Elamur Research⌕
☰
home / papers / 2002 / A Probabilistic Method for Analyzing Japanese Anaphora Integrating Zero Pronoun Detection and Resolution

A Probabilistic Method for Analyzing Japanese Anaphora Integrating Zero Pronoun Detection and Resolution

Summary

Japanese often omits obligatory arguments of verbs. These "zero pronouns" do not appear in the text, so a system must find them before it can link them to antecedents. Earlier corpus-based work mostly addressed antecedent selection alone and relied on costly annotated data. Seki, Fujii and Ishikawa propose one probabilistic framework that combines both steps. The first factor estimates whether a missing case of a verb is a zero pronoun, using verb–case co-occurrence counts. The second factor scores candidate antecedents. It is split into a syntactic model (particle, sentence distance, relative-clause constraint) trained on a small annotated set, and a semantic model trained on large unannotated newspaper text. A certainty score lets the system output only confident analyses. In evaluation on 30 hand-annotated Kyotodaigaku Text Corpus articles, the combined model selected the correct antecedent first in 50.7% of cases. Adding the detection factor raised accuracy at every coverage level. The approach shows that cheap unannotated text can offset the scarcity of anaphora annotation.

Key points

  • Pipeline:
    • JUMAN handles segmentation and POS tagging, and the KNP parser handles syntax.
    • Candidate zero pronouns (ga/wo/ni cases) are filtered with the IPAL verb dictionary, which covers case frames for 911 verbs.
    • Nominative ga is always treated as obligatory.
  • Features:
    • Six features are used: verb, case, particle, distance, relative-clause flag, and semantic class.
    • Semantic classes come from the 544 classes of the Bunruigoihyou thesaurus, which covers about 55K nouns.
  • Training data:
    • The semantic model was built from six years of Mainichi Shimbun text, yielding about 12M verb–noun co-occurrences for 6,194 verb types.
    • The evaluation set has 449 zero pronouns and was tested with leave-one-out cross-validation over 30 articles.
  • Learning curves:
    • Accuracy rose steeply with the first ~5 annotated articles.
    • More unannotated text gave only modest, but supervision-free, gains.
  • Results:
    • The best combination (Both2) reached 50.7% top-1 accuracy.
    • A rule-based baseline in the results table reached 40.1% at top-1 and 58.6% at top-3.
    • Lowering coverage via the certainty threshold (t = 0.5) increased accuracy.

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

23 papers of our library cite it (of the 140673 whose bibliographies we read); 14 of them have a page here, 9 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. arXiv cs/0206030

@inproceedings{seki2002probabilistic,
  title = {{A Probabilistic Method for Analyzing Japanese Anaphora Integrating Zero Pronoun Detection and Resolution}},
  author = {Seki, Kazuhiro and Fujii, Atsushi and Ishikawa, Tetsuya},
  booktitle = {COLING 2002: The 19th International Conference on Computational Linguistics},
  year = {2002},
  eprint = {cs/0206030},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/cs/0206030}
}

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. Chinatsu Aone and Scott William Bennett. 1995. Evaluating automated and manual acquisition of anaphora resolution strategies. In Proceedings of 33th Annual Meeting of the Association for Computational Linguistics, pages 122–129. find our card
  2. Niyu Ge, John Hale, and Eugene Charniak. 1998. A statistical approach to anaphora resolution. In Proceedings of the Sixth Workshop on Very Large Corpora, pages 161–170. find our card
  3. Ralph Grishman and Beth Sundheim. 1996. Message Understanding Conference - 6: A brief history. In Proceedings of the 16th International Conference on Computational Linguistics, pages 466–471. find
  4. Barbara J. Grosz, Aravind K. Joshi, and Scott Weinstein. 1995. Centering: A framework for modeling the local coherence of discourse. Computational Linguistics, 21(2):203–226. find our card
  5. Jerry R. Hobbs. 1978. Resolving pronoun references. Lingua, 44:311–338. find our card
  6. Information-technology Promotion Agency, 1987. IPA Lexicon of the Japanese language for computers (Basic Verbs). (in Japanese). find
  7. Megumi Kameyama. 1986. A property-sharing constraint in centering. In Proceedings of the 24th Annual Meeting of the Association for Computational Linguistics, pages 200–206. find
  8. Yeun-Bae Kim and Terumasa Ehara. 1995. Zero-subject resolution method based on probabilistic inference with evaluation function. In Proceedings of the 3rd Natural Language Processing Pacific-Rim Symposium, pages 721–727. find
  9. Sadao Kurohashi and Makoto Nagao. 1998a. Building a Japanese parsed corpus while improving the parsing system. In Proceedings of The 1st International Conference on Language Resources & Evaluation, pages 719–724. find
  10. Sadao Kurohashi and Makoto Nagao, 1998b. Japanese morphological analysis system JUMAN version 3.6 manual. Department of Informatics, Kyoto University. (in Japanese). find
  11. Sadao Kurohashi, 1998. Japanese Dependency/Case Structure Analyzer KNP version 2.0b6. Department of Informatics, Kyoto University. (in Japanese). find
  12. Mainichi Shimbunsha. 1994–1999. Mainichi Shimbun CD-ROM. find
  13. Ruslan Mitkov, Lamia Belguith, and Malgorzata Stys. 1998. Multilingual robust anaphora resolution. In Proceedings of the 3rd Conference on Empirical Methods in Natural Language Processing, pages 7–16. find our card
  14. Hiromi Nakaiwa and Satoshi Shirai. 1996. Anaphora resolution of Japanese zero pronouns with deictic reference. In Proceedings of the 16th International Conference on Computational Linguistics, pages 812–817. find our card
  15. Hiromi Nakaiwa. 2000. An environment for extracting resolution rules of zero pronouns from corpora. In COLING-2000 Workshop on Semantic Annotation and Intelligent Content, pages 44–52. find
  16. National Language Research Institute. 1964. Bunruigoihyou. Shuei publisher. (in Japanese). find
  17. Manabu Okumura and Kouji Tamura. 1996. Zero pronoun resolution in Japanese discourse based on centering theory. In Proceedings of the 16th International Conference on Computational Linguistics, pages 871–876. find our card
  18. Manuel Palomar, Antonio Ferrández, Lidia Moreno, Patricio Martı́nez-Barco, Jesús Peral, Maximiliano Saiz-Noeda, and Rafael Mu noz. 2001. An algorithm for anaphora resolution in Spanish texts. Computational Linguistics, 27(4):545–568. find our card
  19. 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
  20. Marilyn Walker, Masayo Iida, and Sharon Cote. 1994. Japanese discourse and the process of centering. Computational Linguistics, 20(2):193–233. find

↑ back to top