Elamur Research⌕
☰
home / papers / 2002 / Acquiring Lexical Knowledge for Anaphora Resolution

Acquiring Lexical Knowledge for Anaphora Resolution

Summary

Resolving bridging references, such as "the kitchen" after "the house," needs lexical and commonsense knowledge that hand-built resources like WordNet often lack. The relations may be missing, depend on context, or sit deep in the hierarchy. Building on an earlier definite-description resolver that combined WordNet 1.6 with heuristics, the authors test whether this knowledge can be learned from corpora instead.

They first tried unsupervised vector-based word association on the BNC. This slightly beat WordNet on overall bridging resolution and matched it for synonymy, but it was weak on hyponymy and meronymy.

For part-of relations, they then parsed the BNC with Abney's CASS parser and collected word pairs found in Hearst-style constructions (of-phrases, possessives, compounds). Candidate anchors were ranked by mutual information. This construction-based method sharply improved resolution of meronymic bridging descriptions.

The work shows that corpus-acquired lexical knowledge can help anaphora resolution. The best results come from using a different acquisition method for each relation type.

Key points

  • The evaluation corpus has 204 bridging descriptions, split into six knowledge-based classes: lexical relations (synonymy, hypernymy, meronymy), proper names, compound nouns, events, discourse topic, and inference. The earlier WordNet-plus-heuristics system reached R=29% and P=38% on these.
  • The best vector-based configuration used a window of 10, lemmatized but untagged text, and cosine similarity. It resolved 22.7% of bridging cases (46/204), compared with 16.7% (34/204) for WordNet alone. Synonymy accuracy was the same for both methods (36%), but hyponymy fell to 14.3% (WordNet: 57.1%) and meronymy to 16.7% (WordNet: 25%).
  • A simple "priming hypothesis" was not supported: the strongest associate within the previous five sentences was the correct anchor in only 22.7% of cases.
  • The construction patterns were "the NP of NP," "NP of NP," possessives "NP's NP," and nominal compounds "NP N." Proper-name antecedents were first mapped to named-entity types using Mikheev et al.'s system.
  • Combining the three non-compound constructions gave 66.7% recall and 72.7% precision on meronymic bridging, compared with 25% for WordNet and 16.7% for the vector method.

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

22 papers of our library cite it (of the 140673 whose bibliographies we read); 9 of them have a page here, 13 not yet.

Cite

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

@inproceedings{poesio2002acquiring,
  title = {{Acquiring Lexical Knowledge for Anaphora Resolution}},
  author = {Poesio, Massimo and Ishikawa, Tomonori and Walde, Sabine Schulte im and Vieira, Renata},
  booktitle = {Proceedings of the Third International Conference on Language Resources and Evaluation (LREC'02)},
  year = {2002},
  doi = {10.63317/3vp7obnb6oct},
  url = {https://doi.org/10.63317/3vp7obnb6oct}
}

