[Corpora-List] LREC 2004 Workshop CFP: Building Lexical Resources from Semantically Annotated Corpora

Katrin Erk erk at CoLi.Uni-SB.DE
Thu Jan 29 11:36:45 UTC 2004


(Apologies for multiple postings)

	       First Call for Papers
	WORKSHOP ANNOUNCEMENT AND CALL FOR PAPERS

	 Building Lexical Resources from Semantically
		  Annotated Corpora

In association with the Fourth International
Conference on Language Resources and Evaluation -
LREC 2004

Centro Cultural de Belem, Lisbon, Portugal
May 30, 2004

BACKGROUND

Over the last several decades, print dictionaries have
largely shifted from being based on previous
dictionaries, citation slips and lexicographer's
intuitions, to using sophisticated corpus searches for
discovering and recording the actual ways in which
words are used.  A similar change is under way in the
field of on-line lexical resources, as people seek a
middle way between massive word crunching at one
extreme and hand-built entries based on native speaker
intuitions at the other.

Since 1997, the FrameNet project has built a body of
more than 120,000 corpus examples annotated with
finely detailed semantic roles (based on frame
semantics) and deriving from these a lexicon with rich
semantic/syntactic descriptions of roughly 7,000
lexical units. Projects with goals very similar to
FrameNet are under way for German at the University of
the Saarland (SALSA), for Spanish in Barcelona, and
for Japanese in Tokyo, and independent but allied
efforts are in progress for English at U Penn
(PropBank) and for Czech in Prague (Prague treebank).

This workshop will bring together researchers building
and using such lexica to report on recent work and to
share ideas for future directions.

PAPER TOPICS

Papers are invited on questions such as:

What sorts of lexical information will be most useful
for NLP applications, commercial lexicography, or
language pedagogy? (And what's a reasonable balance
between what we would like and what we can
cost-effectively get?)

To what extent can manual annotation of corpus
examples be facilitated or replaced by automatic
processes?

How can the lexica be used to guide semantic
parsing/role labeling in unrestricted text?

What are the advantages and disadvantages of the
differing approaches of the various projects?  What
are the possibilities of adopting a common program and
setting up the means of cooperating closely?

How do the semantic role concepts underlying different
annotation practices compare? What are their
respective advantages and disadvantages? Is it
possible to align them and to map between them?

PAPER SELECTION CONSIDERATIONS

We would like the workshop to contain a balance of
papers dealing with

(1) lexicon building of the kind supported by corpus
analysis,  (2) methods of semantic and/or functional
annotation,

(3) ontological issues with frame structures and the
elements of frames,

(4) outreach, including NLP applications, the
suitability of current procedures to the terminologies
of scientific and technical discourse, and the special
problems related to the lexica and grammatical
structures of different languages,

(5) the potential for achieving some level of
cross-linguistic and cross-framework standardization
of annotation and analysis practices.

TARGET AUDIENCE

Builders and potential users of
semantically/functionally annotated corpora.

IMPORTANT DATES

Abstract Submission Deadline:		February 29, 2004
Notification:				March 21, 2004
Camera Ready Papers:			April 21, 2004
Workshop:				May 30, 2004

ABSTRACT SUBMISSION

Abstracts should be no more than 1000 words in length,
in .pdf format (preferred) or plain ASCII text. The
abstract should be emailed to Collin Baker
(collinb at icsi.berkeley.edu) with the subject line
"CORPUS ANNOTATION WORKSHOP". The body of the email
should begin with the author's name(s) and
affiliation(s), and the email address of the contact
person. The attachment should contain only the text of
the submission.

WORKSHOP CHAIRS

Charles Fillmore
Manfred Pinkal
Collin Baker
Katrin Erk


PROGRAM COMMITTEE

Collin Baker
Katrin Erk
Charles Fillmore
Daniel Gildea
Eva Hajicova
Ulrich Heid
Mirella Lapata
Martha Palmer
Manfred Pinkal



More information about the Corpora mailing list