Elamur Research⌕
☰
home / papers / 2004 / Learning Noun Phrase Anaphoricity to Improve Conference Resolution: Issues in Representation and Optimization

Learning Noun Phrase Anaphoricity to Improve Conference Resolution: Issues in Representation and Optimization

Summary

In principle, a coreference system could skip noun phrases that are not anaphoric. However, earlier attempts to add automatically predicted anaphoricity to such systems often lowered performance. Ng studies two design choices that had received little attention. The first is representation: anaphoricity can act as a hard filter (constraint-based) or as one more feature for the coreference learner. The second is optimization: the anaphoricity classifier can be tuned on its own (local) or tuned for end coreference performance (global). For global optimization, he proposes a conservativeness-based anaphoricity model with a single tunable parameter, chosen on held-out development data. That parameter is either a misclassification cost ratio (RIPPER) or a probability threshold (MaxEnt). He tests all four combinations on the three ACE data sets, using both RIPPER and MaxEnt as coreference learners. The constraint-based, globally optimized variant (CBGO) works best. It beats a no-anaphoricity baseline and the usual locally optimized filter without needing extra labeled data. This shows that anaphoricity filtering pays off only when it is tuned to the downstream task.

Key points

  • Data and learners: The experiments use the ACE data sets BNEWS, NPAPER and NWIRE. Coreference follows the classification-plus-clustering framework of Soon et al. (2001) and Ng and Cardie (2002b).
  • Tuning the filter: Local optimization uses default settings (cost ratio cr = 1, threshold t = 0.5). The globally tuned values were always more conservative, which means the default filters label too many NPs as non-anaphoric.
  • Precision–recall trade-off: On NPAPER development data with RIPPER/RIPPER, raising cr increases recall and lowers precision. F-measure peaks at cr = 4.
  • Feature-based variant: The feature-based global variant (FBGO) did worse than CBGO in 11 of 12 settings. It significantly improved the baseline only four times and was no better than the local approaches.
  • Most informative features: By information gain, the strongest anaphoricity features are HEAD MATCH, STR MATCH and PRONOUN. PROPER NOUN, BARE SINGULAR/BARE PLURAL and ARTICLE features, which reflect definiteness, also carry substantial information.

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

10 papers of our library cite it (of the 140673 whose bibliographies we read); 8 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. DOI 10.3115/1218955.1218975

@inproceedings{ng2004learning,
  title = {{Learning Noun Phrase Anaphoricity to Improve Conference Resolution: Issues in Representation and Optimization}},
  author = {Ng, Vincent},
  booktitle = {Proceedings of the 42nd Annual Meeting of the Association for Computational Linguistics (ACL-04)},
  year = {2004},
  doi = {10.3115/1218955.1218975},
  url = {https://doi.org/10.3115/1218955.1218975}
}

References

22 references, 8 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. David Bean and Ellen Riloff. 1999. Corpus-based identification of non-anaphoric noun phrases. In Proceedings of the ACL, pages 373–380. find our card
  2. Adam L. Berger, Stephen A. Della Pietra, and Vincent J. Della Pietra. 1996. A maximum entropy approach to natural language processing. Computational Linguistics, 22(1):39–71. find our card
  3. Stanley Chen and Ronald Rosenfeld. 2000. A survey of smoothing techniques for ME models. IEEE Transactions on Speech on Audio Processing, 8(1):37–50. find
  4. William Cohen. 1995. Fast effective rule induction. In Proceedings of ICML. find
  5. Stephen Della Pietra, Vincent Della Pietra, and John Lafferty. 1997. Inducing features of random fields. IEEE Transactions on Pattern Analysis and Machine Intelligence, 19(4):380–393. find
  6. Michel Denber. 1998. Automatic resolution of anaphora in English. Technical report, Eastman Kodak Co. find
  7. Richard Evans. 2001. Applying machine learning toward an automatic classification of it. Literary and Linguistic Computing, 16(1):45–57. find
  8. Christopher Kennedy and Branimir Boguraev. 1996. Anaphor for everyone: Pronominal anaphora resolution without a parser. In Proceedings of COLING, pages 113–118. find
  9. Shalom Lappin and Herbert Leass. 1994. An algorithm for pronominal anaphora resolution. Computational Linguistics, 20(4):535–562. find our card
  10. Ruslan Mitkov, Richard Evans, and Constantin Orasan. 2002. A new, fully automatic version of Mitkov’s knowledge-poor pronoun resolution method. In Al. Gelbukh, editor, Computational Linguistics and Intelligent Text Processing, pages 169–187. find
  11. MUC-6. 1995. Proceedings of the Sixth Message Understanding Conference (MUC-6). find
  12. MUC-7. 1998. Proceedings of the Seventh Message Understanding Conference (MUC-7). find
  13. Vincent Ng and Claire Cardie. 2002a. Identifying anaphoric and non-anaphoric noun phrases to improve coreference resolution. In Proceedings of COLING, pages 730–736. find
  14. Vincent Ng and Claire Cardie. 2002b. Improving machine learning approaches to coreference resolution. In Proceedings of the ACL, pages 104–111. find our card
  15. Eric W. Noreen. 1989. Computer Intensive Methods for Testing Hypothesis: An Introduction. John Wiley & Sons. find
  16. Chris Paice and Gareth Husk. 1987. Towards the automatic recognition of anaphoric features in English text: the impersonal pronoun ’it’. Computer Speech and Language, 2. find
  17. J. Ross Quinlan. 1993. C4.5: Programs for Machine Learning. San Mateo, CA: Morgan Kaufmann. find our card
  18. 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
  19. Michael Strube and Christoph Müller. 2003. A machine learning approach to pronoun resolution in spoken dialogue. In Proceedings of the ACL, pages 168–175. find our card
  20. Renata Vieira and Massimo Poesio. 2000. An empirically-based system for processing definite descriptions. Computational Linguistics, 26(4):539– 593. find
  21. Marc Vilain, John Burger, John Aberdeen, Dennis Connolly, and Lynette Hirschman. 1995. A model-theoretic coreference scoring scheme. In Proceedings of the Sixth Message Understanding Conference (MUC-6), pages 45–52. find our card
  22. Xiaofeng Yang, Guodong Zhou, Jian Su, and Chew Lim Tan. 2003. Coreference resolution using competitive learning approach. In Proceedings of the ACL, pages 176–183. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/ng-2004-learning-noun-phrase-anaphoricity-improve/