Elamur Research⌕
☰
home / papers / 2010 / Extending BART to Provide a Coreference Resolution System for German

Extending BART to Provide a Coreference Resolution System for German

Summary

Supervised coreference resolution has mostly been developed for English. Several strong German systems exist, but none was publicly released, which makes it hard for others to start research on German. The authors address this by adapting the modular BART toolkit to German. They convert version 4 of the TüBa-D/Z treebank into MMAX2 standoff XML and derive markables from the gold parse trees. Predicative NPs, appositions, name-part modifiers, and non-referential es are filtered out. Number, gender, and semantic class are assigned using GermaNet, gazetteers, and morphological heuristics.

The baseline follows the pairwise classification and closest-first clustering scheme of Soon et al. (2001), with features modelled on Klenner and Ailloud (2009). The authors then add new features, including speaker person, quoted speech, tree distance, partial head match, and GermaNet relatedness.

MaxEnt outperforms J48, and training separate classifiers for pronouns and non-pronouns helps further. The full system does not reach Klenner and Ailloud's scores, but it shows that machine-learned coreference can be ported robustly beyond English. It also provides a freely available starting point for German.

Key points

  • Data: TüBa-D/Z v4 contains 32,945 sentences, from which the authors extract 144,942 markables, 52,386 coreferential links, and 14,073 coreference sets.
  • Evaluation setup:
    • Metrics are MUC and CEAF, computed on true (gold) mentions only.
    • The first 1100 documents are evaluated with 5-fold cross-validation, and the remaining documents serve as the development set.
  • Learners: The system is tested with WEKA's J48 (C4.5) and MaxEnt with feature combination. MaxEnt with a "split" pronoun/non-pronoun architecture gives the best results.
  • Feature contributions:
    • Used alone, only PARTIAL MATCH helps clearly (+1.4 MUC F1, +1.3 CEAF F1).
    • Combining all new features gives +2.1 MUC F1 and +3.3 CEAF F1.
    • GermaNet relatedness is computed with the Pathfinder library using the Wu & Palmer measure.
  • Gap to prior work: The system remains below Klenner and Ailloud (2009). The authors attribute this to closest-first clustering instead of aggressive merging and to a limited instance window.
  • Availability: BART is released at bart-anaphora.org, with ports to Italian and Spanish and an end-to-end preprocessing pipeline planned.

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

5 papers of our library cite it (of the 140673 whose bibliographies we read); 3 of them have a page here, 2 not yet.

Cite

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

@inproceedings{broscheit2010extending,
  title = {{Extending BART to Provide a Coreference Resolution System for German}},
  author = {Broscheit, Samuel and Ponzetto, Simone Paolo and Versley, Yannick and Poesio, Massimo},
  booktitle = {Proceedings of the Seventh International Conference on Language Resources and Evaluation (LREC'10)},
  year = {2010},
  doi = {10.63317/2o32xdc2fiyt},
  url = {https://doi.org/10.63317/2o32xdc2fiyt}
}

