Elamur Research⌕
☰
home / papers / 2004 / Using Word Similarity Lists for Resolving Indirect Anaphora

Using Word Similarity Lists for Resolving Indirect Anaphora

Summary

The paper tackles indirect anaphora in Portuguese. These are coreferent definite descriptions whose head noun differs from the antecedent's head but is semantically related to it, for example through synonymy or hyponymy. Such cases cannot be resolved by head matching and need lexical knowledge.

The authors built that knowledge automatically. They parsed a 1.4-million-word Folha de São Paulo news corpus with PALAVRAS. They then grouped nouns by the syntactic contexts they share, using a weighted Jaccard measure in a method extending Grefenstette's approach. The output was a list of similar words for each noun.

A heuristic accepts a preceding NP as antecedent in three cases:

  • its head appears in the anaphor's list,
  • the anaphor's head appears in the candidate's list,
  • or the two lists overlap.

When several candidates qualify, the winner is chosen by a penalty for the matching route plus the distance in words.

Evaluated on 95 manually annotated indirect anaphors, the noun-only lists gave 52.7% precision and 33.3% recall. Lists that also contained proper names performed worse. The work shows that cheap, corpus-tuned resources can support anaphora resolution for a language with fewer resources than English, with results comparable to WordNet-based work on English.

Key points

  • Similarity lists: two sets were built, each list holding the 15 most similar words. The noun-only set has 8019 lists; the nouns-plus-proper-names set has 12275.
  • Evaluation data: 95 indirect anaphors were drawn from 680 hand-classified definite descriptions, annotated with the MMAX tool. Of these, 14 are synonymy, 43 hyponymy and 38 involve proper names.
  • Ranking scheme: the matching routes are penalised 0, 40 and 200 respectively, and the penalty is added to the word distance. The heuristic is implemented as an XSL stylesheet in the Anaphora Resolution Tool (ART).
  • Results by list set and relation:
    • With noun-only lists, 64% of synonymy cases and 28% of hyponymy cases were resolved.
    • With proper names added, precision was 36.8% and recall 22.1% (21 correct). No proper-name case was resolved.
  • Comparison with English work:
    • WordNet resolved 39% of synonymy, hypernymy and meronymy bridging cases (Poesio et al., 1997).
    • Schulte im Walde's corpus-based method on the British National Corpus resolved 22% of the same cases.

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

6 papers of our library cite it (of the 140673 whose bibliographies we read); 2 of them have a page here, 4 not yet.

Cite

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

@inproceedings{gasperin2004using,
  title = {{Using Word Similarity Lists for Resolving Indirect Anaphora}},
  author = {Gasperin, Caroline and Vieira, Renata},
  booktitle = {Proceedings of the Conference on Reference Resolution and Its Applications},
  year = {2004},
  url = {https://aclanthology.org/W04-0706/}
}

References

18 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. D. Bean and E. Riloff. 1999. Corpus-based identification of non-anaphoric noun phrases. In Proceedings of the 37th Annual Meeting of the Association for Computational Linguistics (ACL-99). find our card
  2. Eckhard Bick. 2000. The Parsing System PALAVRAS: Automatic Grammatical Analysis of Portuguese in a Constraint Grammar Framework. Ph.D. thesis, Århus University, Århus. find
  3. Razvan Bunescu. 2003. Associative anaphora resolution: A web-based approach. In Proceedings of EACL 2003 - orkshop on The Computational Treatment of Anaphora, Budapest. find
  4. Caroline Gasperin, Pablo Gamallo, Alexandre Agustini, Gabriel Lopes, and Vera Lima. 2001. Using syntactic contexts for measuring word similarity. In Proceedings of the Workshop on Semantic Knowledge Acquisition and Categorisation, Helsink, Finland. find
  5. Caroline Gasperin, Renata Vieira, Rodrigo Goulart, and Paulo Quaresma. 2003. Extracting xml syntactic chunks from portuguese corpora. In Traitement automatique des langues minoritaires - TALN 2003, Btaz-sur-mer, France. find
  6. Caroline Varaschin Gasperin. 2001. Extração automática de relações semânticas a partir de relações sintáticas. Master’s thesis, PUCRS, Porto Alegre. find
  7. Gregory Grefenstette. 1994. Explorations in Automatic Thesaurus Discovery. Kluwer Academic Publishers, USA. find
  8. John A. Hawkins. 1978. Definiteness and Indefiniteness. Humanities Press, Atlantic Highland, NJ. find
  9. Christoph Müller and Michael Strube. 2001. MMAX: A tool for the annotation of multi-modal corpora. In Proceedings of the IJCAI 2001, pages 45–50, Seattle. find
  10. Massimo Poesio and Renata Vieira. 1998. A corpus-based investigation of definite description use. Computational Linguistics, 24(2):183–216. find
  11. Massimo Poesio, Renata Vieira, and Simone Teufel. 1997. Resolving bridging descriptions in unrestricted texts. In Proceedings of the Practical, Robust, Anaphora Resolution for Unrestricted Texts, Workshop on Operational Factors, Madrid. find
  12. Massimo Poesio, Tomonori Ishikawa, Sabine Schulte Im Walde, and Renata Vieira. 2002. Acquiring lexical knowledge for anaphora resolution. In Proceedings of LREC 2002, Las Palmas De Gran Canaria. find our card
  13. Susanne Salmon-Alt and Renata Vieira. 2002. Nominal expressions in multilingual corpora: Definites and demonstratives. In Proceedings of the LREC 2002, Las Palmas de Gran Canaria. find
  14. Sabine Schulte im Walde. 1997. Resolving Bridging Descriptions in High-Dimensional Space. Master’s thesis, Institut für Maschinelle Sprachverarbeitung, University of Stuttgart, and Center for Cognitive Science, University of Edinburgh. find
  15. K. van Deemter and R. Kibble. 2000. On coreferring: Coreference in muc and related annotation schemes. Computational Linguistics, 26(4). find our card
  16. Renata Vieira, Susanne Salmon-Alt, and Emmanuel Schang. 2002. Multilingual corpora annotation for processing definite descriptions. In Proceedings of the PorTAL 2002, Faro. find
  17. Renata Vieira, Caroline Gasperin, and Rodrigo Goulart. 2003. From manual to automatic annotation of coreference. In Proceedings of the International Symposium on Reference Resolution and Its Applications to Question Answering and Summarization, Venice. find
  18. Vieira et al. 2000. Extração de sintagmas nominais para o processamento de co-referência. In Anais do V Encontro para o processamento computacional da Língua Portuguesa escrita e falada - PROPOR, Atibaia. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/gasperin-2004-word-similarity-lists-resolving-indirect/