Elamur Research⌕
☰
home / papers / 2004 / Learning to Resolve Bridging References

Learning to Resolve Bridging References

Summary

The paper addresses mereological bridging references in English: definite noun phrases such as "the top" that refer to a part of an entity mentioned earlier. Earlier work had tested two kinds of evidence separately. Salience cues, such as choosing the first-mentioned entity of the previous sentence, and lexical relatedness, measured with WordNet, vector lexicons or construction patterns, each stayed below roughly 40% accuracy. The authors combine both kinds of evidence by treating anchor identification as a binary classification task over candidate antecedents. The features are first mention, utterance distance, and a lexical distance score computed either from Google hit counts or from WordNet, where hypernym paths compensate for missing meronymy links. They train their own MLP, plus Weka's neural network and Naive Bayes, on the GNOME corpus. The combined features clearly beat all baselines and previously published results. WordNet distance performs about as well as Google, and slightly better for definite descriptions. This shows that pairing salience with lexical knowledge is effective, and that WordNet can still be useful for part-of bridging.

Key points

  • Data: 58 mereological bridging descriptions from the GNOME corpus. A second experiment adds all bridging references, including indefinite NPs, for 153 positive instances.
  • On balanced data, the authors' MLP reaches 79.3% correct with WN distance and 76.7% with Google distance. The F-values for positive instances are 79.6% (WN) and 77.6% (Google).
  • On the extended set of all mereological references, the F-measure is 76.3% with WN and 75.8% with Google, so the two resources perform essentially the same.
  • In a harder test, classifiers pick anchors from all 346 candidate antecedents in context. The best result, F=.5, combines Google and WN distance with a larger, unbalanced training set. This is about 66% above the strongest baseline (first mention only).
  • Planned work: enlarge the dataset, integrate the method into the GUITAR anaphora resolution system, and detect bridging descriptions automatically.

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

46 papers of our library cite it (of the 140673 whose bibliographies we read); 27 of them have a page here, 19 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.3115/1218955.1218974

@inproceedings{poesio2004learning,
  title = {{Learning to Resolve Bridging References}},
  author = {Poesio, Massimo and Mehta, Rahul and Maroudas, Axel and Hitzeman, Janet},
  booktitle = {Proceedings of the 42nd Annual Meeting of the Association for Computational Linguistics (ACL-04)},
  year = {2004},
  doi = {10.3115/1218955.1218974},
  url = {https://doi.org/10.3115/1218955.1218974}
}

