Elamur Research⌕
☰
home / papers / 2018 / SzegedKoref: A Hungarian Coreference Corpus

SzegedKoref: A Hungarian Coreference Corpus

Summary

Hungarian had only a small manually annotated coreference dataset (Miháltz, 2012), which was not enough to build machine-learning coreference resolvers. The authors created SzegedKoref by adding coreference annotation to selected documents of the Szeged Treebank, which already contains POS, constituency and dependency layers. They kept longer texts (student essays and newspaper articles) and dropped the very short business news. Two linguistically trained annotators, supervised by an expert, worked in MMAX2. They marked mentions, linked them to antecedents and labelled each relation as pronominal, nominal, adverbial, verbal or derivational. Nominal links also received semantic subtypes such as repetition, synonym, hypernym or apposition. The corpus explicitly handles two Hungarian-specific phenomena. Dropped subject, object and possessive pronouns were inserted automatically by rules before annotation. Pronouns that refer to subordinate clauses, whether overt or dropped, were linked to those clauses. Zero pronouns turned out to be central to Hungarian coreference. The resource is intended as training and test data for future Hungarian coreference systems, with possible uses in information extraction and gender-aware machine translation.

Key points

  • Size: 400 texts, 4,021 sentences and 55,763 tokens. The student essays contribute 45,981 tokens and the newspaper subcorpus 9,782 tokens. The corpus contains 2,456 anaphoric chains.
  • There are 5,385 annotated anaphors. Pronominal anaphors are the most common (34.37%), followed by repetition (23.44%) and synonym (10.79%).
  • 1,243 zero pronouns take part in anaphoric chains: 713 subjects, 190 objects and 340 possessives. Roughly 67% of all pronominal anaphors contain a dropped (zero) pronoun.
  • Inter-annotator agreement on mention identification was 0.95 F-score, measured on 10 doubly annotated documents.
  • The paper describes the corpus only. No coreference resolution model is trained or evaluated; this is left for future work.

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: Coreference resolution

Cited by

9 papers of our library cite it (of the 140673 whose bibliographies we read); 7 of them have a page here, 2 not yet.

Cite

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

@inproceedings{vincze2018szegedkoref,
  title = {{SzegedKoref: A Hungarian Coreference Corpus}},
  author = {Vincze, Veronika and Hegedús, Klára and Sliz-Nagy, Alex and Farkas, Richárd},
  booktitle = {Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)},
  year = {2018},
  doi = {10.63317/5hghr677923d},
  url = {https://doi.org/10.63317/5hghr677923d}
}

References

9 references, 6 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Björkelund, A., Eckart, K., Riester, A., Schauffler, N., and Schweitzer, K. (2014). The Extended DIRNDL Corpus as a Resource for Coreference and Bridging Resolution. In Proceedings of the Ninth International Conference on Language Resources and Evaluation (LREC’14), pages 3222–3228, Reykjavik, Icel… find
  2. Springer. Ghaddar, A. and Langlais, P. (2016). Wikicoref: An english coreference-annotated corpus of wikipedia articles. In Nicoletta Calzolari (Conference Chair), et al., editors, Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC 2016), Paris, France, may.… find our card
  3. Hendrickx, I., Bouma, G., Coppens, F., Daelemans, W., Hoste, V., Kloosterman, G., Mineur, A.-M., Vloet, J. V. D., and Verschelde, J.-L. (2008). A Coreference Corpus and Resolution System for Dutch. In Proceedings of the Sixth International Conference on Language Resources and Evaluation (LREC’08), M… find our card
  4. Iida, R., Komachi, M., Inui, K., and Matsumoto, Y. (2007). *Figure 2: Frequency of anaphor types.* Annotating a Japanese Text Corpus with Predicate-Argument and Coreference Relations. In Proceedings of the Linguistic Annotation Workshop, pages 132–139, Prague, Czech Republic, June. ACL. Miháltz, M.… find our card
  5. Muzerelle, J., Lefeuvre, A., Schang, E., Antoine, J.-Y., Pelletier, A., Maurel, D., Eshkol, I., and Villaneau, J. (2014). ANCOR Centre, a large free spoken French coreference corpus: description of the resource and reliability measures. In Proceedings of the Ninth International Conference on Languag… find our card
  6. Pradhan, S. S., Ramshaw, L., Weischedel, R. M., MacBride, J., and Micciulla, L. (2007). Unrestricted Coreference: Identifying Entities and Events in OntoNotes. In ICSC, pages 446–453. find
  7. IEEE Computer Society. Pradhan, S., Ramshaw, L., Marcus, M., Palmer, M., Weischedel, R., and Xue, N. (2011). CoNLL-2011 Shared Task: Modeling Unrestricted Coreference in OntoNotes. In Proceedings of the Fifteenth Conference on Computational Natural Language Learning: Shared Task, CONLL Shared Task ’… find our card
  8. Pradhan, S., Moschitti, A., Xue, N., Uryupina, O., and Zhang, Y. (2012). CoNLL-2012 shared task: Modeling multilingual unrestricted coreference in OntoNotes. In Proceedings of the Sixteenth Conference on Computational Natural Language Learning (CoNLL 2012), Jeju, Korea. find our card
  9. Weischedel, R., Hovy, E., Marcus, M., Palmer, M., Belvin, R., Pradan, S., Ramshaw, L., and Xue, N. (2011). OntoNotes: A Large Training Corpus for Enhanced Processing. In Handbook of Natural Language Processing and Machine Translation. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/vincze-2018-szegedkoref-hungarian-coreference-corpus/