Comparing Corpora

A workshop to be held in conjunction with
The 38th Annual Meeting of the Association for Computational Linguistics
7th October 2000
Hong Kong Convention and Exhibition Centre
Hong Kong University of Science and Technology

PROGRAMME here
Registration details: http://www.cs.ust.hk/acl2000/.

Theme

Anyone who has worked with corpora will be all too aware of differences between them. Depending on the differences, it may, or may not, be reasonable to expect results based on one corpus to also be valid for another. It may, or may not, be appropriate for a grammar, or parser, based on one to perform well on another. It may, or may not, be straightforward to port an application from a domain of the first text type to a domain of the second. Currently, characterisations of corpora are mostly textual and at different levels of generality. A corpus is described as ``Wall Street Journal'' or ``transcripts of business meetings'' or ``foreign learners' essays (intermediate grade)''. It would be desirable to be able to place a new corpus in relation to existing ones, and to be able to quantify similarities and differences.

Allied to corpus-similarity is corpus-homogeneity. An understanding of homogeneity is a prerequisite to a measure of the similarity -- it makes little sense to compare a corpus sampled across many genres, like the Brown, with a corpus of weather forecasts, without first accounting for the one being broad, the other narrow.

Given the centrality of corpora to contemporary language engineering, it is remarkable how little research there has been to date on the question. Biber's work, coming from sociolinguistics, has made a considerable impact, with various researchers in computational lingustics taking forward the model (Biber 1989, 1995). Studies in text classification, genre and sublanguage are also salient, but it is rarely evident how well the technologies ddeveloped in these fields are suited to measuring corpus similarity or homogeneity.

The workshop will welcome contributions concerned with measuring and comparing corpora using quantitative methods, from any field.

Where and when

The workshop will last half a day and will be on the afternoon of the 7th Oct, the main ACL conference being 3rd-6th October. The venue will be the Hong Kong University of Science and Technology.

Submissions

Submissions are limited to original, unpublished work. Papers may not exceed 3200 words (exclusive of title page and references). They must be received by July 8, 2000, in hard copy (4 copies) OR postscript OR rtf format. Electronic delivery is to

compcorp@itri.brighton.ac.uk

and hard copies are to be mailed to

COMPCORP submission
ITRI
University of Brighton
Lewes Road
Brighton BN2 4GJ
United Kingdom

Important Dates

July 8, 2000	Submission (of full-length paper)
August 17, 2000	Acceptance notice
September 1, 2000	Camera-ready paper received
October 7 or 8	Workshop date

Co-ordinators

Adam Kilgarriff - University of Brighton, UK
Tony Berber Sardinha - Catholic University of Sao Paulo, Brazil

Programme committee

Douglas Biber	Northern Arizona University
Jeremy Clear	University of Birmingham
Ted Dunning	MusicMatch Software, Inc.
Tomaz Erjavec	Jozef Stefan Institute, Slovenia
Pascale Fung	University of Science and Technology, Hong Kong
Greg Grefenstette	Xerox Research Centre Europe
Benoit Habert	LIMSI, France
Przemek Kaszubski	Adam Mickiewicz University, Poland
Adam Kilgarriff	University of Brighton
David Lee	University of Lancaster
Oliver Mason	University of Birmingham
Doug Oard	University of Maryland
Tony Rose	Canon Research
Tony Berber Sardinha	Catholic University of Sao Paulo, Brazil
George Tambouratzis	ILSP, Athens
Christopher Tribble	King's College, London University

Adam Kilgarriff

Last modified: Tue Sep 5 22:33:34 BST 2000