Event coreference for information extraction

Summary

Information extraction (IE) systems need to recognise when several clauses or noun phrases describe one and the same event. Otherwise they produce duplicate event records and miss role fillers scattered across sentences. The authors treat event coreference as an extension of noun-phrase coreference inside LaSIE, the system they built for the MUC-6 tasks. LaSIE stores each text's discourse entities in an inheritance-based semantic graph written in the XI knowledge representation language, so domain knowledge and the evolving discourse model share one structure. Each newly added instance is compared with earlier ones under three checks. Their types must be connected by a valid ontology path; for events, both must be sub-events of the template's key event, as hire and retire are under succession. Single-valued attributes such as time must agree, and type-specific "distinct" constraints must not apply. No event-annotated corpus existed, so the authors evaluated indirectly by toggling the mechanism on the MUC-6 management succession task. Overall gains were modest, but succession events improved more clearly. This suggests event coreference is worthwhile in IE.

Key points

Objects and events are handled by one coreference procedure, with different constraints per type. Semantic similarity is the inverse of the path length (in nodes) between two types; for example, company vs. organisation scores 0.5.
Event coreference depends on object coreference: two events can only be merged if the entities filling their roles are themselves coreferent.
On the MUC-6 coreference task, LaSIE scored 51% recall and 71% precision. The top systems reached 63%/63% (best recall) and 59%/72% (best precision).
In the indirect evaluation on the 100-article MUC-6 scenario test corpus, enabling event coreference raised succession-event precision by 5 points and lowered recall by 1 point.
The mechanism prevented 11 spurious succession events out of 196 possible events.

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

Topics: Coreference resolution

Cited by

35 papers of our library cite it (of the 140673 whose bibliographies we read); 18 of them have a page here, 17 not yet.

Cite

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

@inproceedings{humphreys1997event,
  title = {{Event coreference for information extraction}},
  author = {Humphreys, Kevin and Gaizauskas, Robert and Azzam, Saliha},
  booktitle = {Operational Factors in Practical, Robust Anaphora Resolution for Unrestricted Texts},
  year = {1997},
  doi = {10.3115/1598819.1598830},
  url = {https://doi.org/10.3115/1598819.1598830}
}

References

15 references, 0 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

Allen, J. 1987. Natural Language Understanding. Benjamin/Cummings, Menlo Park, CA, 1st edition. find
Appelt, D., J. Hobbs, J. Bear, D. Israel, M. Kameyama, A. Kehler, D. Martin, K. Myers, and M. Tyson. 1995. SRI International FAS-TUS system: MUC-6 Test Results and Analysis. In Proceedings of the Sixth Message Understanding Conference (MUC-6), pages 237-248. Morgan Kaufmann. find
Botley, S. and T. McEnery, editors. 1997. Discourse Anaphora and Anaphor Resolution. University College London Press, London. In press. find
Cowie, J. and W. Lehnert. 1996. Information Extraction. Communications o/the ACM, 39(1):80-91. find
Crowe, J. 1996. Constraint Based Event Recognition for Information Extraction. Ph.D. thesis, University of Edinburgh. find
DARPA. 1995. Proceedings o/ the Sixth Message Understanding Conference (MUC-6). Morgan Kaufmann. find
Gaizauskas, R. 1995. XI: A Knowledge Representation Language Based on Cross-Classification and Inheritance. Technical Report CS-95-24, Department of Computer Science, University of Sheffield. find
Gaizauskas, R. and K. Humphreys. 1997a. Quantative Evaluation of Coreference Algorithms in an Information Extraction System. In S. Botley and T. McEnery, editors, Discourse Anaphora and Anaphor Resolution. University College London Press. In press. find
Gaizauskas, R. and K. Humphreys. 1997b. Using a Semantic Network for Information Extraction. Journal of Natural Language Engineering. In press. find
Gaizauskas, R., T. Wakao, K Humphreys, H. Cunningham, and Y. Wilks. 1995. Description of the LaSIE System as Used for MUC-6. In Proceedings of the Sixth Message Understanding Conference (MUC-6). Morgan Kaufmann. find
Grishman, R. and B. Sundheim. 1996. Message Understanding Conference - 6: A Brief History. In Proceedings of the 16th International Conference on Computational Linguistics, Copenhagen, June. find
Kamp, H. 1981. A Theory of Truth and Semantic Representation. In Formal Methods in The Study o/ Language. J. Groenendijk, Jannsen, T, and Stokhof, M. find
Sowa, J.F. 1984. Conceptual Structures : Information Processing in Mind and Machine. Reading (MA): Addison-Wesley. find
Whittemore, S and M. Macpherson. 1991. Event-building through Role-Filling and Anaphora Resolution. In Proceedings of the 29th meeting of the Asssociation for Computational Linguistics, Berkeley, CA. find
Zarri, G.P. 1992. Semantic Modeling of the Content of (Normative) Natural Language Documents. In Proceedings of the specialized conference on Natural Language Processing and its applications, Avignon. find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/humphreys-1997-event-coreference-information-extraction/

📄 Event coreference for information extraction