Elamur Research⌕
☰
home / papers / 2004 / Unsupervised Learning of Contextual Role Knowledge for Coreference Resolution

Unsupervised Learning of Contextual Role Knowledge for Coreference Resolution

Summary

The paper addresses coreference resolution, specifically choosing the right antecedent when an anaphor's own features, such as lexical match or agreement, are not enough. The authors argue that the role a noun phrase plays in an event (for example, kidnapper versus victim) is a strong clue. They present BABAR, a resolver that learns this contextual role knowledge from raw text without annotations. BABAR first gathers training pairs using lexical and syntactic heuristics that resolve easy cases reliably. It then runs AutoSlog to produce extraction patterns, which it normalizes into agent/patient "caseframes." From co-occurrence statistics over these pairs, it builds four contextual role knowledge sources. A Dempster-Shafer model combines their evidence with seven general knowledge sources. Experiments used the MUC-4 terrorism corpus and Reuters natural-disaster texts. The contextual knowledge raised F-measure in both domains, mainly by improving pronoun recall, while definite NPs barely changed. The authors conclude that pronoun resolution relies chiefly on context, whereas definite NP resolution relies more on the meaning of the noun itself.

Key points

  • Automatically generated training resolutions: 5,078 for terrorism (2,386 lexical seeding, 2,692 syntactic seeding) and 20,479 for disasters (11,652 lexical, 8,827 syntactic).
  • Syntactic seeding uses seven heuristics, including reflexive and relative pronouns, appositives, and "NP said it/they" constructions. Lexical seeding links matching proper names and repeated existential definite NPs such as "the FBI."
  • The four contextual role knowledge sources are the Caseframe Network (CFNet), Lexical Caseframe Expectations (CFLex), and two semantic caseframe sources (CFSem-CFSem and WordSem-CFSem). Using all of them together worked best.
  • Pronoun recall improved by 13 points in terrorism and 15 points in disasters. Pronoun F rose from .59 to .68 for terrorism and from .56 to .66 for disasters.
  • For definite NPs, the gains were negligible: F went from .55 to .55 for terrorism and from .58 to .59 for disasters.

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

38 papers of our library cite it (of the 140673 whose bibliographies we read); 16 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{bean2004unsupervised,
  title = {{Unsupervised Learning of Contextual Role Knowledge for Coreference Resolution}},
  author = {Bean, David and Riloff, Ellen},
  booktitle = {Proceedings of the Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics: HLT-NAACL 2004},
  year = {2004},
  url = {https://aclanthology.org/N04-1038/}
}

References

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

  1. J. Allen. 1995. Natural Language Understanding. jamin/Cummings Press, Redwood City, CA. find
  2. C. Aone and S. Bennett. 1995. Applying Machine Learning to Anaphora Resolution. In IJCAI-95 Workshop on New Approaches to Learning for NLP. find
  3. D. Bean and E. Riloff. 1999. Corpus-Based Identification of Non-Anaphoric Noun Phrases. In Proc. of the 37th Annual Meeting of the Association for Computational Linguistics. find our card
  4. I. Dagan and A. Itai. 1990. Automatic Processing of Large Corpora for the Resolution of Anaphora References. In Proceedings of the Thirteenth International Conference on Computational Linguistics (COLING-90), pages 330–332. find our card
  5. T. Dunning. 1993. Accurate methods for the statistics of surprise and coincidence. Computational Linguistics, 19(1):61– 74. find our card
  6. B. Grosz and C. Sidner. 1998. Lost Intuitions and Forgotten Intentions. In M. Walker, A. Joshi, and E. Prince, editors, Centering Theory in Discourse, pages 89–112. Clarendon Press. find
  7. B. Grosz, A. Joshi, and S. Weinstein. 1995. Centering: A Framework for Modeling the Local Coherence of Discourse. Computational Linguistics, 21(2):203–226. find our card
  8. J. Hobbs. 1978. Resolving Pronoun References. 44(4):311–338. find our card
  9. Lingua, A. Kehler. 1997. Probabilistic Coreference in Information Extraction. In Proceedings of the Second Conference on Empirical Methods in Natural Language Processing. find our card
  10. S. Lappin and H. Leass. 1994. An algorithm for pronominal anaphora resolution. Computational Linguistics, 20(4):535– 561. find our card
  11. J. McCarthy and W. Lehnert. 1995. Using Decision Trees for Coreference Resolution. In Proc. of the Fourteenth International Joint Conference on Artificial Intelligence. find
  12. G. Miller. 1990. Wordnet: An On-line Lexical Database. International Journal of Lexicography, 3(4). find
  13. MUC-4 Proceedings. 1992. Proceedings of the Fourth Message Understanding Conference (MUC-4). find
  14. MUC-6 Proceedings. 1995. Proceedings of the Sixth Message Understanding Conference (MUC-6). find
  15. MUC-7 Proceedings. 1998. Proceedings of the Seventh Message Understanding Conference (MUC-7). find
  16. V. Ng and C. Cardie. 2002. Improving Machine Learning Approaches to Coreference Resolution. In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics. find our card
  17. G. Niyu, J. Hale, and E. Charniak. 1998. A statistical approach to anaphora resolution. In Proceedings of the Sixth Workshop on Very Large Corpora. find our card
  18. E. Riloff. 1996. An Empirical Study of Automated Dictionary Construction for Information Extraction in Three Domains. Artificial Intelligence, 85:101–134. find
  19. W. Soon, H. Ng, and D. Lim. 2001. A Machine Learning Approach to Coreference of Noun Phrases. Computational Linguistics, 27(4):521–541. find
  20. Ben-M. Stefik. 1995. Introduction to Knowledge Systems. Morgan Kaufmann, San Francisco, CA. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/bean-2004-unsupervised-learning-contextual-role-knowledge/