Elamur Research⌕
☰
home / papers / 2019 / Cross-document coreference: An approach to capturing coreference without context

Cross-document coreference: An approach to capturing coreference without context

Summary

The paper addresses a basic difficulty in building patient timelines from clinical text. Deciding whether two mentions corefer usually relies on a shared discourse context, but linking mentions across separate notes means reasoning over narratives that do not form one coherent text. The authors extended the THYME colon cancer corpus by merging intra-document annotations from two earlier projects, THYME 1 (temporal relations) and Clinical Coreference. They first ran a manual corrections pass, because the merged annotations conflicted with each other. They then double-annotated and adjudicated cross-document coreference and bridging links within each patient's set of notes. A key design choice was a rule-driven CONTAINS-SUBEVENT relation, restricted to specific event categories. It resolves granularity clashes, for example a note that treats "resection" and "surgery" as the same event while another note distinguishes them. The authors found that relying on explicit rules rather than annotator intuition gave the most consistent links. Agreement was very high for identity relations but low for bridging types. They argue that rule-based schemas suit context-poor tasks and are easier to automate.

Key points

  • The gold corpus covers 198 patient sets of three de-identified notes each (594 notes in total). Every set includes an initial clinical note, followed by a pathology note and then a subsequent clinical note, spanning weeks or months.
  • Four cross-document relation types are annotated: IDENTICAL, SET-SUBSET, WHOLE-PART (entities only) and CONTAINS-SUBEVENT (events only). CONTAINS-SUBEVENT is stored as a TLINK.
  • Cross-document IAA was 93.77% for IDENTICAL, 36.43% for CON-SUB, 6.88% for SET-SUBSET and 0.00% for WHOLE-PART. Intra-document CON-SUB agreement was 34.14%.
  • WHOLE-PART agreement was zero mainly because few such links existed: anatomical entities were linked across documents only via IDENTICAL.
  • Future work includes refining the cross-document linking method and systematically incorporating medical domain expertise into annotation rules.

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

9 papers of our library cite it (of the 140673 whose bibliographies we read); 2 of them have a page here, 7 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/d19-6201

@inproceedings{wrightbettner2019cross,
  title = {{Cross-document coreference: An approach to capturing coreference without context}},
  author = {Wright-Bettner, Kristin and Palmer, Martha and Savova, Guergana and Groen, Piet de and Miller, Timothy},
  booktitle = {Proceedings of the Tenth International Workshop on Health Text Mining and Information Analysis (LOUHI 2019)},
  year = {2019},
  doi = {10.18653/v1/d19-6201},
  url = {https://doi.org/10.18653/v1/d19-6201}
}

References

12 references, 5 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Daniel Albright, Arrick Lanfranchi, Anwen Fredriksen, William F Styler IV, Colin Warner, Jena D Hwang, Jinho D Choi, Dmitriy Dligach, Rodney D Nielsen, James Martin, Wayne Ward, Martha Palmer, Guergana K Savova. 2013. Towards comprehensive syntactic and semantic annotations of the clinical narrative… — [doi:10.1136/amiajnl-2012-001317](https://doi.org/10.1136/amiajnl-2012-001317) find
  2. Yu Hong, Tongtao Zhang, Tim O’Gorman, Sharone Horowit-Hendler, Heng Ji, Martha Palmer. 2016. Building a Cross-document Event-Event Relation Corpus. In Proceedings of the 10th Linguistic Annotation Workshop held in conjunction with ACL 2016 (LAW-X 2016), Association for Computational Linguistics, pag… find
  3. Edward Hovy, Teruko Mitamura, Felisa Verdejo, Jun Araki, Andrew Philpot. 2013. Events are Not Simple: Identity, Non-Identity, and Quasi-Identity. In Workshop on Events: Definition, Detection, Coreference, and Representation, Association for Computational Linguistics, pages 21–28. https://www.aclweb.… ↗
  4. Alex Baron and Marjorie Freedman. 2008. Who is Who and What is What: Experiments in Cross-Document Co-Reference. In Proceedings of the 2008 Conference on Empirical Methods in Natural Language Processing, Association for Computational Linguistics, pages 274–283. https://www.aclweb.org/anthology/D08-1… ↗ our card
  5. Tim O’Gorman, Kristin Wright-Bettner, Martha Palmer. 2016. Richer Event Description: Integrating event coreference with temporal, causal and bridging annotation. In Proceedings of the 2nd Workshop on Computing News Storylines (CNS 2016), Association for Computational K. Bretonnel Cohen, Arrick Lanfr… find our card
  6. James Pustejovsky and Amber Stubbs. 2011. Increasing informativeness in temporal annotation. In Proceedings of the 5th Linguistic Annotation Workshop, Association for Computational Linguistics, pages 152–160. https://www.aclweb.org/anthology/W11-0419. reethi Raghavan, Eric osler- ussier, No mie Elha… ↗ our card
  7. Marta Recasens. 2010. “Coreference: Theory, Annotation, Resolution and Evaluation.” hD Thesis. University of Barcelona. http://stel.ub.edu/cba2010/phd/phd.pdf. ↗
  8. Marta Recasens, Eduard Hovy, M. Ant nia Mart . 2011. Identity, Non-identity, and Near-identity: Addressing the complexity of coreference. Lingua, 121(6): 1138-1152. https://doi.org/10.1016/j.lingua.2011.02.004. — [doi:10.1016/j.lingua.2011.02.004](https://doi.org/10.1016/j.lingua.2011.02.004) ↗
  9. Marta Recasens, M. Ant nia Mart , Constantin Orasan. 2012. Annotating Near-Identity from Coreference Disagreements. In Proceedings of the Eighth International Conference on Language Resources and Evaluation (LREC-2012), European Languages Resources Association, pages 165-172. http://www.lrec-conf.or… ↗ our card
  10. Zhiyi Song, Ann Bies, Justin Mott, Xuansong Li, Stephanie Strassel, Christopher Caruso. 2018. Cross-Document, Cross-Language Event Coreference Annotation Using Event Hoppers. In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC-2018), European Languages… find our card
  11. Amber Stubbs. 2013. "A Methodology for Using Professional Knowledge in Corpus Annotation." PhD Thesis. Brandeis University. http://amberstubbs.net/docs/AmberStubbs_disserta tion.pdf. ↗
  12. William F. Styler IV, Steven Bethard, Sean Finan, Martha Palmer, Sameer Pradhan, Piet C de Groen, Brad Erickson, Timothy Miller, Chen Lin, Guergana Savova and James Pustejovsky. 2014. Temporal Annotation in the Clinical Domain. Transactions of the Association for Computational Linguistics, 2, pages… find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/bettner-2019-cross-document-coreference-approach-capturing/