Elamur Research⌕
☰
home / papers / 2017 / CEVO: Comprehensive EVent Ontology Enhancing Cognitive Annotation

CEVO: Comprehensive EVent Ontology Enhancing Cognitive Annotation

Paper · 2017 · Saeedeh Shekarpour, Faisal Alshargi, Valerie Shalin et al.

Summary

Relations between named entities get far less attention than the entities themselves. There is also no abstract scheme for grouping relations, so the same event phrased with different verbs, or encoded by different ontologies, is hard to recognise as the same thing. The authors address this by turning Beth Levin's classification of English verbs into an OWL ontology called CEVO. Levin groups verbs that share both meaning and syntactic patterns. In CEVO, each of her classes becomes an event class and each verb becomes an individual that can belong to several classes. The resource is published under the w3id.org/cevo namespace and can be queried with SPARQL. The paper shows three uses:

  • tagging verbs in text with event types, layered over NIF/OLiA annotations;
  • attaching CEVO classes to properties of other ontologies via the Web Annotation Data Model;
  • linking a textual relation to a knowledge-base property through a shared CEVO class.

There is no quantitative evaluation. The contribution is a fine-grained, cognitively motivated upper layer for relations that could support relation extraction and ontology alignment.

Key points

  • CEVO holds more than 230 event classes and over 3,000 English verbs as individuals. All of it derives from Levin's verb hierarchy.
  • Existing event models (LODE, Schema.org, the DBpedia ontology) offer only one generic event concept or a shallow hierarchy. CEVO aims at fine-grained event types such as communication, meet and murder, illustrated on news headlines from Twitter.
  • Motivating example: "visit" and "consult" can both express a meeting event, yet WordNet synonymy would not connect them.
  • Use case 1: in the headlines, "announce" and "says" are anchored as NIF strings tagged Olia:MainVerb, then both are typed cevo:Communication.
  • Use cases 2 and 3: dbp:spouse is annotated with cevo:Amalgamate via WADM. The verb "marry" in a BBC headline gets the same class, which lets it be linked to dbp:spouse through itsrdf:taIdentRef.

Interesting ideas

  • I1. Tag a relation by its verb class rather than by its lemma or synonym set, so verbs that are not lexical synonyms (announce/say, visit/consult) can still land on one event type.
  • I2. Use a shared upper class as a pivot: label both the text mention and the knowledge-base property with the same event class, then join them, instead of learning a direct text-to-property mapping.
  • I3. Let a verb be a member of several event classes and resolve the choice at query time, rather than forcing a single sense into the lexicon (the paper's "cook" belongs to more than one class, including change of state).
  • I4. Build annotation in layers on the same character-offset anchor: first the syntactic category, then the semantic event type, each layer added without rewriting the previous one.
  • I5. Classify each sentence by its head verb's coarse event type (communication, meeting, killing), as the paper does for news headlines.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (frames): Mova could map the predicate of each frame to its VerbNet/Levin class as a normalised relation label, letting later steps treat "tell" and "announce" as the same kind of act.
  • I2 (world): Mova's compiled ontology could pre-annotate Wikidata/FrameNet relations with verb classes, so a frame from text is grounded in a property by simple class lookup, and every link stays explainable.
  • I3 (world): Mova's knowledge base could keep all class memberships of a verb and let the world builder pick one using argument types, recording which alternative it chose and why.
  • I4 (frames): Mova could attach frame and event-class layers to the exact token spans of the UD tree, so any answer can be traced back from event type to span and part of speech.
  • I5 (answer): the fable and tale readers could index story events by coarse type, making questions like "who spoke to whom" or "who was killed" direct lookups over the world of the text.

