Elamur Research⌕
☰
home / papers / 2019 / Neural Coreference Resolution with Limited Lexical Context and Explicit Mention Detection for Oral French

Neural Coreference Resolution with Limited Lexical Context and Explicit Mention Detection for Oral French

Summary

This paper asks whether knowledge-poor neural coreference models, which perform strongly on CoNLL-2012/OntoNotes, carry over to other annotation paradigms, languages and domains. The test case is ANCOR, a corpus of transcribed spoken French that includes singleton mentions. The author adapts the E2EC architecture of K. Lee et al. (2018) in two main ways. First, each span is encoded using only its nearby left and right context rather than the whole document, which reduces memory cost and allows mini-batching and shuffling. Second, mention detection becomes a separate classification step, so singletons can be distinguished from non-mention spans and gold mentions can be supplied for evaluation. Other changes include character-level GRU word encodings, two-head soft-head attention, a context vector from the LSTM, and a learned score for new mentions. With gold mentions, the system scores close to earlier ANCOR results that relied on far richer linguistic input. It also sets a fully evaluated baseline for French coreference resolution.

Key points

  • Data: ANCOR has roughly 418,000 words and is the only publicly available French coreference corpus large enough for machine learning. The author's own split is 59% train, 12% development and 29% test.
  • Training: Mention detection and antecedent scoring are trained one after the other with AdamW. Mention detection uses class-weighted cross-entropy (weights 1 and 3), and non-mention spans in training are capped at 90%.
  • Coreference results (averaged over 5 runs):
    • MUC F 57.3, B³ F 79.2, CEAFe F 79.4, BLANC F 65.7.
    • CoNLL average 72.0, compared with 75.3 for Désoyer et al. (2015), whose system relies on gold features.
  • Evaluation: Mention detection and coreference are scored separately, using the CoNLL-2012 metrics plus BLANC. The original E2EC could not be compared directly because it cannot easily accept gold mentions.
  • Future work: The author proposes testing on ARRAU, the Polish Coreference Corpus, DEMOCRAT and CoNLL-2012.

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); 2 of them have a page here, 6 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/w19-2802

@inproceedings{grobol2019neural,
  title = {{Neural Coreference Resolution with Limited Lexical Context and Explicit Mention Detection for Oral French}},
  author = {Grobol, Loïc},
  booktitle = {Proceedings of the Second Workshop on Computational Models of Reference, Anaphora and Coreference},
  year = {2019},
  doi = {10.18653/v1/w19-2802},
  url = {https://doi.org/10.18653/v1/w19-2802}
}

