User Tools

Site Tools


SPMRL 2014 SHARED TASK on Parsing of Morphologically-Rich Languages

1. Shared task description
Update Feb. 2015: the link to the spmrl's evalb version pointed to an outdated version. please update!!

2. How to obtain licenses for the shared task data
(update: Korean unlabeled data name and url fix)

3. Schedule

4. Download (requires an account, see 2.)
July 19: release of test data
July 9; Patches for French (Adding mwe predictions on pred files)
July 9: Patches for Korean (fix some missing sentences separator and remove 55k faulty sentences – beyond repair –)
July 6: Patches for Polish (add parses to unlabeled data and explicit feature names)
June 28: French unlabeled data set (morph+mwe annotation, parses upcoming)
June 24: patch for Korean unlabeled data available (fix some tokenization errors)
June 17: - Hebrew unlabeled data correction (tokenization, POS tagset)
- German unlabeled data correction (annotation scheme)
- Swedish unlabeled data correction (feature names) + scripts
- Hungarian, script to fix incompatibilities between orig FPOS and Spmrl 2013 Hungarian data set

5. Frequently Asked Questions

6. Committee page (for organizers only)

7. Mailing List Archives

8. ** Official Results page **

10. Paper Submission Guidelines

11. Acknowledgments

start.txt · Last modified: 2017/03/10 14:53 by seddah