CoNLL Shared Task 2007
The shared task at the Conference on Computational Natural Language Learning (CoNLL) 2007 will for the second year running be devoted to dependency parsing. Unlike in 2006, there will be two separate tracks: a Multilingual Track and a Domain Adaptation Track. The task will be part of the joint EMNLP-CoNLL conference to be held in conjunction with ACL-07 in Prague, June 28-30, 2007.
For more information please read the third and final call and check out the following subpages with
descriptions of the Multilingual Track and Domain Adaptation Track,
instructions on how to download the data,
how to upload your results,
and how to submit the final version of your paper,
an overview table for the languages in the multilingual track,
answers to frequently asked questions,
the data format we are using,
software for evaluation and data conversion,
background information on dependency parsing and domain adaptation,
possible parsing frameworks to choose from,
the timeline,
the scores and outputs of all participating systems, and
the program for the shared task session at EMNLP-CoNLL-2007.
News
Read here for announcements and news regarding the shared task.
May 30 2007
Program for shared task session at EMNLP-CoNLL-2007 published.
May 14 2007
Outputs of all systems added to the score page.
April 25 2007
Instructions for paper submission updated for final version.
April 23 2007
The complete score table published on the website. Corrected gold standard test set for Basque on the data download page.
April 10 2007
Bibliographical references available in both BiBTeX (with minor corrections) and Word on the PaperSubmission page.
April 3 2007
Results sent to participants. Top three scores published on the website.
March 29 2007
Instructions for paper submission published.
March 28 2007
- Deadline for submission of test results extended to April 1, 18.00 EDT (23.00 GMT). The CHILDES test set made optional in the domain adaptation track.
New version of the English and Hungarian test sets released. More information on the data download page.
March 26 2007
Blind test data released.
March 24 2007
Information about download of test data, submission of test results, publication of results, and submission of papers sent to participants.
February 9 2007
New version of the Catalan data set released. More information on the data download page.
New version of the Basque data set released. More information on the data download page.
Updated the overview table.
February 7 2007
New version of the Basque data set released. More information on the data download page.
February 6 2007
New version of the Chinese data set released. More information on the data download page.
February 5 2007
New versions of the Arabic and Czech data sets released. More information on the data download page.
February 2 2007
Added a page with an overview table fo the languages in the multilingual track.
February 1 2007
Basque training data now available at data download page.
January 29 2007
Corrected fax number for LDC on data download page.
January 29 2007
Training data released for all languages but Basque.
January 23 2007
Third and definitely final call sent out. Deadline for registration extended to January 26.
January 15 2007
- Problems with e-mail address (see below).
January 11 2007
Second and final call sent out.
December 20 2006
First call sent out.
December 13 2006
- Webpage has been updated at several places.
Registration
In order to take part in the shared task, participants should register their intent to participate, in either or both of the two tracks, by sending an e-mail to conll07st@uvt.nl before January 26, 2007.
(We have previously had problems with the e-mail address conll07st@uvt.nl. Everything should be working now, but if you do not get a reply to your message please send a message to skuebler@indiana.edu or nivre@msi.vxu.se.)
The Organizers
- Joakim Nivre
- Johan Hall
- Sandra Kübler
Ryan McDonald
- Jens Nilsson
- Sebastian Riedel
- Deniz Yuret
DepParse Wiki
This page is a part of the DepParse wiki which has been set up to provide all kinds of resources (tools, information, links to data) that are useful for research in dependency parsing. As it is a Wiki site, you are encouraged to contribute to it.