References

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

  1. Olivier Baude and Céline Dugua. 2011. (Re)faire le corpus d’Orléans quarante ans après : quoi de neuf, linguiste ? Corpus. Varia, 10, 2011: 99– 118. find
  2. In Proceedings of Workshop for NLP Open Source Software, pages 1–6. Association for Computational Linguistics. Elisabeth Godbert and Benoît Favre. 2017. Détection de coréférences de bout en bout en français. In Actes de la 24e Conférence sur le Traitement Automatique des Langues Naturelles. Associat… find
  3. Orléans, France. Edouard Grave, Piotr Bojanowski, Prakhar Gupta, Armand Joulin, and Tomas Mikolov. 2018. Learning Word Vectors for 157 Languages. In Proceedings of the 11th International Conference on Language Resources and Evaluation. find our card
  4. European Language Resource Association. Loïc Grobol, Isabelle Tellier, Éric De La Clergerie, Marco Dinarelli, and Frédéric Landragin. 2017. Apports des analyses syntaxiques pour la détection automatique de mentions dans un corpus de français oral. In Actes de la 24e Conférence sur le Traitement Auto… find our card
  5. Orléans, France. Frédéric Landragin. 2016. Description, Modélisation et Détection Automatique Des Chaînes de Référence (DEMOCRAT). Bulletin de l’AFIA, 92, 2016: 11–15. find
  6. Heeyoung Lee, Angel Chang, Yves Peirsman, Nathanael Chambers, Mihai Surdeanu, and Dan Jurafsky. 2013. Deterministic Coreference Resolution Based on Entity-Centric, Precision-Ranked Rules. Computational Linguistics, 39.4, Dec. 2013: 885–916. find our card
  7. Kenton Lee, Luheng He, Mike Lewis, and Luke Zettlemoyer. 2017. End-to-End Neural Coreference Resolution. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 188–197. Association for Computational Linguistics. find our card
  8. København, Danmark. Kenton Lee, Luheng He, and Luke Zettlemoyer. 2018. Higher-Order Coreference Resolution with Coarse-to-Fine Inference. In Proceedings of the 2018 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 687–692.… find our card
  9. New Orleans, Louisiana. Ilya Loshchilov and Frank Hutter. 2019. Decoupled Weight Decay Regularization. In International Conference on Learning Representations. arXiv: 1711.05101. find
  10. Chao Ma, Janardhan Rao Doppa, J. Walker Orr, Prashanth Mannem, Xiaoli Fern, Tom Dietterich, and Prasad Tadepalli. 2014. Prune-and-Score: Learning for Greedy Coreference Resolution. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 2115–2126. Ass… find our card
  11. Xuezhe Ma and Eduard Hovy. 2016. End-to-End Sequence Labeling via Bi-Directional LSTM-CNNs-CRF. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics, pages 1064–1074. Volume 1. Association for Computational Linguistics. find our card
  12. Berlin, Deutschland. Judith Muzerelle, Anaïs Lefeuvre, Emmanuel Schang, Jean-Yves Antoine, Aurore Pelletier, Denis Maurel, Iris Eshkol, and Jeanne Villaneau. 2014. ANCOR Centre, a Large Free Spoken French Coreference Corpus: Description of the Resource and Reliability Measures. In Proceedings of the… find our card
  13. European Language Resources Association. Reykjavík, Ís-land. Alexis Nasr, Frédéric Béchet, Jean-François Rey, Benoit Favre, and Joseph Le Roux. 2011. MACAON : An NLP Tool Suite for Processing Word Lattices. In Proceedings of the 49th Annual Meeting of the Association for Computational Linguistics, p… find
  14. United States. Maciej Ogrodniczuk, Katarzyna Głowińska, Mateusz Kopeć, Agata Savary, and Magdalena Zawisławska. 2016. Polish Coreference Corpus. In Zygmunt Vetulani, Hans Uszkoreit, and Marek Kubis, editors, Human Language Technology. Challenges for Computer Science and Linguistics, pages 215–226. L… find
  15. Springer International Publishing. Sankaran Panchapagesan, Ming Sun, Aparna Khare, Spyros Matsoukas, Arindam Mandal, Björn Hoffmeister, and Shiv Vitaladevuni. 2016. Multi-Task Learning and Weighted Cross-Entropy for DNN-Based Keyword Spotting. In Proceedings of Interspeech, pages 760– 764. find
  16. Massimo Poesio and Ron Artstein. 2008. Anaphoric Annotation in the ARRAU Corpus. In Proceedings of the International Conference on Language Resources and Evaluation. find our card
  17. Marrakech, Morocco. Massimo Poesio, Ron Stuckardt, and Yannick Versley. 2016. Anaphora Resolution: Algorithms, Resources, and Applications. Theory and Applications of Natural Language Processing. find
  18. Springer Berlin Heidelberg. Massimo Poesio et al. 2018. Anaphora Resolution with the ARRAU Corpus. In Proceedings of the First Workshop on Computational Models of Reference, Anaphora and Coreference, pages 11– 22. Association for Computational Linguistics. find our card
  19. New Orleans, Louisiana. Sameer Pradhan, Xiaoqiang Luo, Marta Recasens, Eduard Hovy, Vincent Ng, and Michael Strube. 2014. Scoring Coreference Partitions of Predicted Mentions: A Reference Implementation. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics, page… find our card
  20. Baltimore, Maryland. Sameer Pradhan, Alessandro Moschitti, Nianwen Xue, Olga Uryupina, and Yuchen Zhang. 2012. CoNLL-2012 Shared Task: Modeling Multilingual Unrestricted Coreference in OntoNotes. In Proceedings of the Joint Conference on EMNLP and CoNLL, pages 1–40. Association for Computational Lin… find our card
  21. Jeju Island, Korea. Marta Recasens. 2010. find
  22. Coreference: Theory, Annotation, Resolution and Evaluation. Universitat de Barcelona, 2010. find
  23. Marta Recasens and Eduard Hovy. 2011. BLANC: Implementing the Rand Index for Coreference Evaluation. Natural Language Engineering, 17.4, Oct. 2011: 485–510. find
  24. Susanne Salmon-Alt, Laurent Romary, Andrei Popescu-Belis, and Loïs Rigouste. 2004. Online Evaluation of Coreference Resolution. In 4th International Conference on Language Resources and Evaluation. find
  25. European Language Resources Association. Lisboa, Portugal. Victor Sanh, Thomas Wolf, and Sebastian Ruder. 2018. A Hierarchical Multi-Task Approach for Learning Embeddings from Semantic Tasks, Nov. 14, 2018: arXiv: 1811.06031 [cs]. — [arXiv:1811.06031](https://arxiv.org/abs/1811.06031) find
  26. Agnès Tutin, François Trouilleux, Catherine Clouzot, Éric Gaussier, Annie Zaenen, Stéphanie Rayot, and Georges Antoniadis. 2000. Annotating a Large Corpus with Anaphoric Links. In Proceedings of the Third International Conference on Discourse Anaphora and Anaphora Resolution. find
  27. United Kingdom. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Łukasz Kaiser, and Illia Polosukhin. 2017. Attention Is All You Need. In I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, editors, Advances in Neural Inf… find
  28. Curran Associates, Inc. Long Beach, California. Sam Wiseman, Alexander M. Rush, and Stuart M. Shieber. 2016. Learning Global Features for Coreference Resolution. In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Techn… find our card
  29. Sam Wiseman, Alexander M. Rush, Stuart M. Shieber, and Jason Weston. 2015. Learning Anaphoricity and Antecedent Ranking Features for Coreference Resolution. In Proceedings of the 53rd Annual Meeting of the Association for Computational Linguistics and the 7th International Joint Conference on Natura… find our card

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/papers/grobol-2019-neural-coreference-resolution-limited-lexical/