Elamur Research⌕
☰
home / papers / 2003 / An Empirical Model of Multiword Expression Decomposability

An Empirical Model of Multiword Expression Decomposability

Summary

The paper looks for a way to estimate how far a multiword expression's meaning derives from its parts. The goal is to separate "simple decomposable" MWEs (e.g. traffic light) from idiosyncratically decomposable and non-decomposable ones (e.g. spill the beans, kick the bucket). The authors build a latent semantic analysis (LSA) vector space and measure how similar each MWE is to each of its constituent words. Their hypothesis is that high similarity signals simple decomposability. Because LSA ignores syntax and language-specific structure, the method can in principle apply to any construction type and any language.

They test it on English noun-noun compounds and verb-particle constructions, using WordNet 1.7 as the reference. LSA scores agree poorly with WordNet-based similarity measures. However, when pairs are ranked by LSA similarity, the higher-ranked pairs more often show the MWE as a WordNet hyponym of its constituent, and this pattern holds clearly for low-frequency items. The work offers an unsupervised, construction-independent first step toward modelling decomposability, with the longer-term aim of constraining MWE syntax in grammars.

Key points

  • Data: NN compounds were extracted from the 1996 Wall Street Journal (WSJ, about 31m words) and verb-particles from the British National Corpus (BNC, about 90m words). Evaluation used WordNet 1.7, which lists about 47,000 compound nouns and 2,600 multiword verbs.
  • LSA model: The model follows Schütze's approach. The 50,000 most frequent terms are represented by how often they share a sentence with 1,000 frequent content words, and singular-value decomposition then reduces the dimensions.
  • Similarity correlation: Correlation with WordNet-derived similarity measures, including Hirst and St-Onge (HSO), was very low for NN compounds. It was somewhat higher, but still weak, for verb-particles.
  • Hyponymy test: Pairs were ranked by similarity and split into 3 partitions, with a frequency threshold of 100 occurrences separating low- from high-frequency MWEs. Mean hyponymy fell steadily across partitions for low-frequency items, while the high-frequency curves were erratic.
  • Head vs. modifier: The head noun predicts decomposability better than the modifier. In WordNet, 71.4% of NN compounds fall under their head noun, compared with only 13.7% under their modifier.

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

Cited by

90 papers of our library cite it (of the 140673 whose bibliographies we read); 10 of them have a page here, 80 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.3115/1119282.1119294

@inproceedings{baldwin2003empirical,
  title = {{An Empirical Model of Multiword Expression Decomposability}},
  author = {Baldwin, Timothy and Bannard, Colin and Tanaka, Takaaki and Widdows, Dominic},
  booktitle = {Proceedings of the ACL 2003 Workshop on Multiword Expressions: Analysis, Acquisition and Treatment},
  year = {2003},
  doi = {10.3115/1119282.1119294},
  url = {https://doi.org/10.3115/1119282.1119294}
}

