Elamur Research⌕
☰
home / papers / 2009 / EM Works for Pronoun Anaphora Resolution

EM Works for Pronoun Anaphora Resolution

Summary

The paper tackles English personal-pronoun anaphora resolution: deciding whether a pronoun refers back to an earlier noun phrase and, if so, which one. The authors build a generative model in which a pronoun is either non-anaphoric or produced by an antecedent NP from the current or two preceding sentences. The antecedent then generates the pronoun's person, number, gender and its governor plus syntactic relation. Almost all parameters are estimated with Expectation Maximization on a large body of parsed but otherwise unlabeled text. Training starts from uniform distributions and moves gradually from simple to richer models. Non-anaphoric pronouns are detected jointly with resolution, and gender is learned without any external lexicon. Evaluated on the Ge et al. (1998) annotated WSJ data against publicly available resolvers, the system clearly comes out on top. The result is notable because EM has failed on many NLP tasks, and the authors released a fast, working pronoun resolver for others to use.

Key points

  • Accuracy on the test data: 0.686 for the proposed system, versus 0.593 for OpenNLP, 0.534 for GuiTAR (which cannot handle possessives) and 0.529 for JavaRAP (third-person pronouns only). BART was run, but its scores were implausibly low, so it was dropped from the comparison.

  • To keep the comparison fair, every system was run on Charniak parse trees with gold sentence segmentation, using each system's standard distributed models.

  • Several modelling details made training work. The first, simplest model (antecedent probability by sentence distance) is trained for four EM iterations before more parameters are introduced. Governor/relation probabilities use Kneser-Ney smoothing and are tempered with an nth root to control their very wide dynamic range.

  • Unlike Cherry and Bergsma (2005), the model uses syntactic role distinctions and covers first- and second-person pronouns. It also needs no separate preprocessing step for non-anaphoric pronouns or gender.

  • Known limitations include:

    • no handling of cataphora;
    • a two-sentence window for antecedents;
    • no conjoined-NP antecedents for plural pronouns;
    • no speaker identification in quotes, which causes many first-person errors.

    The code, model and Ge corpus are distributed by BLLIP.

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

35 papers of our library cite it (of the 140673 whose bibliographies we read); 13 of them have a page here, 22 not yet.

Cite

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

@inproceedings{charniak2009em,
  title = {{EM Works for Pronoun Anaphora Resolution}},
  author = {Charniak, Eugene and Elsner, Micha},
  booktitle = {Proceedings of the 12th Conference of the European Chapter of the ACL (EACL 2009)},
  year = {2009},
  doi = {10.3115/1609067.1609083},
  url = {https://doi.org/10.3115/1609067.1609083}
}

