Elamur Research⌕
☰
home / papers / 2013 / Modeling Semantic Relations Expressed by Prepositions

Modeling Semantic Relations Expressed by Prepositions

Paper · 2013 · Vivek Srikumar and Dan Roth

Summary

Prepositions are highly ambiguous: one preposition can signal many meanings, and many prepositions can signal the same meaning. Existing sense inventories, such as the Preposition Project used in SemEval 2007, define labels separately for each preposition. This produces many labels and prevents generalization across words. Srikumar and Roth instead define 32 abstract relations by merging related senses across 34 prepositions. Because each sense maps to exactly one relation, the SemEval 2007 sense-annotated data can be converted into a relation-labeled corpus.

Each relation is treated as a predicate whose arguments are the preposition's governor and object, together with the arguments' semantic types (e.g., WordNet hypernyms). The data label only the relation, so the arguments and types are hidden. The authors compare two approaches:

  • Model 1 pools features over all candidate arguments and types.
  • Model 2 models arguments and types as latent variables and learns them with an extended form of latent structural SVM training.

Model 2 reaches 89.43% relation accuracy. Joint inference with a sense predictor raises this to 90.23% and also improves sense prediction. The resulting labels are more transferable, which is useful for question answering and entailment.

Key points

  • Relation inventory: 32 relation labels (including an "Other" overflow class) cover 34 single-word transitive prepositions. They replace the 332 per-preposition sense labels of the Preposition Project/SemEval 2007.
  • Argument candidates: Governor and object candidates come from a dependency parser plus heuristics, such as the previous noun, adjective or verb for the governor and the next noun for the object. Using all candidate generators performs best.
  • Feature sharing: Sharing features across prepositions raises accuracy from 87.17% to 88.01% (p < 0.01).
  • Model size: Model 2 is much sparser than Model 1. In the typed setting it has about 1.0M weights versus 2.57M; in the typed+gen setting it has 2.21M non-zero weights versus 5.41M.
  • Interpretability: Predicted argument types help explain the chosen label. For example, "experience" + "disease" points to Cause, while "change" + "disease" points to StartState.

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 Illinois Urbana-ChampaignUrbana · United States184 works · 69,372 citations

 Fields, topics and concepts

📚 Shelves 4The 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 works

Cite

Cite as Vivek Srikumar and Dan Roth. 2013. Modeling Semantic Relations Expressed by Prepositions. Transactions of the Association for Computational Linguistics.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.1162/tacl_a_00223

@article{srikumar2013modeling,
  title = {{Modeling Semantic Relations Expressed by Prepositions}},
  author = {Vivek Srikumar and Dan Roth},
  journal = {Transactions of the Association for Computational Linguistics},
  year = {2013},
  doi = {10.1162/tacl\_a\_00223},
  url = {https://doi.org/10.1162/tacl_a_00223}
}

