Elamur Research⌕
☰
home / papers / 1992 / Automatic Acquisition of Hyponyms from Large Text Corpora

Automatic Acquisition of Hyponyms from Large Text Corpora

Paper · 1992 · Marti A. Hearst

Summary

The paper addresses the task of automatically extracting the hyponymy relation, i.e. the "X is a kind of Y" relation, from unrestricted text. The method requires neither pre-encoded knowledge nor a complex syntactic parser. The author proposes a set of lexico-syntactic patterns with the markers "such as", "and other", "or other", "including" and "especially". Such constructions occur frequently in texts of various genres, almost always indicate hyponymy, and are easy to recognize. An iterative procedure for discovering new patterns is also described. First, known pairs of terms are taken and places in the corpus where they occur near each other are found. Then the shared contexts are generalized, and the new patterns are applied to collect even more pairs. Part of the algorithm was implemented only for the "such as" pattern. It was tested on Grolier's encyclopedia and on New York Times texts, and the pairs found were compared against the WordNet noun hierarchy. The approach proved suitable both for extending and for critically checking manually built thesauri. The paper laid the foundation for the method now known as "Hearst patterns". It still influences automatic taxonomy construction.

Key points

  • Six patterns are proposed: NP such as NP…, such NP as NP…, NP … or other NP, NP … and other NP, NP including NP…, NP especially NP…. The England–country pair helped to find the patterns with "and other" and "including", and the tank–vehicle pair a productive pattern with "especially".

  • In the Grolier's American Academic Encyclopedia corpus of 8.6 million words, 7067 sentences with "such as" were found. Under the strict condition that neither NP has modifiers, 152 relations were obtained. After relaxing the condition (allowing two nouns or a participle with a noun), their number rose to 330.

  • In New York Times texts of about 20 million words, 3178 sentences with "such as" were found, but under the strict criterion only 46 relations.

  • Checking the 152 relations from Grolier's against WordNet: of the 226 unique words, 180 are in the hierarchy. Among the 106 pairs in which both terms are present in WordNet, 61 relations are already recorded.

  • Problems identified:

    • metonymy (king–institution);
    • insufficient specificity (steatornis–species);
    • context dependence (aircraft–target);
    • overly general hypernyms (substance, form).

    Ideas for automatically adding pairs to WordNet with word sense disambiguation are only proposed; the author did not test them.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓University of California, BerkeleyBerkeley · United States175 works · 227,440 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksBiochemistry, Genetics and Molecular Biology · 234 worksMolecular Biology · 231 works

💡 Concepts 2Key ideas and methods explained on our concept pages.Natural language processing · 131 worksInformation extraction · 50 works

Cite

Cite as Marti A. Hearst. 1992. Automatic Acquisition of Hyponyms from Large Text Corpora. COLING 1992 Volume 2: The 14th International Conference on Computational Linguistics.

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

@inproceedings{hearst1992automatic,
  title = {{Automatic Acquisition of Hyponyms from Large Text Corpora}},
  author = {Marti A. Hearst},
  booktitle = {COLING 1992 Volume 2: The 14th International Conference on Computational Linguistics},
  year = {1992},
  doi = {10.3115/992133.992154},
  url = {https://doi.org/10.3115/992133.992154}
}

References

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

  1. Ahlswede, T. & M. Evens (1988). Parsing vs. text processing in the analysis of dictionary definitions. Proceedings of the 26th Annual Meeting of the Association for Computational Linguistics, pages 217-224. find
  2. Alshawi, H. (1987). Processing dictionary definitions with phrasal pattern hierarchies. American Journal of Computational Linguistics, 13(3):195 202.
  3. Batali, J. (1991). Automatic Acquisition and Use of Some of the Knowledge in Physics Tezts. PhD thesis, Massachusetts Institute of Technology, Artificial Intelligence Laboratory. find
  4. Brent, M. R. (1991). Automatic acquisition ofsubcat-egorization frames from untagged, free-text corpora. In Proceedings of the 29th Annual Meeting of the Association fo'e Computational Linguistics. find
  5. Calzolari, N. & R. Bindi (1990). Acquisition of lexical information from a large textual italian corpus. In Proceedings of the Thirteenth International Conference on Computational Linguistics, Helsinki. find
  6. Coates-Stephens, S. (1991). Coping with lexical inadequacy - the automatic acquisition of proper nouns from news text. In The Proceedings of the 7th Annual Conference of the UW Centre for the New OED and Tezt Research: Using Corpora, pages 154-169, Oxford. find
  7. Cutting, D., J. Kupiec, J. Pedersen, & P. Sibun (1991). A practical part-of-speech tagger. Submitted to The 3rd Conference on Applied Natural Language Process*ng.
  8. Grolier (1990). Academic American Encyclopedia. Grolier Electronic Publishing, Danbury, Con-neeticut. find
  9. Hearst, M. A. (1991). Noun homograph disambiguation using local context in large text corpora. In The Proceedings of the 7th Annual Conference of the UW Centre for the New OED and Tezt Research: Using Corpora, Oxford. find
  10. Hindle, D. (1990). Noun classification from predicate-argument structures. Proceedings of the 28th Annual Meeting of the Association for Computational Linguistics, pages 268-275.
  11. Jacobs, P. & U. Zernik (1988). Acquiring lexical knowledge from text: A case study. In Proceedings of AAAI88, pages 739-744. find
  12. ACrF~ DE COLING-92, NANTES, 23-28 AOt~T 1992 find
  13. Jensen, K. & J.-L. Binot (1987). Disambiguating prepositional phrase attachments by using online dictionary definitions. American Journal of Computational Linguistics, 13(3):251-260. find
  14. Markowitz, J., T. Ahlswede, & M. Evens (1986). Semantically significant patterns in dictionary definitions. Proceedings of the 24th Annual Meeting of the Assoczation for Computational Linguistics, pages 112-119. find
  15. Miller, G. A., R. Beckwith, C. Fellbaum, D. Gross, & K. J. Miller (1990). Introduction to wordnet: An on-line lexieal database. Journal of Le~xieography, 3(4):235-244. find
  16. Morris, J. & G. Hirst (1991). Lexical cohesion computed by tbesaural relations as an indicator of the structure of text. Computational Lzngmstics, 17(1):21-48. find
  17. Nakamura, J. & M. Nagao (1988). Extraction of semantic inlbrmation t¥om an ordinary english dictionary and its evaluation. In Proceedings of the Twelfth International Conference on Computational Linguistics, pages 459-464, Budapest. find
  18. Smadja, F. A. & K. R. McKeown (1990). Automatically extracting and representing collocations for language generation. Proceedings of the 28th Annual Meeting of the Association for Computational Linguistics, pages 252-259. find
  19. Velardi, P. & M. T. Pazienza (1989). Computer aided interpretation of lexical cooccurrences. Proceedings of the 27th Annual Meeting of the Association for Computational Linguistics, pages 185-192. find
  20. Wilks, Y. A., D. C. Fass, C. ruing Guo, J. E. McDonald, T. Plate, & B. M. Slator (1990). Providing machine tractable dictionary tools. Journal of Machzne Translation, 2. find
  21. PRoc. OF COLING-92, NANTnS. AUO. 23-28, 1992 find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/hearst-1992-automatic-acquisition-hyponyms-large-text/