References

36 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Abney, S., and M. Light, 1999. Hiding a Semantic Class Hierarchy in a Markov Model . In Proc. of ACL Workshop on Unsupervised Learning in NLP. University of Maryland. find
  2. Our original experiments were part of (Ishikawa, 1998). find
  3. Berland, M. and Charniak, E. 1999. Finding Parts in Very Large Corpora. In Proc. of the 37th ACL. University of Maryland. find
  4. Brent, 1993. . From grammar to lexicon: unsupervised learning of lexical syntax. Computational Linguistics, 19(3):243–262. find
  5. Brown, P., V. J. D. Della Pietra, P. V. DeSouza, J. C. Lai, and R. L. Mercer, 1992. Class-based n-grams models of natural language. Computational Linguistics, 18(4):467–479. find
  6. Caraballo, S. A., 1999. Automatic acquisition of a hypernym-labeled noun hierarchy from text. In Proc. of the ACL. find
  7. Carter, D. M., 1987. Interpreting Anaphors in Natural Language Texts. Chichester, UK: Ellis Horwood. find
  8. Clark, H. H., 1977. Bridging. In P. N. Johnson-Laird and P.C. Wason (eds.), Thinking: Readings in Cognitive Science. London and New York: Cambridge University Press. find
  9. Fellbaum, Christiane, editor. 1998. WordNet: An electronic lexical database. The MIT Press. find
  10. Grefenstette, G., 1993. SEXTANT: extracting semantics from raw text. Heuristics. find
  11. Harabagiu, S., 1997. WordNet-Based Inference of Textual Context, Cohesion and Coherence. PhD Dissertation, University of Southern California. find
  12. Hawkins, J. A., 1978. Definiteness and Indefiniteness. London: Croom Helm. find
  13. Hearst, M. A., 1998. Automated discovery of wordnet relations. In C. Fellbaum (ed.), WordNet: An Electronic Lexical Database. MIT Press. find
  14. Ishikawa, T., 1998. Acquisition of Associative Information and Resolution of Bridging Descriptions. MSc thesis, University of Edinburgh, Department of Linguistics. find
  15. Lappin, S. and H. J. Leass, 1994. An algorithm for pronominal anaphora resolution. Computational Linguistics, 20(4):535–562. find our card
  16. Lund, K., C. Burgess, and R. A. Atchley, 1995. Semantic and associative priming in high-dimensional semantic space. In Proc. of the 17th Annual Conference of the Cognitive Science Society. find
  17. Manning, C. D., 1993. Automatic acquisition of a large subcategorization dictionary from corpora. In Proc. of the 31st ACL. p. 235–242. find
  18. Mikheev, A., M. Moens, and C. Grover, 1999. Named Entity recognition without gazetteers. In Proc. of EACL. Bergen, Norway: EACL. find
  19. Mitkov, R., 1998. Robust pronoun resolution with limited knowledge. In Proc. of the 18th COLING. Montreal. find our card
  20. Morton, T., 2000. Coreference for NLP applications. In Proc. of the 38th ACL. Hong Kong. find
  21. Moss, H. E., R. K. Ostrin, L. K. Tyler, and W. D. Marslen-Wilson, 1995. Accessing different types of lexical semantic information: Evidence from priming. Journal of Experimental Psychology: Learning, Memory and Cognition:863–883. find
  22. Poesio, M., H. Cheng, B. Di Eugenio, J. M. Hitzeman, and R. Stevenson, 2002. A corpus-based evaluation of centering theory. Submitted. find
  23. Poesio, M., H. Cheng, R. Henschel, J. M. Hitzeman, R. Kibble, and R. Stevenson, 2000. Specifying the parameters of Centering Theory: a corpus-based evaluation using text from application-oriented domains. In Proc. of the 38th ACL. Hong Kong. find
  24. Poesio, M., 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. Stanford, CA: AAAI. find
  25. Poesio, M. and R. Vieira, 1998. A corpus-based investigation of definite description use. Computational Linguistics, 24(2):183–216. Also available as Research Paper CCS-RP-71, Centre for Cognitive Science, University of Edinburgh. find
  26. Poesio, M., R. Vieira, and S. Teufel, 1997. Resolving bridging references in unrestricted text. In R. Mitkov (ed.), Proc. of the ACL Workshop on Operational Factors in Robust Anaphora Resolution. Madrid. Also available as HCRC Research Paper HCRC/RP-87, University of Edinburgh. find our card
  27. Resnik, P., 1993. Selection and information: a class-based approach to lexical relationships. PhD Dissertation, University of Pennsylvania. find
  28. Schulte im Walde, S., 1997. Resolving bridging descriptions in high-dimensional space. Studienarbeit, Universities of Stuttgart and Edinburgh. find
  29. Schütze, H., 1997. Ambiguity Resolution in Language Learning. Stanford: CSLI. find
  30. Sidner, C. L., 1979. Towards a computational theory of definite anaphora comprehension in English discourse. Ph.D. thesis, MIT. find
  31. Strand, K., 1996. A taxonomy of linking relations. Manuscript. A preliminary version presented at the Workshop on Indirect Anaphora, Lancaster University, 1996. find
  32. Tetreault, J. R., 2001. A corpus-based evaluation of centering and pronoun resolution. Computational Linguistics, 27(4):507–520. find our card
  33. Vieira, R., 1998. Definite Description Resolution in Unrestricted Texts. Ph.D. thesis, University of Edinburgh, Centre for Cognitive Science. find
  34. Vieira, R. and M. Poesio, 2000a. An empirically-based system for processing definite descriptions. Computational Linguistics, 26(4). find
  35. Vieira, R. and M. Poesio, 2000b. Processing definite descriptions in corpora. In S. Botley and T. McEnery (eds.), Corpus-based and Computational Approaches to Discourse Anaphora, chapter 10. Amsterdam / New York: John Benjamins, pages 189–212. find
  36. Vieira, R. and S. Teufel, 1997. Towards resolution of bridging descriptions. In Proc. of the 35th Joint Meeting of the Association for Computational Linguistics. Madrid. find our card

↑ back to top

This page in Markdown →