References

30 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. E. Agirre, T. Baldwin, and D. Martinez. 2008. Improving parsing and PP attachment performance with sense information. In Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL), pages 317–325, Columbus, USA. find
  2. T. Baldwin, V. Kordoni, and A. Villavicencio. 2009. Prepositions in applications: A survey and introduction to the special issue. Computational Linguistics, 35(2):119–149. find
  3. M. Chang, D. Goldwasser, D. Roth, and V. Srikumar. 2010a. Discriminative learning over constrained latent representations. In Proceedings of the Annual Meeting of the North American Association of Computational Linguistics (NAACL), pages 429–437, Los Angeles, USA. find
  4. M. Chang, V. Srikumar, D. Goldwasser, and D. Roth. 2010b. Structured output learning with indirect supervision. In Proceedings of the International Conference on Machine Learning (ICML), pages 199–206, Haifa, Israel. find
  5. J. Clarke, V. Srikumar, M. Sammons, and D. Roth. 2012. An NLP Curator (or: How I Learned to Stop Worrying and Love NLP Pipelines). In Proceedings of the International Conference on Language Resources and Evaluation (LREC), pages 3276–3283, Istanbul, Turkey. find
  6. J. Cohen. 1960. A coefficient of agreement for nominal scales. Educational and Psychological Measurement, 20:37–46. find
  7. D. Dahlmeier, H. T. Ng, and T. Schultz. 2009. Joint learning of preposition senses and semantic roles of prepositional phrases. In Proceedings of the Conference on Empirical Methods for Natural Language Processing (EMNLP), pages 450–458, Singapore. find
  8. D. Das, N. Schneider, D. Chen, and N. Smith. 2010. Probabilistic frame-semantic parsing. In Proceedings of Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics, pages 948– 956, Los Angeles, USA.
  9. C. Fillmore, C. Johnson, and M. Petruck. 2003. Background to FrameNet. International Journal of Lexicography, 16(3):235–250. find
  10. D. Gildea and D. Jurafsky. 2002. Automatic labeling of semantic roles. Computational Linguistics, 28(3):245–288.
  11. Y. Goldberg and M. Elhadad. 2010. An efficient algorithm for easy-first non-directional dependency parsing. In Proceedings of Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics, pages 742–750, Los Angeles, USA. find
  12. D. Hovy, S. Tratz, and E. Hovy. 2010. What’s in a preposition? dimensions of sense disambiguation for an interesting word class. In Coling 2010: Posters, pages 454–462, Beijing, China. find
  13. D. Lin. 1998. Automatic retrieval and clustering of similar words. In Proceedings of the Annual Meeting of the Association for Computational Linguistics (ACL), pages 768–774, Montreal, Canada.
  14. K. Litkowski and O. Hargraves. 2005. The Preposition Project. In ACL-SIGSEM Workshop on the Linguistic Dimensions of Prepositions and their Use in Computational Linguistics Formalisms and Applications, pages 171–179, Colchester, UK. find
  15. K. Litkowski and O. Hargraves. 2007. SemEval-2007 Task 06: Word-Sense Disambiguation of Prepositions. In Proceedings of the Fourth International Workshop on Semantic Evaluations (SemEval-2007), pages 24– 29, Prague, Czech Republic. find
  16. K. Litkowski. 2012. Proposed Next Steps for The Preposition Project. Technical Report 12-01, CL Research. find
  17. A. Meyers, R. Reeves, C. Macleod, R. Szekely, V. Zielinska, B. Young, and R. Grishman. 2004. The Nom-Bank project: An interim report. In HLT-NAACL 2004 Workshop: Frontiers in Corpus Annotation, pages 24– 31, Boston, USA. find
  18. T. O’Hara and J. Wiebe. 2009. Exploiting semantic role resources for preposition disambiguation. Computational Linguistics, 35(2):151–184. find
  19. M. Palmer, D. Gildea, and P. Kingsbury. 2005. The Proposition Bank: An Annotated Corpus of Semantic Roles. Computational Linguistics, 31(1):71–106.
  20. M. Palmer, D. Gildea, and N. Xue. 2010. Semantic Role Labeling, volume 3. find
  21. Morgan & Claypool Publishers. V. Punyakanok, D. Roth, and W. Yih. 2008. The importance of syntactic parsing and inference in semantic role labeling. Computational Linguistics, 34(2).
  22. D. Roth and W. Yih. 2004. A linear programming formulation for global inference in natural language tasks.
  23. In Proceedings of the Annual Conference on Computational Natural Language Learning (CoNLL), pages 1–8, Boston, USA. D. Roth and W. Yih. 2007. Global inference for entity and relation identification via a linear programming formulation. Introduction to Statistical Relational Learning. find
  24. V. Srikumar and D. Roth. 2011. A joint model for extended semantic role labeling. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP), Edinburgh, Scotland. find
  25. S. Tratz and D. Hovy. 2009. Disambiguation of preposition sense using linguistically motivated features. In Proceedings of Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, Companion Volume: Student Research Worksh… find
  26. S. Tratz. 2011. Semantically-enriched Parsing for Natural Language Understanding. Ph.D. thesis, University of Southern California. find
  27. I. Tsochantaridis, T. Hofmann, T. Joachims, and Y. Altun. 2004. Support vector machine learning for interdependent and structured output spaces. In Proceedings of the International Conference on Machine Learning (ICML), pages 104–111, Banff, Canada. find
  28. P. Ye and T. Baldwin. 2006. Semantic role labeling of prepositional phrases. ACM Transactions on Asian Language Information Processing (TALIP), 5(3):228– 244. find
  29. P. Ye and T. Baldwin. 2007. MELB-YB: Preposition sense disambiguation using rich semantic features. In Proceedings of the Fourth International Workshop on Semantic Evaluations (SemEval-2007), pages 241– 244, Prague, Czech Republic. find
  30. C. Yu and T. Joachims. 2009. Learning structural SVMs with latent variables. In Proceedings of the International Conference on Machine Learning (ICML), pages 1–8, Montreal, Canada. B. Zapirain, E. Agirre, L. Màrquez, and M. Surdeanu. 2012. Selectional preferences for semantic role classification. C…

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/srikumar-2013-modeling-semantic-relations-expressed-prepositions/