Elamur Research⌕
☰
home / papers / 2014 / Bidirectionnal converter between syntactic annotations : from French Treebank Dependencies to PASSAGE annotations, and back

Bidirectionnal converter between syntactic annotations : from French Treebank Dependencies to PASSAGE annotations, and back

Summary

Several freely available statistical dependency parsers exist for French. Their output uses the FTB-DEP format, a CoNLL variant derived from the French Treebank. However, there is little gold-standard data in this scheme; in practice, there is only the SEQUOIA corpus. The PASSAGE campaigns produced a much larger manually annotated corpus, but it uses a different formalism with non-recursive chunks and its own set of grammatical relations in XML. The authors present the first version of a converter that works in both directions between these schemes. The FTB-DEP → PASSAGE direction is rule-based and written in Python 2. Each rule specifies a dependency pattern, the corresponding PASSAGE annotation and optional constraints on the arguments. The reverse direction uses pairs of aligned, underspecified template sentences. In these templates, words are replaced by variables constrained by form, lemma or part of speech. Passive constructions get the most detailed treatment, because FTB-DEP distinguishes types of auxiliary verbs and PASSAGE does not. With this converter, parsers can be evaluated on the large PASSAGE gold standard even though they work in a different formalism.

Key points

  • Why it is needed: SEQUOIA contains 3,204 sentences in FTB-DEP format. The PASSAGE corpus has at least 14,000 manually annotated sentences, and 8,200 of them are free of copyright restrictions.
  • Differences between the schemes: PASSAGE has 6 chunk types (GN, GP, NV, GA, GR, PV) and 14 relations. FTB-DEP has no chunks. Instead, it has a virtual ROOT and finer distinctions: AUX-TPS and AUX-PASS, A-OBJ/DE-OBJ/P-OBJ, and P-OBJ-AGT.
  • Passive in the reverse converter: four templates are enough. Action verbs are identified with the LVF lexicon (Dubois & Dubois-Charlier). This makes it possible to recover AUX-PASS and P-OBJ-AGT from the PASSAGE relations AUX-V and CPL-V.
  • Preliminary evaluation of the FTB-DEP → PASSAGE direction: the Berkeley Parser adapted to French was run on 1,584 sentences of the PASSAGE corpus (EP & JRC). AUX-V gave the best result: precision 0.88, recall 0.75, F-measure 0.81.
  • Reverse converter: the 184 examples from the FTB-DEP guidelines include 6 passive sentences, and all 6 are handled correctly. No full evaluation was available at the time of publication. The authors recommend building the mapping between schemes around linguistic phenomena rather than individual relations.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

Fields: Linguistics & languages

Topics: Treebanks

Cite

The paper itself, not this page. BibTeX from the library's record of the work.

@inproceedings{asadullah2014bidirectionnal,
  title = {{Bidirectionnal converter between syntactic annotations : from French Treebank Dependencies to PASSAGE annotations, and back}},
  author = {Asadullah, Munshi and Paroubek, Patrick and Vilnat, Anne},
  booktitle = {Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC'14)},
  year = {2014},
  doi = {10.63317/2ma8vubtr4yy},
  url = {https://doi.org/10.63317/2ma8vubtr4yy}
}