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

🎓Leipzig UniversityLeipzig · Germany26 works · 8,622 citations🎓University of DaytonDayton · United States3 works · 169 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

🧭 Topics 1Research subjects and languages the work is about.Knowledge graphs & ontologies · 192 works

Cite

Cite as Saeedeh Shekarpour, Faisal Alshargi, Valerie Shalin, Krishnaprasad Thirunarayan and Amit P. Sheth. 2017. CEVO: Comprehensive EVent Ontology Enhancing Cognitive Annotation. arXiv:1701.05625.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.48550/arxiv.1701.05625 · arXiv 1701.05625

@misc{shekarpour2017cevo,
  title = {{CEVO: Comprehensive EVent Ontology Enhancing Cognitive Annotation}},
  author = {Saeedeh Shekarpour, Faisal Alshargi, Valerie Shalin, Krishnaprasad Thirunarayan and Amit P. Sheth},
  year = {2017},
  doi = {10.48550/arxiv.1701.05625},
  eprint = {1701.05625},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.48550/arxiv.1701.05625}
}

References

15 references, 1 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Andreas Both, Jens Lehmann, Sören Auer, and Martin Brümmer. Integrating NLP using linked data. In The Semantic Web - ISWC 2013 - 1th International Extended Semantic Web Conference, Sydney, NSW, Australia, October 21-25, 2013, Proceedings, Part II, pages 98–113, 2013. find
  2. Christian Chiarcos. Ontologies of linguistic annotation: Survey and perspectives. In Proceedings of the Eight International Conference on Language Resources and Evaluation (LREC’12), Istanbul, Turkey, may 2012. European Language Resources Association (ELRA). find
  3. Christian Chiarcos and Maria Sukhareva. Olia - ontologies of linguistic annotation. Semantic Web, 6(4):379–386, 2015. ↗
  4. LA Doumas and John E Hummel. Approaches to modeling human mental representations: What works, what doesn’t and why. The Cambridge handbook of thinking and reasoning, ed. KJ Holyoak & RG Morrison, pages 73–94, 2005. find
  5. Leonidas AA Doumas and John E Hummel. A symbolic-connectionist model of relation discovery. In Proceedings of the Annual Meeting of the Cognitive Science Society, volume 27, 2005. find
  6. Paolo Ferragina and Ugo Scaiella. Fast and accurate annotation of short texts with wikipedia pages. IEEE Software, 29(1):70–75, 2012. ↗
  7. Jenny Rose Finkel, Trond Grenager, and Christopher Manning. Incorporating non-local information into information extraction systems by gibbs sampling. In Proceedings of the 43rd Annual Meeting on Association for Computational Linguistics, ACL ’05, 2005.
  8. Sebastian Hellmann, Jens Lehmann, Sören Auer, and Martin Brümmer. Integrating NLP using linked data. In The Semantic Web - ISWC 2013 - 12th International Semantic Web Conference, Sydney, NSW, Australia, October 21-25, 2013, Proceedings, Part II, pages 98–113, 2013. find
  9. Beth Levin. English Verb Classes and Alternations:~A Preliminary Investigation. University of Chicago Press, 1993. find
  10. John P. McCrae, Dennis Spohr, and Philipp Cimiano. Linking lexical resources and ontologies on the semantic web with lemon. In The Semantic Web: Research and Applications - 8th Extended Semantic Web Conference, ESWC 2011. find
  11. Pablo N. Mendes, Max Jakob, Andrés García-Silva, and Christian Bizer. Dbpedia spotlight: shedding light on the web of documents. In Proceedings the 7th International Conference on Semantic Systems, I-SEMANTICS 2011, Graz, Austria, September 7-9, 2011, pages 1–8, 2011. find
  12. R. C. Schank. Conceptual dependency: A theory of natural language processing. 3, 1972. find
  13. R.C. Schank. Conceptual Information Processing. North-Holland, Amsterdam, 1975. find
  14. Kuldeep Singh, Andreas Both, Dennis Diefenbach, and Saeedeh Shekarpour. Towards a message-driven vocabulary for promoting the interoperability of question answering systems. In Tenth IEEE International Conference on Semantic Computing, ICSC 2016, Laguna Hills, CA, USA, February 4-6, 2016, pages 386–389, 2016. find
  15. Christina Unger, John P. McCrae, Sebastian Walter, Sara Winter, and Philipp Cimiano. A lemon lexicon for dbpedia. In Proceedings of the NLP & DBpedia workshop co-located with the 12th International Semantic Web Conference (ISWC 2013), Sydney, Australia, October 22, 2013. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/shekarpour-2017-cevo-comprehensive-event-ontology-enhancing/