Elamur Research⌕
☰
home / papers / 2014 / First steps towards a Predicate Matrix

First steps towards a Predicate Matrix

Paper · 2014 · Maddalen López de Lacalle, Egoitz Laparra and German Rigau

Summary

Predicate lexicons such as FrameNet, VerbNet and PropBank support semantic role labelling and other language-understanding tasks. However, they are expensive to build, cover far fewer items than WordNet, and exist mostly for English. SemLink connects these resources through VerbNet, but its mappings have large gaps.

The authors start from SemLink and measure how much of WordNet, PropBank and FrameNet is actually linked to VerbNet, both for lexical entries and for role structures. They then reorganise the data around WordNet senses. In the resulting Predicate Matrix, each row holds the partial predicate information for one sense and role. To grow coverage, they apply two simple heuristics: passing VerbNet class membership on to WordNet synonyms, and a method based on monosemous words. This preliminary matrix is roughly twice the size of the version derived from SemLink alone.

The work matters because a predicate lexicon anchored in WordNet can expose inconsistencies between resources. It could also be carried over to other languages through local wordnets aligned to the English WordNet.

Key points

  • WordNet–VerbNet gap: WordNet has 25,047 verb senses, while VerbNet classes contain 6,293 predicates. Of the WordNet senses, 18,559 (9,995 lemmas) have no VerbNet link, and 1,077 VerbNet predicates have no WordNet sense in SemLink.
  • PropBank–VerbNet gap: Only 3,558 of 6,181 PropBank predicates map to VerbNet. About half of PropBank arguments (7,915 of 15,871) are tied to VerbNet thematic roles.
  • Synonym propagation: This heuristic proposes up to 5,075 new VerbNet class members (4,616 WordNet senses). For example, abandon, forsake and desolate join leave-51.2.1 via desert. The authors treat these links as suggestions that still need manual checking.
  • Matrix growth: The matrix grew from 36,174 rows (6,556 senses) from SemLink alone, to 69,508 rows (10,984 senses) with synonyms, to 70,391 rows (11,146 senses) after the monosemous step.
  • Row contents: Each row combines VerbNet class and role, FrameNet frame and frame element, PropBank predicate and argument, and MCR data (Adimen-SUMO, WordNet domains, Base Level Concept). It also records the sense's frequency and its number of relations. The matrix is released via the adimen.si.ehu.es site.

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 the Basque CountryLeioa · Spain62 works · 16,683 citations

 Fields, topics and concepts

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 19,503 worksArtificial Intelligence · 18,805 worksArts and Humanities · 17,959 worksLanguage and Linguistics · 17,789 works

Cite

Cite as Maddalen López de Lacalle, Egoitz Laparra and German Rigau. 2014. First steps towards a Predicate Matrix. Proceedings of the Seventh Global Wordnet Conference.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/w14-0150

@inproceedings{lacalle2014first,
  title = {{First steps towards a Predicate Matrix}},
  author = {Maddalen López {de Lacalle}, Egoitz Laparra and German Rigau},
  booktitle = {Proceedings of the Seventh Global Wordnet Conference},
  year = {2014},
  doi = {10.18653/v1/w14-0150},
  url = {https://doi.org/10.18653/v1/w14-0150}
}

References

16 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Javier Álvez, Paqui Lucio, and German Rigau. 2012. Adimen-sumo: Reengineering an ontology for first-order reasoning. International Journal on Semantic Web and Information Systems (IJSWIS), 8(4):80– 116. find
  2. Jordi Atserias, Luı́s Villarejo, German Rigau, Eneko Agirre, John Carroll, Bernardo Magnini, and Piek Vossen. 2004. The meaning multilingual central repository. In Proceedings of GWC, Brno, Czech Republic. find
  3. Collin F. Baker, Charles J. Fillmore, and John B. Lowe. 1997. The berkeley framenet project. In COL-ING/ACL’98, Montreal, Canada.
  4. Christian Bizer, Jens Lehmann, Georgi Kobilarov, Sören Auer, Christian Becker, Richard Cyganiak, and Sebastian Hellmann. 2009. Dbpedia-a crystallization point for the web of data. Web Semantics: Science, Services and Agents on the World Wide Web, 7(3):154–165. find
  5. Kurt Bollacker, Colin Evans, Praveen Paritosh, Tim Sturge, and Jamie Taylor. 2008. Freebase: a collaboratively created graph database for structuring human knowledge. In Proceedings of the 2008 ACM SIGMOD international conference on Management of data, pages 1247–1250. ACM. find
  6. Aljoscha Burchardt, Katrin Erk, and Anette Frank. 2005. A WordNet Detour to FrameNet. In Proceedings of the GLDV 2005 GermaNet II Workshop, pages 408–421, Bonn, Germany. find
  7. Diego De Cao, Danilo Croce, Marco Pennacchiotti, and Roberto Basili. 2008. Combining word sense and usage for modeling frame semantics. In Proceedings of The Symposium on Semantics in Systems for Text Processing (STEP 2008), Venice, Italy.
  8. Katrin Erk and Sebastian Pado. 2004. A powerful and versatile xml format for representing role-semantic annotation. In Proceedings of LREC-2004, Lisbon. find
  9. Christiane Fellbaum, editor. 1998. WordNet. An Electronic Lexical Database. The MIT Press. find
  10. Charles J. Fillmore. 1976. Frame semantics and the nature of language. In Annals of the New York Academy of Sciences: Conference on the Origin and Development of Language and Speech, volume 280, pages 20–32, New York. find
  11. Ana-Maria Giuglea and Alessandro Moschitti. 2006. Semantic role labeling via framenet, verbnet and propbank. In Proceedings of COLING-ACL 2006, pages 929–936, Morristown, NJ, USA. ACL. find
  12. Aitor Gonzalez-Agirre, Egoitz Laparra, and German Rigau. 2012a. Multilingual central repository version 3.0. In LREC, pages 2525–2529. find
  13. Aitor González-Agirre, German Rigau, and Mauro Castillo. 2012b. A graph-based method to improve wordnet domains. In CICLING, pages 17–28. Springer. find
  14. Iryna Gurevych, Judith Eckle-Kohler, Silvana Hartmann, Michael Matuschek, Christian M Meyer, and Christian Wirth. 2012. Uby: A large-scale unified lexical-semantic resource based on lmf. In Proceedings of EACL, pages 580–590. find
  15. Rubén Izquierdo, Armando Suárez, and German Rigau. 2007. Exploring the automatic selection of basic level concepts. In Proceedings of RANLP, volume 7. Citeseer. find
  16. Richard Johansson and Pierre Nugues. 2007. Using WordNet to extend FrameNet coverage. In Proceedings of the Workshop on Building Frame-semantic Resources for Scandinavian and Baltic Languages, at NODALIDA, Tartu, Estonia, May 24. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/lacalle-2014-first-steps-predicate-matrix/