References

19 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Abeillé, A. and Barrier, N. (2004). Enriching a french treebank. In Proceedings of the Fourth International Conference on Language Resources and Evaluation (LREC), Lisbon, Portugal, May. European Language Resources Association. find
  2. Asadullah, M., Vilnat, A., and Paroubek, P. (2013). Converting from the french treebank dependencies into passage syntactic annotations. In Proceedings of the 6th Language and Technology Conference, pages 188–192, Poznan, December. find
  3. Buchholz, S. and Marsi, E. (2006). Conll–x shared task on multilingual dependency parsing. In Proceedings of the Tenth Conference on Computational Natural Language Learning, CoNLL-X ’06, pages 149–164, Stroudsburg, PA, USA. Association for Computational Linguistics. find our card
  4. Candito, M. and Seddah, D. (2012). Le corpus Sequoia : annotation syntaxique et exploitation pour l’adaptation d’analyseur par pont lexical. In Actes de la 19e conférence sur le Traitement Automatique des Langues Naturelles (TALN’2012), pages 321–334, Grenoble, France. find
  5. Candito, M., Crabbé, B., and Denis, P. (2010a). Statistical french dependency parsing: Treebank conversion and first results. In Calzolari, N., Choukri, K., Maegaard, B., Mariani, J., Odijk, J., Piperidis, S., Rosner, M., and Tapias, D., editors, Proceedings of the Seventh International Conference… find
  6. Candito, M., Nivre, J., Denis, P., and Henestroza Anguiano, E. (2010b). Benchmarking of statistical dependency parsers for french. In Proceedings of the 23rd International Conference on Computational Linguistics: Posters, COLING ’10, pages 108–116, Stroudsburg, PA, USA. Association for Computational… find
  7. Candito, M., Nivre, J., Denis, P., and Henestroza Anguiano, E. (2010c). Benchmarking of statistical dependency parsers for french. In Proceedings of the 23rd International Conference on Computational Linguistics: Posters, COLING ’10, pages 108–116, Stroudsburg, PA, USA. Association for Computational… find
  8. Candito, M., Crabbé, B., and Falco, M., (2011). Dépendances syntaxiques de surface pour le français – Schéma d’annotation pour un corpus en dépendances obtenu par conversion du FrenchTreebank, may. find
  9. Carroll, J., Briscoe, T., and Sanfilippo, A. (1998). Parser Evaluation: A Survey and a New Proposal. In Proceedings of the 1st International Conference on Language Resources and Evaluation, pages 447–454, Granada, Spain. find
  10. De la Clergerie, E. V., Hamon, O., Mostefa, D., Ayache, C., Paroubek, P., and Vilnat, A. (2008). find
  11. PASSAGE: from French Parser Evaluation to Large Sized Treebank. In Calzolari, N., Choukri, K., Maegaard, B., Mariani, J., Odijk, J., Piperidis, S., Rosner, M., and Tapias, D., editors, Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC’08), pages 3570– 3577,… find
  12. De Marneffe, M.-C. and Manning, C. D. (2008). The stanford typed dependencies representation. In Coling 2008: Proceedings of the workshop on Cross-Framework and Cross-Domain Parser Evaluation., CrossParser ’08, pages 1–8, Stroudsburg, PA, USA. Association for Computational Linguistics. find our card
  13. Dubois, J. and Dubois-Charlier, F. (1997). Les verbes français. Larousse-Bordas. find
  14. Lin, D. (1998). A dependency-based method for evaluating broad-coverage parsers. Natural Language Engineering, 4(02):97–114, 5. find
  15. Nivre, J. (2006). Inductive Dependency Parsing (Text, Speech and Language Technology). Springer-Verlag New York, Inc., Secaucus, NJ, USA. find
  16. Paroubek, P., Robba, I., Vilnat, A., and Ayache, C. (2006). Data, Annotations and Measures in EASY - the Evaluation Campaign for Parsers of French. In proceedings of the fifth international conference on Language Resources and Evaluation (LREC 2006), pages 315–320, Genoa, Italy, May. ELRA. find
  17. Paroubek, P., De la Clergerie, É., Loiseau, S., Vilnat, A., and Francopoulo, G. (2009). The PASSAGE Syntactic Representation. In Proceedings of the 7th International Workshop on Treebanks and Linguistic Theories, pages 91–102, Gröningenm Pays-Bas, January. Netherlands Graduate Schools of Linguisti… find
  18. Tesnière, L. (1959). Éléments de Syntaxe Structurale. Éditions Klinksieck, Paris, France. find
  19. Vilnat, A., Paroubek, P., de la Clergerie, E. V., Francopoulo, G., and Guénot, M.-L. (2010). PASSAGE Syntactic Representation: a Minimal Common Ground for Evaluation. In Calzolari, N., Choukri, K., Maegaard, B., Mariani, J., Odijk, J., Piperidis, S., Rosner, M., and Tapias, D., editors, Proceedings… find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/asadullah-2014-bidirectionnal-converter-between-syntactic-annotations/