References

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

  1. Adam Berger, Stephen A. Della Pietra, and Vincent J. Della Pietra. 1996. A maximum entropy approach to natural language processing. Computational Linguistics, 22(1):39–71. ↗ our card
  2. Christian Biemann. 2002. Finden von semantischen Relationen in natürlichsprachlichen Texten mit Hilfe maschinellem Lernens. Diplomarbeit, Universität Leipzig. find
  3. Adriane Boyd, Whitney Gegg-Harrison, and Donna Byron. 2005. Identifying non-referential it: a machine learning approach incorporating linguistically motivated features. In Proceedings of the ACL-05 Workshop on Feature Engineering for Machine Learning in Natural Language Processing, pages 40–47. find
  4. Irene Cramer and Marc Finthammer. 2008. An evaluation procedure for WordNet based lexical chaining: Methods and issues. In Proc. of GWC-08, pages 120–147. find
  5. Marc Finthammer and Irene Cramer. 2008. Exploring and navigating: Tools for GermaNet. In Proc. of LREC ’08. find
  6. Kilian Foth and Wolfgang Menzel. 2006. Hybrid parsing: Using probabilistic models as predictors for a symbolic parser. In Proc. of COLING-ACL-06, pages 321–328. find
  7. Iryna Gurevych and Hendrik Niederlich. 2005. Accessing GermaNet data and computing semantic relatedness. In Comp. Vol. to Proc. of ACL-05, pages 5–8. find
  8. Sven Hartrumpf, Herrmann Helbig, and Rainer Osswald. 2003. The semantically based corpus HaGenLex - structure and technological environment. Traitement automatique des langues, 44(2):81–105. find
  9. Sven Hartrumpf. 2001. Coreference resolution with syntactico-semantic rules and corpus statistics. In Proc. of CoNLL-01, pages 137–144. find
  10. Erhard Hinrichs, Katja Filippova, and Holger Wunsch. 2005a. What treebanks can do for you: Rule-based and machine-learning approaches to anaphora resolution in German. In Proceedings of the 4th Workshop on Treebanks and Linguistic Theories, pages 77–88. find
  11. Erhard Hinrichs, Sandra Kübler, and Karin Naumann. 2005b. A unified representation for morphological, syntactic, semantic and referential annotations. In Proceedings of the ACL-05 Workshop on Frontiers in Corpus Annotation II: Pie in the Sky, pages 13–20. find
  12. Manfred Klenner and Étienne Ailloud. 2009. Optimization in coreference resolution is not needed: A nearly-optimal algorithm with intensional constraints. In Proc. of EACL-09, pages 442–450. find our card
  13. Beata Kouchnir. 2004. A machine learning approach to German pronoun resolution. In Proc. of ACL-04, pages 55–60. Student Session. find
  14. Lothar Lemnitzer and Claudia Kunze. 2002. GermaNet – representation, visualization, application. In Proc. of LREC ’02, pages 1485–1491. find
  15. J48 MaxEnt MaxEnt split Klenner and Ailloud (2009) Table 1: Performance for different classifiers baseline (MaxEnt split) + 1/2 P ERSON , S PEECH + N ODE D IST + P ARTIAL M ATCH + G ERMA N ET R ELATEDNESS + all features Klenner and Ailloud (2009) Table 2: Performance for different feature sets Xiaoqiang Luo, Abe Ittycheriah, Hongyan Jing, Nanda Kambhatla, and Salim Roukos. 2004. A mention-synchronous coreference resolution algorithm based on the Bell Tree. In Proc. of ACL-04, pages 136–143. find our card
  16. Xiaoqiang Luo. 2005. On coreference resolution performance metrics. In Proc. HLT-EMNLP ’05, pages 25–32. find our card
  17. Christoph Müller and Michael Strube. 2006. Multi-level annotation of linguistic data with MMAX2. In Sabine Braun, Kurt Kohn, and Joybrato Mukherjee, editors, Corpus Technology and Language Pedagogy: New Resources, New Tools, New Methods, pages 197–214. Peter Lang: Frankfurt a.M., Germany. find