References

26 references, 12 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Roxana Angheluta, Patrick Jeuniaux, Rudradeb Mitra, and Marie-Francine Moens. 2004. Clustering algorithms for noun phrase coreference resolution. In Proceedings of the 7es Journes internationales d’Analyse statistique des Donnes Textuelles, pages 60–70, Louvain La Neuve, Belgium, March 10–12. find
  2. Amit Bagga and Breck Baldwin. 1998. Algorithms for scoring coreference chains. In In The First International Conference on Language Resources and Evaluation Workshop on Linguistics Coreference, pages 563–566. find our card
  3. P. Brown, S. Della Pietra, V. Della Pietra, and R. Mercer. 1993. The mathematics of statistical machine translation: Parameter estimation. Computational Linguistics, 19(2). find
  4. Donna K. Byron. 2002. Resolving pronominal reference to abstract entities. In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics (ACL2002), pages 80– 87, Philadelphia, PA, USA, July 6–12. find
  5. Claire Cardie and Kiri Wagstaff. 1999. Noun phrase coreference as clustering. In In Proceedings of EMNLP, pages 82–89. find our card
  6. Eugene Charniak and Mark Johnson. 2005. Coarse-to-fine n-best parsing and MaxEnt discriminative reranking. In Proc. of the 2005 Meeting of the Assoc. for Computational Linguistics (ACL), pages 173–180. find our card
  7. Colin Cherry and Shane Bergsma. 2005. An Expectation Maximization approach to pronoun resolution. In Proceedings of the Ninth Conference on Computational Natural Language Learning (CoNLL-2005), pages 88–95, Ann Arbor, Michigan, June. Association for Computational Linguistics. find our card
  8. Michael Collins and Yorav Singer. 1999. Unsupervised models for named entity classification. In Proceedings of the Joint SIGDAT Conference on Empirical Methods in Natural Language Processing and Very Large Corpora (EMNLP 99). find
  9. Niyu Ge, John Hale, and Eugene Charniak. 1998. A statistical approach to anaphora resolution. In Proceedings of the Sixth Workshop on Very Large Corpora, pages 161–171, Orlando, Florida. Harcourt Brace. find our card
  10. Aria Haghighi and Dan Klein. 2007. Unsupervised coreference resolution in a nonparametric Bayesian model. In Proceedings of the 45th Annual Meeting of the Association for Computational Linguistics, pages 848–855. Association for Computational Linguistics. find our card
  11. Andrew Kehler, Douglas Appelt, Lara Taylor, and Aleksandr Simma. 2004a. Competitive self-trained pronoun interpretation. In Daniel Marcu Susan Dumais and Salim Roukos, editors, HLT-NAACL 2004: Short Papers, pages 33–36, Boston, Massachusetts, USA, May 2 - May 7. Association for Computational Linguis… find
  12. Andrew Kehler, Douglas E. Appelt, Lara Taylor, and Aleksandr Simma. 2004b. The (non)utility of predicate-argument frequencies for pronoun interpretation. In Proceedings of the 2004 Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics, pa… find
  13. Shalom Lappin and Herber J. Leass. 1994. An algorithm for pronouminal anaphora resolution. Computational Linguistics, 20(4):535–561. find
  14. Min-Yen Kan Long Qiu and Tat-Seng Chua. 2004. A public reference implementation of the RAP anaphora resolution algorithm. In Proceedings of the Fourth International Conference on Language Resources and Evaluation, volume I, pages 291– 294. find our card
  15. David McClosky, Eugene Charniak, and MarkJohnson. 2008. BLLIP North American News Text, Complete. Linguistic Data Consortium. LDC2008T13. find
  16. Bernard Merialdo. 1991. Tagging text with a probabilistic model. In International Conference on Speech and Signal Processing, volume 2, pages 801–818. find
  17. Ruslan Mitkov, Richard Evans, and Constantin Orăsan. 2002. A new, fully automatic version of Mitkov’s knowledge-poor pronoun resolution method. In Proceedings of the Third International Conference on Intelligent Text Processing and Computational Linguistics (CICLing-2002), Mexico City, Mexico, Febr… find
  18. Thomas Morton, Joern Kottmann, Jason Baldridge, and Gann Bierner. 2005. Opennlp: A java-based nlp toolkit. http://opennlp.sourceforge.net. ↗
  19. Massimo Poesio and Mijail A. Kabadjov. 2004. A general-purpos, of-the-shelf anaphora resolution module: implementataion and preliminary evaluation. In Proceedings of the 2004 international Conference on Language Evaluation and Resources, pages 663,668. find
  20. Massimo Poesio and Renata Vieira. 1998. A corpus-based investigation of definite description use. Computational Linguistics, 24(2):183–216. find
  21. Wee Meng Soon, Hwee Tou Ng, and Daniel Chung Yong Lim. 2001. A machine learning approach to coreference resolution of noun phrases. Computational Linguistics, 27(4):521–544. find our card
  22. Joel R. Tetreault. 2001. A corpus-based evaluation of centering and pronoun resolution. Computational Linguistics, 27(4):507–520. find our card
  23. Yannick Versley, Simone Ponzetto, Massimo Poesio, Vladimir Eidelman, Alan Jern, Jason Smith, Xiaofeng Yang, and Alessandro Moschitti. 2008. Bart: A modular toolkit for coreference resolution. In Companion Volume of the Proceedings of the 46th Annual Meeting of the Association for Computational Lingu… find our card
  24. Marilyn A. Walker. 1989. Evaluating discourse processing algorithms. In ACL, pages 251–261. find
  25. Xiaofeng Yang, Jian Su, Guodong Zhou, and Chew Lim Tan. 2004. Improving pronoun resolution by incorporating coreferential information of candidates. In Proceedings of the 42nd Annual Meeting of the Association for Computational Linguistics (ACL2004), pages 127–134, Barcelona, Spain, July 21–26. find our card
  26. Xiaofeng Yang, Jian Su, and Chew Lim Tan. 2006. Kernel-based pronoun resolution with structured syntactic knowledge. In Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, pages 41–48, Sydney, Austral… find our card

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/charniak-2009-em-works-pronoun-anaphora-resolution/