CoNLL-X Shared Task: Multi-lingual Dependency Parsing

Tenth Conference on Computational Natural Language Learning - New York City, June 8-9, 2006

Contents

Final release notes

Back

Release notes

Final release (January 9 2006)

We are happy to announce the final release of the CoNLL-X Shared Task material. This release includes training data sets and documentation for the following languages:

  1. Arabic
  2. Bulgarian
  3. Czech
  4. Chinese
  5. Danish
  6. Dutch
  7. German
  8. Japanese
  9. Portuguese
  10. Slovene
  11. Spanish
  12. Swedish
  13. Turkish

For some of these data, you need to send us a signed copy of the relevant license first. Once we have received all your signed licenses, we will email you instructions for downloading these data sets. See Obtain data for detailed information.

Included in this release are scripts for evaluation and conversion; see Download software. See also our Framework page, which was recently extended with a link to download the CaboCha parser. We expect more software to become available soon.

If you encounter any bugs or errors, you can check the BugReport page to see if they are already known. If not, please notify us at conll06st@uvt.nl. If anything else is unclear, do not hesitate to contact us.

We hope you will enjoy working on this year's shared task. We certainly did! Good luck to all of you,

Sabine Buchholz
Amit Dubey
Yuval Krymolowski
Erwin Marsi