  18. Vincent Ng and Claire Cardie. 2002. Improving machine learning approaches to coreference resolution. In Proc. of ACL-02, pages 104–111. find our card
  19. Massimo Poesio and Mijail A. Kabadjov. 2004. A general-purpose, off-the-shelf anaphora resolution module: Implementation and preliminary evaluation. In Proc. of LREC ’04, pages 663–666. find our card
  20. Simone Paolo Ponzetto and Michael Strube. 2006. Exploiting semantic role labeling, WordNet and Wikipedia for coreference resolution. In Proc. of HLT-NAACL-06, pages 192–199. find our card
  21. J. Ross Quinlan. 1993. C4.5: Programs for Machine Learning. Morgan Kaufman, San Mateo, Cal. find our card
  22. Michael Schiehlen. 2004. Optimizing algorithms for pronoun resolution. In Proc. of COLING-04, pages 390– 396. find
  23. Wee Meng Soon, Hwee Tou Ng, and Daniel Chung Yong Lim. 2001. A machine learning approach to coreference resolution of noun phrases. Computational Linguistics, 27(4):521–544. ↗ our card
  24. Michael Strube, Stefan Rapp, and Christoph Müller. 2002. The influence of minimum edit distance on reference resolution. In Proc. EMNLP-02, pages 312–319. find
  25. Michael Strube. 2007. Corpus-based and machine learning approaches to coreference resolution. In M. Schwarz-Friesel, M. Consten, and M. Knees, editors, Anaphors in Text. Cognitive, formal and applied approaches to MUC scorer R P F 1 60.9 70.7 65.4 71.2 78.0 74.4 75.6 80.8 78.1 − − − CEAF P 56.1 64.0 67.0 73.8 R 52.9 60.4 63.2 69.3 F 1 54.4 62.2 65.0 71.5 CEAF P 67.0 67.4 67.1 68.3 66.8 70.3 73.8 MUC scorer R P F 1 75.6 80.8 78.1 76.2 80.9 78.4 75.7 80.9 78.2 77.8 81.3 79.5 76.4 80.6 78.5 78.4 82.2 80.2 − − − R 63.2 63.6 63.3 64.4 63.0 66.3 69.3 F 1 65.0 65.4 65.1 66.3 64.8 68.3 71.5 anaphoric reference, pages 207–222. Amsterdam, The Netherlands: John Benjamins. find
  26. Roland Stuckardt. 2004. Three algorithms for competence-oriented anaphor resolution. In Proceedings of the 5th Discourse Anaphora and Anaphor Resolution Colloquium, pages 157–163. find
  27. Yannick Versley, Simone Paolo Ponzetto, Massimo Poesio, Vladimir Eidelman, Alan Jern, Jason Smith, Xiaofeng Yang, and Alessandro Moschitti. 2008. BART: A modular toolkit for coreference resolution. In Proc. of LREC ’08. find our card
  28. Yannick Versley. 2006. A constraint-based approach to noun phrase coreference resolution in German newspaper text. In Proceedings of Konferenz zur Verarbeitung Natürlicher Sprache, pages 143–150. find
  29. Marc Vilain, John Burger, John Aberdeen, Dennis Connolly, and Lynette Hirschman. 1995. A model-theoretic coreference scoring scheme. In Proceedings of the 6th Message Understanding Conference (MUC-6), pages 45–52, San Mateo, Cal. Morgan Kaufmann. find our card
  30. Christopher Walker, Stephanie Strassel, Julie Medero, and Kazuaki Maeda. 2006. ACE 2005 multilingual training corpus. LDC2006T06, Philadelphia, Penn.: Linguistic Data Consortium. ↗
  31. Ralph Weischedel, Sameer Pradhan, Lance Ramshaw, Martha Palmer, Nianwen Xue, Mitchell Marcus, Ann Taylor, Craig Greenberg, Eduard Hovy, Robert Belvin, and Ann Houston. 2008. Ontonotes release 2.0. LDC2008T04, Philadelphia, Penn.: Linguistic Data Consortium. ↗
  32. Ian H. Witten and Eibe Frank. 2005. Data Mining: Practical Machine Learning Tools and Techniques. Morgan Kaufmann, San Francisco, Cal., 2nd edition. find
  33. Zhibiao Wu and Martha Palmer. 1994. Verb semantics and lexical selection. In Proc. of ACL-94, pages 133–138. ↗
  34. Xiaofeng Yang, Guodung Zhou, Jian Su, and Chew Lim Tan. 2003. Coreference resolution using competition learning approach. In Proc. of ACL-03, pages 176–183. find our card

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/broscheit-2010-extending-bart-provide-coreference-resolution/