Elamur Research⌕
☰
home / papers / 2015 / Rule-based Coreference Resolution in German Historic Novels

Rule-based Coreference Resolution in German Historic Novels

Summary

Automatically tracking characters in fiction requires linking every mention of a person, including names, descriptions and pronouns, into chains. Off-the-shelf coreference tools, built mostly on news data, perform poorly on 19th-century German novels. These texts have longer sentences, more pronouns, extensive dialogue, and a few main characters who recur constantly. Krug and colleagues adapt the multi-pass sieve design of Lee et al. (2011) to this genre, expanding it to 11 rule passes restricted to person entities. The passes cover string and name-variant matching, head and title matching, GermaNet synonyms, constraint-based pronoun resolution, and the identification of speakers and addressees in direct speech. The authors annotated their own corpus of 130-sentence novel excerpts. On it, the system clearly outperforms CorZu and stays stable on a separate unseen test set. Because the system is rule-based, its decisions can be traced, which supports systematic error analysis. The authors also intend to use it to bootstrap a larger annotated corpus for later machine-learning work.

Key points

  • Corpus: The data consists of excerpts of 130 consecutive sentences, each taken from a different novel. Five excerpts were used for development. The first test set has 48 excerpts (about 143,000 tokens and 19,000 references). A second, unseen test set has 30 excerpts (about 104,000 tokens and 11,600 entities). Roughly 70% of mentions are pronouns.
  • Pipeline: Preprocessing uses TreeTagger, OpenNLP and the MATE parser. Gender and number are assigned through a precision-ordered cascade of name lists, a maximum entropy model trained on TIGER, MATE morphology and SMOR.
  • Comparison with CorZu (48 excerpts): The system reaches a MUC F1 of 85.5% against CorZu's 65.9%, and a B³ F1 of 56.0% against 33.6%.
  • Robustness and pass analysis: On the 30 unseen excerpts, scores are MUC F1 86% and B³ F1 55.5%. Adding the pronoun pass alone raises MUC F1 from 38.9% to 83.3%.
  • Error analysis (5 excerpts): Semantic or world-knowledge errors account for 37% of mistakes. Heuristic errors, such as salience and parser problems, make up 35%. Gender, number and person errors contribute 14%, and direct-speech errors another 14%.

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

8 papers of our library cite it (of the 140673 whose bibliographies we read); 5 of them have a page here, 3 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.3115/v1/w15-0711

@inproceedings{krug2015rule,
  title = {{Rule-based Coreference Resolution in German Historic Novels}},
  author = {Krug, Markus and Puppe, Frank and Jannidis, Fotis and Macharowsky, Luisa and Reger, Isabella and Weimar, Lukas},
  booktitle = {Proceedings of the Fourth Workshop on Computational Linguistics for Literature},
  year = {2015},
  doi = {10.3115/v1/w15-0711},
  url = {https://doi.org/10.3115/v1/w15-0711}
}