References

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

  1. Ricardo Baeza-Yates and Berthier Ribiero-Neto. 1999. Modern Information Retrieval. Addison Wesley / ACM press. find
  2. Timothy Baldwin and Aline Villavicencio. 2002. Extracting the unextractable: A case study on verb-particles. In Proc. of the 6th Conference on Natural Language Learning (CoNLL-2002), Taipei, Taiwan. find
  3. Colin Bannard, Timothy Baldwin, and Alex Lascarides. 2003. A statistical approach to the semantics of verb-particles. In Proc. of the ACL-2003 Workshop on Multiword Expressions: Analysis, Acquisition and Treatment. (this volume). find
  4. Colin Bannard. 2002. Statistical techniques for automatically inferring the semantics of verb-particle constructions. LinGO Working Paper No. 2002-06. find
  5. Alexander Budanitsky and Graeme Hirst. 2001. Semantic distance in WordNet: An experimental, application-oriented evaluation of five measures. In Workshop on Wordnet and Other Lexical Resources, Second meeting of the NAACL, Pittsburgh, USA. find
  6. Lou Burnard. 2000. User Reference Guide for the British National Corpus. Technical report, Oxford University Computing Services. find
  7. Nicoletta Calzolari, Charles Fillmore, Ralph Grishman, Nancy Ide, Alessandro Lenci, Catherine MacLeod, and Antonio Zampolli. 2002. Towards best practice for multiword expressions in computational lexicons. In Proceedings of the Third International Conference on Language Resources and Evaluation (LRE… find
  8. Ann Copestake, Fabre Lambeau, Aline Villavicencio, Francis Bond, Timothy Baldwin, Ivan A. Sag, and Dan Flickinger. 2002. Multiword expressions: Linguistic precision and reusability. In Proc. of the 3rd International Conference on Language Resources and Evaluation (LREC 2002), pages 1941–7, Las Palma… find our card
  9. Scott Deerwester, Susan Dumais, George Furnas, Thomas Landauer, and Richard Harshman. 1990. Indexing by latent semantic analysis. Journal of the American Society for Information Science, 41(6):391–407. 2002. Understanding Morphology. find
  10. Graeme Hirst and David St-Onge. 1998. Lexical chains as representations of context for the detection and correction of malapropism. In Christiane Fellbaum, editor, WordNet: An Electronic Lexical Database, pages 305–32. MIT Press, Cambridge, USA. find
  11. Thomas Landauer and Susan Dumais. 1997. A solution to Plato’s problem: The latent semantic analysis theory of acquisition. Psychological Review, 104(2):211–240. find
  12. Dekang Lin. 1998a. Automatic retrieval and clustering of similar words. In Proceedings of the 36th Annual Meeting of the ACL and 17th International Conference on Computational Linguistics (COLING/ACL-98). find
  13. Dekang Lin. 1998b. Extracting collocations from text corpora. In First Workshop on Computational Terminology. find
  14. Dekang Lin. 1998c. An information-theoretic definition of similarity. In Proceedings of the 15th International Conference on Machine Learning. find
  15. Dekang Lin. 1999. Automatic identification of non-compositional phrases. In Proc. of the 37th Annual Meeting of the ACL, pages 317–24, College Park, USA. find our card
  16. Diana McCarthy, Bill Keller, and John Carroll. 2003. Detecting a continuum of compositionality in phrasal verbs. In Proc. of the ACL-2003 Workshop on Multiword Expressions: Analysis, Acquisition and Treatment. (this volume). find
  17. I. Dan Melamed. 1997. Automatic discovery of non-compositional compounds in parallel data. In Proc. of the 2nd Conference on Empirical Methods in Natural Language Processing (EMNLP-97), Providence, USA. find
  18. George A. Miller, Richard Beckwith, Christiane Fellbaum, Derek Gross, and Katherine J. Miller. 1990. Introduction to WordNet: an on-line lexical database. International Journal of Lexicography, 3(4):235–44. find
  19. Grace Ngai and Radu Florian. 2001. Transformation-based learning in the fast lane. In Proc. of the 2nd Annual Meeting of the North American Chapter of Association for Computational Linguistics (NAACL2001), pages 40–7, Pittsburgh, USA. find
  20. Geoffrey Nunberg, Ivan A. Sag, and Tom Wasow. 1994. Idioms. Language, 70:491–538. find
  21. Siddharth Patwardhan, Satanjeev Banerjee, and Ted Pedersen. 2003. Using measures of semantic relatedness for word sense disambiguation. In Proc. of the 4th International Conference on Intelligent Text Processing and Computational Linguistics (CICLing-2003), Mexico City, Mexico. find
  22. Darren Pearce. 2001a. Synonymy in collocation extraction. In Proc. of the NAACL 2001 Workshop on WordNet and Other Lexical Resources: Applications, Extensions and Customizations, Pittsburgh, USA. find
  23. Darren Pearce. 2001b. Using conceptual similarity for collocation extraction. In Proc. of the 4th UK Special Interest Group for Computational Linguistics (CLUK4). find
  24. Philip Resnik. 1995. Using information content to evaluate semantic similarity. In Proceedings of the 14th International Joint Conference on Artificial Intelligence. find
  25. Susanne Riehemann. 2001. A Constructional Approach to Idioms and Word Formation. Ph.D. thesis, Stanford. find
  26. Ivan A. Sag, Timothy Baldwin, Francis Bond, Ann Copestake, and Dan Flickinger. 2002. Multiword expressions: A pain in the neck for NLP. In Proc. of the 3rd International Conference on Intelligent Text Processing and Computational Linguistics (CICLing-2002), pages 1–15, Mexico City, Mexico. find our card
  27. Patrick Schone and Dan Jurafsky. 2001. Is knowledge-free induction of multiword unit dictionary headwords a solved problem? In Proc. of the 6th Conference on Empirical Methods in Natural Language Processing (EMNLP 2001), pages 100–108. find
  28. Hinrich Schütze. 1998. Automatic word sense discrimination. Computational Linguistics, 24(1):97–124. find our card
  29. Dominic Widdows, Beate Dorow, and Chiu-Ki Chan. 2002. Using parallel corpora to enrich multilingual lexical resources. In Third International Conference on Language Resources and Evaluation, pages 240–245, Las Palmas, Spain, May. find
  30. Dominic Widdows. 2003. Unsupervised methods for developing taxonomies by combining syntactic and statistical information. In Proc. of the 3rd International Conference on Human Language Technology Research and 4th Annual Meeting of the NAACL (HLT-NAACL 2003). (to appear). find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/baldwin-2003-empirical-model-multiword-expression-decomposability/