corpora:
from parallel to non-parallel corpora

   ACL-IJCNLP 2009
     Singapore, August 6th, 2009
          Suntec, WS7, MR301



 Kenneth Ward Church (Chief Scientist, Human Language Technology
   Center of Excellence, Johns Hopkins University, USA)


 Kenneth Ward Church (HLTCOE, Johns Hopkins University, USA)
 Hitoshi Isahara (National Institute of Information and Communications
   Technology, Japan)
 Kevin Knight (Information Sciences Institute, USC, USA)
 Franz Joseph Och (Google Research, USA)
 Richard Sproat (OGI School of Science & Technology, OHSU, USA)


 Research in comparable corpora has been motivated by two main reasons
 in the language engineering and the linguistics communities.  In
 language engineering, it is chiefly motivated by the need to use
 comparable corpora as training data for statistical NLP applications
 such as statistical machine translation or cross-language information
 retrieval.  In linguistics, on the other hand, comparable corpora are
 of interest themselves in providing intra-linguistic discoveries and
 comparisons.  It is generally accepted in both communities that
 comparable corpora are documents in one to many languages, that are
 comparable in content and form in various degrees and dimensions.  It
 was pointed out that parallel corpora are at one end of the spectrum
 of comparability whereas quasi-comparable corpora are at the other
 end.  We believe that the linguistic definitions and observations in
 comparable corpora can improve methods to mine such corpora for
 applications to statistical NLP.  As such, it is of great interest to
 bring together builders and users of such corpora.

 Parallel corpora are a key resource as training data for statistical
 machine translation, and for building or extending bilingual lexicons
 and terminologies.  However, beyond a few language pairs such as
 English-French or English-Chinese and a few contexts such as
 parliamentary debates or legal texts, they remain a scarce resource,
 despite the creation of automated methods to collect parallel corpora
 from the Web.  Interest in non-parallel forms of comparable corpora
 in language engineering primarily ensued from the scarcity of
 parallel corpora.  This has motivated research into the use of
 comparable corpora: pairs of monolingual corpora selected according
 to the same set of criteria, but in different languages or language
 varieties.  Non-parallel yet comparable corpora overcome the two
 limitations of parallel corpora, since sources for original,
 monolingual texts are much more abundant than translated texts.
 However, because of their nature, mining translations in comparable
 corpora is much more challenging than in parallel corpora.  What
 constitutes a good comparable corpus, for a given task or per se,
 also requires specific attention: while the definition of a parallel
 corpus is fairly straightforward, building a non-parallel corpus
 requires control over the selection of source texts in both

 With the advent of online data, the potential for building and
 exploring comparable corpora is growing exponentially.  Comparable
 documents in languages that are very different from each other pose
 special challenges as very often, the non-parallel-ness in sentences
 can result from cultural and political differences.


 Following the success of the first workshop on Building and Using
 Comparable Corpora at LREC 2008 in Marrakech, this second workshop
 again brings together language engineers as well as linguists
 interested in the constitution and use of comparable corpora, ranging
 from parallel to non-parallel corpora.  In the larger context of the
 joint ACL-IJCNLP conference, this time the workshop specifically
 aimed to solicit contributions from researchers in different
 geographical regions, in order to highlight in particular the issues
 with comparable corpora across languages that are very different from
 each other, such as across Asian and European languages.  Research in
 minority languages is also of particular interest.  We are very glad
 to include papers on languages as varied as Arabic, Chinese, English,
 French, Japanese, Uyghur and even sign language.

Thursday, August 6, 2009

8:45 Welcome and Introduction

9:00-10:00 Session 1: Invited Presentation
(session chair: P Fung)
9:00 Repetition and Language Models and Comparable Corpora
Ken Church (HLTCOE, Johns Hopkins University, USA)

10:00-10:30 Coffee break

10:30-12:10 Session 2: Information Extraction and Summarization
(session chair: P Zweigenbaum)
10:30 Extracting Lay Paraphrases of Specialized Expressions from
Monolingual Comparable Medical Corpora
Louise Deléger and Pierre Zweigenbaum
10:55 An Extensible Crosslinguistic Readability Framework
Jesse Kirchner, Justin Nuger and Yi Zhang
11:20 An Analysis of the Calque Phenomena Based on Comparable
Marie Garnier and Patrick Saint-Dizier
11:45 Active Learning of Extractive Reference Summaries for Lecture
Speech Summarization
Jian Zhang and Pascale Fung

12:10-13:50 Lunch break

13:50-15:30 Session 3: Statistical Machine Translation
(session chair: R Rapp)
13:50 Train the Machine with What It Can Learn---Corpus Selection
for SMT
Xiwu Han, Hanzhang Li and Tiejun Zhao
14:15 Mining Name Translations from Comparable Corpora by Creating
Bilingual Information Networks
Heng Ji
14:40 Chinese-Uyghur Sentence Alignment: An Approach Based on Anchor
Samat Mamitimin and Min Hou
15:05 Exploiting Comparable Corpora with TER and TERp
Sadaf Abdul Rauf and Holger Schwenk

15:30-16:00 Coffee break

16:00-16:50 Session 4: Building Comparable Corpora / Discussion
(session chair: P Fung)
16:00 Compilation of Specialized Comparable Corpora in French and
Lorraine Goeuriot, Emmanuel Morin and Béatrice Daille
16:25 Toward Categorization of Sign Language Corpora
Jérémie Segouat and Annelies Braffort

16:50   Panel Session
(session chair: P Fung)

Multilingual Information Processing:
from Parallel to Comparable Corpora

Kenneth Ward Church (HLTCOE, Johns Hopkins University, USA)
Hitoshi Isahara (National Institute of Information and
  Communications Technology, Japan)
Kevin Knight (Information Sciences Institute, USC, USA)
Franz Joseph Och (Google Research, USA)
Richard Sproat (OGI School of Science & Technology, OHSU, USA)

17:50 End of Workshop


 Pascale Fung, Hong Kong University of Science & Technology (HKUST)
 Pierre Zweigenbaum, LIMSI-CNRS (France)
 Reinhard Rapp, University of Mainz (Germany)
   and University of Tarragona (Spain)


 Askar Hamdulla (Xinjiang University, China)
 Srinivas Bangalore (AT&T Labs, USA)
 Lynne Bowker (University of Ottawa, Canada)
 Éric Gaussier (Université Joseph Fourier, Grenoble, France)
 Gregory Grefenstette (Exalead, Paris, France)
 Hitoshi Isahara (National Institute of Information and Communications
   Technology, Japan)
 Min-Yen Kan (National University of Singapore)
 Adam Kilgarriff (Lexical Computing Ltd, UK)
 Philippe Langlais (Université de Montréal, Canada)
 Rada Mihalcea (University of North Texas, USA)
 Dragos Stefan Munteanu (Language Weaver, Inc., USA)
 Grace Ngai (Hong Kong Polytechnic University, Hong Kong)
 Carol Peters (ISTI-CNR, Pisa, Italy)
 Serge Sharoff (University of Leeds, UK)
 Richard Sproat (OGI School of Science & Technology, OHSU, USA)
 Mandel Shi (Xiamen University, China)
 Yujie Zhang (National Institute of Information and Communications
   Technology, Japan)


 Ricky Chan Ho Yin (Hong Kong University of Science & Technology)