References

25 references, 14 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Aone, C. and Bennett, S. 1995. Evaluating Automated and Manual Acquisition of Anaphora Resolution Strategies. Proceedings of the 33rd annual meeting on Association for Computational Linguistics (ACL '95). Association for Computational Linguistics, 122-129. find our card
  2. Bohnet, B. 2010. Very High Accuracy and Fast Dependency Parsing is not a Contradiction. The 23rd Int. Conf. on Computational Linguistics (COLING 2010), Beijing, China. find our card
  3. Brants, S., Dipper, S., Eisenberg, P., Hansen, S., König, E., Lezius, W., Rohrer, C., Smith, G. and Uszkoreit, H. 2004. TIGER: Linguistic Interpretation of a German Corpus. Journal of Language and Computation 2 (4), 597-620. ↗
  4. Brunner, A. 2015. Automatische Erkennung von Redewiedergabe: Ein Beitrag zur Quantitativen Narratologie (Narratologia). [Automatic Recognition of Recorded Speech: a Contribution to Quantitative Narratologogia] Walter De Gruyter Inc. find
  5. Durrett G., Hall D., and Klein D. 2013. Decentralized Entity-Level Modeling for Coreference Resolution. In Proceedings of the 51st Annual Meeting of the Association for Computational Linguistics (Volume 1), 114-124. find our card
  6. Ferrucci, D. and Lally, A. 2004. UIMA: An Architectural Approach to Unstructured Information Processing in the Corporate Research Environment. Nat. Lang. Eng. 10, 327-348. http://dx.doi.org/10.1017/S1351324904003523 ↗
  7. Fitschen, A., Schmid, H. and Heid, U. 2004. SMOR: A German computational morphology covering derivation, composition, and inflection. Proceedings of the IVth International Conference on Language Resources and Evaluation (LREC 2004), 1263-1266. find our card
  8. Haghighi, A. and Klein, D. 2009. Simple Coreference Resolution with Rich Syntactic and Semantic Features. In Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing. Association for Computational Linguistics, Singapore, 1152-1161. find our card
  9. Hinrichs, E., Kübler, S. and Naumann, K. 2005. A unified representation for morphological, syntactic, semantic, and referential annotations. Proceedings of the ACL Workshop on Frontiers in Corpus Anotation II: Pie in the Sky, pages 13–20, Ann Arbor, MI. find
  10. Hobbs, J.1976. Pronoun Resolution. Technical report, Dept. of Computer Science, CUNY, Technical Report TR761. find
  11. Iosif, E. and Mishra, T. 2014. From Speaker Identification to Affective Analysis: A Multi-Step System for Analyzing Children`s Stories. Proceedings of the 3rd Workshop on Computational Linguistics for Literature. Gothenburg, Sweden, 40-49. find
  12. Jannidis, F., Krug, M., Reger, I. Toepfer, M. Weimer, L, Puppe, F. 2015. Automatische Erkennung von Figuren in deutschsprachigen Romanen. [Automatic recognition of Characters in German novels] Digital Humanities im deutschsprachigen Raum (Dhd 2015), Graz, Austria, 2015. find
  13. Klenner, M; Tuggener, D. 2011. An Incremental Entity-mention Model for Coreference Resolution with Restrictive Antecedent Accessibility. Recent Advances in Natural Language Processing (RANLP 2011), Hissar, Bulgaria, 178-185. find our card
  14. Lee, H., Peirsman, Y., Chang, A., Chambers, N., Surdeanu, M. and Jurafsky, D. 2011. Stanford's Multi-pass Sieve Coreference Resolution System at the CoNLL-2011 Shared Task. Proc. of the 15th Conference on Computational Natural Language Learning: Shared Task (CONLL Shared Task '11). Association for Computational Linguistics, 28-34. find our card
  15. Luo, X. 2005. On Coreference Resolution Performance Metrics. Proc. of the conference on Human Language Technology and Empirical Methods in Natural Language Processing (HLT '05). Association for Computational Linguistics, 25-32. find our card
  16. Ng, V. 2010. Supervised Noun Phrase Coreference Research: The First Fifteen Years. Proceedings of the 48th Annual Meeting of the Association for Computational Linguistics (ACL '10). Association for Computational Linguistics, 1396-1411. find our card
  17. Pascal, D. and Baldridge, J. 2008. Specialized models and ranking for coreference resolution. In Proceedings of the Conference on Empirical Methods in Natural Language Processing (EMNLP '08). Association for Computational Linguistics, Stroudsburg, PA, USA, 660-669. find our card
  18. Pradhan, S., Moschitti, A., Xue, N., Uryupina, O. and Zhang, Y. 2012. CoNLL-2012 Shared Task: Modeling Multilingual Unrestricted Coreference in Onto-Notes. In Proceedings of EMNLP and CoNLL-2012: Shared Task, 1-40. find
  19. Pradhan, S., Ramshaw, L., Marcus, M., Palmer, M., Weischedel, R., and Xue, N. 2011. CoNLL-2011 Shared Task: Modeling Unrestricted Coreference in OntoNotes. Proceedings of the Fifteenth Conference on Computational Natural Language Learning: Shared Task (CONLL Shared Task '11). Association for Computational Linguistics, 1-27. find our card
  20. Rahman, A and Ng, V. 2009. Supervised Models for Coreference Resolution. Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing, 968-977. find our card
  21. Recasens, M., Martí, T., Taulé, M., Màrquez, L., and Sapena, E. 2009. Coreference Resolution in Multiple Languages. Proceedings of the Workshop on Semantic Evaluations: Recent Achievements and Future Directions (SEW-2009), 70–75. find
  22. Schmid, H. 1995. Improvements in Part-of-Speech Tagging with an Application to German. Proceedings of the ACL SIGDAT-Workshop. Dublin, Ireland. find
  23. Sennrich, R., Schneider, G., Volk, M., Warin, M. 2009. A New Hybrid Dependency Parser for German. Proceedings of GSCL Conference, Potsdam. ↗
  24. Soon, W, Ng, H. and Lim, D. 2001. A Machine Learning Approach to Coreference Resolution of Noun Phrases. Comput. Linguist. 27 (4), 521-544. ↗ our card
  25. Yang, J., Mao, Q., Xiang, Q., Tsang, I., Chai, K., Chieu, H. 2012. Domain Adaptation for Coreference Resolution: An Adaptive Ensemble Approach. Proceedings of the 2012 Joint Conference on Empirical Methods in NLP and CNLL, 744-753. find our card

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/krug-2015-rule-based-coreference-resolution-german/