CFP: Multiword expressions: hard going or plain sailing?

m.b.villada.moiron at rug.nl m.b.villada.moiron at rug.nl
Tue Sep 18 19:56:41 UTC 2007


[Apologies for multiple postings]




Multiword expressions: hard going or plain sailing?


Special issue of the International Journal of Language Resources and
Evaluation   http://www.springer.com/journal/10579/ 


Guest editors:

Paul Rayson (Lancaster University, UK)

Begoña Villada Moirón (University of Groningen, The Netherlands)

Serge Sharoff (University of Leeds, UK)

Scott Piao (University of Manchester, UK)

Stefan Evert (University of Osnabrueck, Germany)


This special issue is concerned with language resources and evaluation
in the area of multiword expressions. For a number of years, the focus
in the natural language processing (NLP) community on the problems
posed by multiword expressions (MWE) was on English. Recently, for
example at the ACL and EACL workshops on multiword expressions,
attention has expanded to other languages including Dutch, Chinese,
Japanese, German, Estonian, Russian, Basque, Turkish and Hindi. This
necessitates a re-evaluation of earlier rule-based, statistical and
hybrid techniques for MWE extraction and classification. In English,
MWE types such as phrasal verbs, light verb constructions, noun
compounds, proper names, and non-compositional idioms, have been
considered. However, in other languages some MWE types can be
represented as compound words, e.g. phrasal verbs in English are
generally expressed as verb-prefix in Russian. At the same time,
research on MWEs for languages other than English is confronted with
new problems, such as the number of word forms per lemma, case
marking, word order or word segmentation.

The focus of this special issue is on the acquisition and analysis of
language resources related to MWE and the methods for evaluation of
the extraction procedures and resulting MWE resources. Language
resources include written or spoken corpora marked up for MWE,
terminology or domain specific databases and dictionaries of MWE, as
well as software tools for their acquisition and analysis.


Topics to be addressed include, but are not limited to:

1.Linguistic analysis: of MWE based on language resources (such as
corpora) and the impact that these studies have on NLP
applications. We also welcome articles which undertake
cross-linguistic analyses of MWE or which identify variation across
languages. Here we include studies which investigate how application
of techniques developed for one language can be transferred to another
language and how successful bilingual (or multilingual) approaches
are.

2.Typologies of MWE: descriptions of different classes of MWE
and their representation in language resources, and in addition to
evaluate how well computational techniques transfer across different
types of MWE. Investigations of the variability of MWEs.  

3.Extraction methods: Do methods generalise across languages? What is
the interaction between linguistic descriptions/analyses of MWE and
extraction methods? (i.e. to what extent are linguistically informed
extraction methods useful?) Is fully automatic extraction of MWE
feasible, or will manual validation/intervention always be necessary?

4.Evaluation strategies: creation of gold standard MWE language
resources. Comparative studies using human subjects including experts
and non-experts and computational evaluations using language resources
derived from the web and elsewhere. Task-based evaluation of MWE
lexical resources in NLP applications.

5.Compositionality: comparative evaluation of how well humans and
machines make reliable judgements on compositionality for the various
classes of MWE. Also, the extent to which compositionality is a key
indicator for extraction and classification of MWE.

6.Applications: of the theories, techniques and tools developed in MWE
research to practical tasks such as IR, Text Mining, etc.




Important Dates

- Deadline for paper submission: 15th November 2007

- Notification of acceptance: 15th May 2008

- Camera-ready version of accepted paper: 15th July 2008

- Target publication date: 4th quarter 2008



Instructions for Authors

Submissions should be not more than 20 pages long, must be in English,
and follow the submission guidelines at

http://www.springer.com/cda/content/document/cda_downloaddocument/instr_print_10579.060421.pdf?SGWID=0-0-45-126854-p35554703

Extended and revised versions of papers accepted at previous ACL and
EACL workshops on multiword expressions, e.g. Workshops held in Trento
(April 2006), Sydney (July 2006) and in Prague (June 2007) are
encouraged.

Authors are advised to use the online manuscript submission for the
journal. Authors are also encouraged to send a brief email to Paul
Rayson (paul at comp.lancs.ac.uk) indicating their intention to
participate as soon as possible, including their contact information
and the topic they intend to address in their submissions. Enquiries
regarding the special issue should be sent to the same address.



More information about the HPSG-L mailing list