Elamur Research⌕
☰
home / papers / 2016 / Coreference in Prague Czech-English Dependency Treebank

Coreference in Prague Czech-English Dependency Treebank

Summary

Manually coreference-annotated parallel corpora are scarce, which hampers cross-lingual discourse research, including work on machine translation that often ignores context beyond the sentence. The authors present PCEDT 2.0 Coref, an extension of the Prague Czech-English Dependency Treebank 2.0. That treebank has about 1.2 million tokens in nearly 50,000 sentence pairs. Its English side is the Wall Street Journal part of the Penn Treebank, and its Czech side was translated by hand. Links are annotated on the tectogrammatical (deep-syntax) layer, so omitted arguments such as Czech null subjects become explicit nodes that can take part in coreference. The earlier release already covered grammatical and pronominal coreference. This version adds manual nominal coreference. For English, a converted BBN-PCETC annotation was checked and corrected by hand. The paper explains the scheme, reports statistics, and compares it with Ontonotes and PDT 3.0. It also introduces a supervised method for aligning coreferential expressions across languages, which yields more reliable Czech–English correspondence statistics. Data and code are public.

Key points

  • Base resource: PCEDT 2.0, a Czech–English parallel treebank with about 1.2M tokens and about 50k sentence pairs. The English part comes from the Penn Treebank WSJ section, and the Czech part is a sentence-by-sentence human translation.
  • The scheme distinguishes two types of coreference:
    • Grammatical coreference: relative and reflexive pronouns, control verbs, reciprocals, and dual-dependency modifiers.
    • Textual coreference: pronouns, textual ellipsis, and nominal coreference. It also includes special links for multi-sentence discourse antecedents and exophora.
  • Textual links follow a chain principle, where each mention points to the nearest preceding antecedent. No ambiguity is encoded, and manual textual links are limited to a 20-sentence distance.
  • Standard unsupervised word aligners handle pronouns and function words poorly. The paper therefore uses a supervised aligner trained on the manual alignment by Novák and Nedoluzhko (2015).
  • The paper compares PCEDT 2.0 Coref with ParCor 1.0 (German–English, pronouns only) and with CzEng 1.0, whose coreference is automatic. It is released at ufal.cz/pcedt2.0-coref, with code at github.com/ufal/pcedt2.0-coref.

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 · Treebanks

Cited by

23 papers of our library cite it (of the 140673 whose bibliographies we read); 8 of them have a page here, 15 not yet.

Cite

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

@inproceedings{nedoluzhko2016coreference,
  title = {{Coreference in Prague Czech-English Dependency Treebank}},
  author = {Nedoluzhko, Anna and Novák, Michal and Cinková, Silvie and Mikulová, Marie and Mírovský, Jiří},
  booktitle = {Proceedings of the Tenth International Conference on Language Resources and Evaluation (LREC'16)},
  year = {2016},
  doi = {10.63317/3yqbenvhok5r},
  url = {https://doi.org/10.63317/3yqbenvhok5r}
}

References

13 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. BBN Technologies, (2006). Co-reference Guidelines for English OntoNotes. Bojar, O., Žabokrtský, Z., Dušek, O., Galuščáková, P., Majliš, M., Mareček, D., Maršı́k, J., Novák, M., Popel, M., and Tamchyna, A. (2012). The Joy of Parallelism with CzEng 1.0. In Proceedings of LREC 2012, Istanbul… find
  2. Friedrich, A., Palmer, A., Sørensen, M. P., and Pinkal, M. (2015). Annotating genericity: a survey, a scheme, and a corpus. In Proceedings of the 9th Linguistic Annotation Workshop (LAW IX), Denver, Colorado, US. Association for Computational Linguistics. find
  3. Guillou, L., Hardmeier, C., Smith, A., Tiedemann, J., and Webber, B. (2014). ParCor 1.0: A Parallel Pronoun-Coreference Corpus to Support Statistical MT. In Proceedings of the 9th International Conference on Language Resources and Evaluation (LREC-2014), Reykjavik, Iceland. European Language Resourc… find our card
  4. Nedoluzhko, A., Mı́rovský, J., and Novák, M. (2013). A Coreferentially annotated Corpus and Anaphora Resolution for Czech. In Computational Linguistics and Intellectual Technologies, Moscow, Russia. find
  5. ABBYY. Nedoluzhko, A. (2013). Generic noun phrases and annotation of coreference and bridging relations in the Prague Dependency Treebank. In Proceedings of the 7th Linguistic Annotation Workshop & Interoperability with Discourse, Sofija, Bulgaria. Omnipress, Inc. Novák, M. and Nedoluzhko, A. (2015… find
  6. Och, F. J. and Ney, H. (2000). Improved Statistical Alignment Models. In Proceedings of the 38th Annual Meeting on Association for Computational Linguistics, Stroudsburg, PA, USA. Association for Computational Linguistics. Pajas, P. and Štěpánek, J. (2008). Recent Advances in a Feature-rich Frame… find our card
  7. Postolache, O., Cristea, D., and Orasan, C. (2006). Transferring Coreference Chains through Word Alignment. In Proceedings of the Fifth International Conference on Language Resources and Evaluation, Genoa, Italy. find our card
  8. European Language Resources Association. Quirk, R., Greenbaum, S., Leech, G., and Svartvik, J. (1985). A Comprehensive Grammar of the English Language. Longman, London, United Kingdom. Sgall, P., Hajičová, E., and Panevová, J. (1986). The Meaning of the Sentence in Its Semantic and Pragmatic Aspe… find
  9. D. Reidel Publishing Company, Dordrecht, Netherlands. Sgall, P. (1967). Generativnı́ popis jazyka a česká dekli-nace. Academia, Prague, Czech Republic. Urešová, Z., Dušek, O., Fučı́ková, E., Hajič, J., and Šindlerová, J. (2015). Bilingual English-Czech Valency Lexicon Linked to a Parallel… find
  10. Hajič, J., Panevová, J., Hajičová, E., Sgall, P., Pajas, P., Štěpánek, J., Havelka, J., Mikulová, M., Žabokrtský, Z., Ševčı́ková-Razı́mová, M., and Urešová, Z. (2006). Prague Dependency Treebank 2.0. Linguistic Data Consortium, URL: http://hdl.handle. net/11858/00-097C-0000-0001-B098… ↗
  11. Marcus, M., Santorini, B., Marcinkiewicz, M. A., and Taylor, A. (1999). Penn Treebank 3. Linguistic Data Consortium, URL: https://catalog.ldc.upenn.edu/LDC99T42. Nedoluzhko, A., Novák, M., Cinková, S., Mikulová, M., and Mı́rovský, J. (2016). Prague Czech-English Dependency Treebank 2.0 Coref. C… ↗
  12. Weischedel, R. and Brunstein, A. (2005). BBN Pronoun Coreference and Entity Type Corpus. Linguistic Data Consortium, ISLRN: 141-282-691-413-2. find
  13. Language Resource References Bejček, E., Hajičová, E., Hajič, J., Jı́nová, P., Kettnerová, V., Kolářová, V., Mikulová, M., Mı́rovský, J., Nedoluzhko, A., Panevová, J., Poláková, L., Ševčı́ková, M., Štěpánek, J., and Zikánová, Š. (2013). Prague Dependency Treebank 3.0. Charles… find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/nedoluzhko-2016-coreference-prague-czech-english-dependency/