References

35 references, 4 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. N. Asher and A. Lascarides. 1998. Bridging. Journal of Semantics, 15(1):83–13. find
  2. M. Berland and E. Charniak. 1999. Finding parts in very large corpora. In Proc. of the 37th ACL. find
  3. H. H. Clark and C. J. Sengul. 1979. In search of referents for nouns and pronouns. Memory and Cognition, 7(1):35–41. find
  4. H. H. Clark. 1977. Bridging. In P. N. Johnson-Laird and P.C. Wason, editors, Thinking: Readings in Cognitive Science. Cambridge. find
  5. C. Fellbaum, editor. 1998. WordNet: An electronic lexical database. The MIT Press. find
  6. A. Garcia-Almanza. 2003. Using WordNet for mereological anaphora resolution. Master’s thesis, University of Essex. find
  7. Lex Res WN Balance 1:1 1:3 1:1 1.3 1:1 1.3 Perc on bal 70.2% 75.9% 64.4% 79.8% 66.3% 77.9% Google WN + Google Table 6: Results using a classifier trained on balanced data on unbalanced ones. find
  8. M. A. Gernsbacher and D. Hargreaves. 1988. Accessing sentence participants. Journal of Memory and Language, 27:699–717. find
  9. P. C. Gordon, B. J. Grosz, and L. A. Gillion. 1993. Pronouns, names, and the centering of attention in discourse. Cognitive Science, 17:311–348. find
  10. G. Grefenstette. 1993. SEXTANT: extracting semantics from raw text. Heuristics. find
  11. B. J. Grosz and C. L. Sidner. 1986. Attention, intention, and the structure of discourse. Computational Linguistics, 12(3):175–204. find
  12. B. J. Grosz, A. K. Joshi, and S. Weinstein. 1995. Centering. Computational Linguistics, 21(2):202–225. find
  13. S. Harabagiu and D. Moldovan. 1998. Knowledge processing on extended WordNet. In (Fellbaum, 1998), pages 379–405. find
  14. M. A. Hearst. 1998. Automated discovery of Wordnet relations. In (Fellbaum, 1998). find
  15. J. R. Hobbs. 1978. Resolving pronoun references. Lingua, 44:311–338. find our card
  16. K. Humphreys, R. Gaizauskas, S. Azzam, C. Huyck, B. Mitchell, and H. Cunningham Y. Wilks. 1997. Description of the LaSIE-II System as used for MUC-7. In Proc. of the 7th Message Understanding Conference (MUC-7). find
  17. T. Ishikawa. 1998. Acquisition of associative information and resolution of bridging descriptions. Master’s thesis, University of Edinburgh. find
  18. F. Keller and M. Lapata. 2003. Using the Web to obtain frequencies for unseen bigrams. Computational Linguistics, 29(3). find
  19. K. Lund, C. Burgess, and R. A. Atchley. 1995. Semantic and associative priming in high-dimensional semantic space. In Proc. of the 17th Conf. of the Cogn. Science Soc., pages 660–665. find
  20. K. Markert, M. Strube, and U. Hahn. 1996. Inferential realization constraints on functional anaphora in the centering model. In Proc. of 18th Conf. of the Cog. Science Soc., pages 609–614. find
  21. K. Markert, M. Nissim, and N.. Modjeska. 2003. Using the Web for nominal anaphora resolution. find
  22. F on bal .7 .4 .7 .5 .6 .4 Perc on Hard 80.2% 91.7% 63.6% 88.4% 65.3% 92.5% F on Hard .2 0 .1 .3 .2 .5 In Proc. of the EACL Workshop on the Computational Treatment of Anaphora, pages 39–46. find
  23. N. Modjeska, K. Markert, and M. Nissim. 2003. Using the Web in ML for anaphora resolution. In Proc. of EMNLP-03, pages 176–183. find
  24. V. Ng and C. Cardie. 2002. Improving machine learning approaches to coreference resolution. In Proceedings of the 40th Meeting of the ACL. find our card
  25. M. Poesio and R. Vieira. 1998. A corpus-based investigation of definite description use. Computational Linguistics, 24(2):183–216, June. find
  26. M. Poesio, R. Vieira, and S. Teufel. 1997. Resolving bridging references in unrestricted text. In R. Mitkov, editor, Proc. of the ACL Workshop on Robust Anaphora Resolution, pages 1–6, Madrid. find our card
  27. M. Poesio, S. Schulte im Walde, and C. Brew. 1998. Lexical clustering and definite description interpretation. In Proc. of the AAAI Spring Symposium on Learning for Discourse, pages 82–89. find
  28. M. Poesio, T. Ishikawa, S. Schulte im Walde, and R. Vieira. 2002. Acquiring lexical knowledge for anaphora resolution. In Proc. of the 3rd LREC. find our card
  29. M. Poesio and M. Alexandrov-Kabadjov. 2004. A general-purpose, off the shelf anaphoric resolver. In Proc. of the 4th LREC, Lisbon. find
  30. M. Poesio, R. Stevenson, B. Di Eugenio, and J. M. Hitzeman. 2004. Centering: A parametric theory and its instantiations. Comp. Linguistics. 30(3). find
  31. M. Poesio. 2003. Associative descriptions and salience. In Proc. of the EACL Workshop on Computational Treatments of Anaphora. find
  32. E. F. Prince. 1981. Toward a taxonomy of given-new information. In P. Cole, editor, Radical Pragmatics, pages 223–256. Academic Press. find
  33. C. L. Sidner. 1979. Towards a computational theory of definite anaphora comprehension in English discourse. Ph.D. thesis, MIT. find
  34. O. Uryupina. 2003. High-precision identification of discourse-new and unique noun phrases. In Proc. of ACL 2003 Stud. Workshop, pages 80–86. find
  35. R. Vieira and M. Poesio. 2000. An empirically-based system for processing definite descriptions. Computational Linguistics, 26(4), December. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/poesio-2004-learning-resolve-bridging-references/