Elamur Research⌕
☰
home / papers / 2018 / Resolving Event Coreference with Supervised Representation Learning and Clustering-Oriented Regularization

Resolving Event Coreference with Supervised Representation Learning and Clustering-Oriented Regularization

Summary

The paper addresses event coreference resolution: deciding which event mentions, within one document or across several, describe the same real-world event. Most earlier systems make pairwise coreference decisions and then combine them into clusters, which can produce inconsistent results. The authors instead learn embeddings that are directly suited to clustering. They train an hourglass-shaped feed-forward network to predict which training-set cluster each mention belongs to, with all singletons grouped into one extra class. The loss adds Clustering-Oriented Regularization (CORE) terms to categorical cross-entropy. These terms push embeddings of coreferent mentions closer together and separate non-coreferent ones under cosine distance. At test time, activations from a middle layer are clustered agglomeratively, optionally starting from a lemma-based initialization. On ECB+ with gold event mentions, the system matches or beats the feature-rich model of Cybulska and Vossen (2015) on all six metrics, while needing less pre-annotated event structure. This suggests that regularizing representations for clustering may also help other clustering problems.

Key points

  • Data: Experiments use the ECB+ corpus, which annotates both within- and cross-document coreference. It adds 502 documents to ECB, for a total of 982, to increase lexical diversity.
  • Model variants: The authors compare CCE, CORE, CORE+CCE and CORE+CCE+LEMMA. The last variant seeds clustering with the lemma-δ baseline of Upadhyay et al. (2016).
  • Training and clustering setup:
    • Training runs for 100 epochs.
    • Each batch has 272 mentions and is forced to include both coreferent and non-coreferent pairs.
    • Clustering uses single-linkage with cosine similarity.
    • The stopping threshold τ is tuned on validation data to maximize B³.
  • Within-document results: CORE+CCE+LEMMA gains about 10 points in MUC over both LEMMA-δ and CORE+CCE. It also improves BLANC by 4–5 points.
  • Release: The code is publicly available at github.com/kiankd/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

32 papers of our library cite it (of the 140673 whose bibliographies we read); 17 of them have a page here, 15 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/s18-2001

@inproceedings{kenyondean2018resolving,
  title = {{Resolving Event Coreference with Supervised Representation Learning and Clustering-Oriented Regularization}},
  author = {Kenyon-Dean, Kian and Cheung, Jackie Chi Kit and Precup, Doina},
  booktitle = {Proceedings of the Seventh Joint Conference on Lexical and Computational Semantics},
  year = {2018},
  doi = {10.18653/v1/s18-2001},
  url = {https://doi.org/10.18653/v1/s18-2001}
}

References

29 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Amit Bagga and Breck Baldwin. 1998. Algorithms for scoring coreference chains. In The first international conference on language resources and evaluation workshop on linguistics coreference, volume 1, pages 563–566. find our card
  2. Jonas Gehring, Yajie Miao, Florian Metze, and Alex Waibel. 2013. Extracting deep bottleneck features using stacked auto-encoders. In Acoustics, Speech and Signal Processing (ICASSP), 2013 IEEE International Conference on, pages 3377–3381. IEEE. find
  3. Amit Bagga and Breck Baldwin. 1999. Cross-document event coreference: Annotations, experiments, and observations. In Proceedings of the Workshop on Coreference and its Applications, pages 1–8. ACL. find our card
  4. Frantisek Grézl, Martin Karafiát, Stanislav Kontár, and Jan Cernocky. 2007. Probabilistic and bottle-neck features for LVCSR of meetings. In Acoustics, Speech and Signal Processing, 2007. find
  5. ICASSP 2007. IEEE International Conference on, volume 4, pages IV–757. IEEE. find
  6. Cosmin Adrian Bejan and Sanda Harabagiu. 2014. Unsupervised Event Coreference Resolution. Computational Linguistics, 40(2):311–347. find our card
  7. Diederik Kingma and Jimmy Ba. 2014. Adam: A method for stochastic optimization. arXiv preprint arXiv:1412.6980. find
  8. Cosmin Adrian Bejan and Sanda M Harabagiu. 2008. A Linguistic Resource for Discovering Event Structures and Resolving Event Coreference. In LREC. find our card
  9. Sebastian Krause, Feiyu Xu, Hans Uszkoreit, and Dirk Weissenborn. 2016. Event linking with sentential features from convolutional neural networks. In Proceedings of the 20th SIGNLL Conference on Computational Natural Language Learning, pages 239–249. find
  10. Yoshua Bengio, Aaron Courville, and Pascal Vincent. 2013. Representation learning: A review and new perspectives. IEEE transactions on pattern analysis and machine intelligence, 35(8):1798–1828. find our card
  11. Jonathan Berant, Vivek Srikumar, Pei-Chun Chen, Abby Vander Linden, Brittany Harding, Brad Huang, Peter Clark, and Christopher D Manning. 2014. Modeling Biological Processes for Reading Comprehension. In Proceedings of the 2014 Conference on EMNLP, pages 1499–1510. find
  12. Linguistic Data Consortium. 2005. ACE (Automatic Content Extraction) English Annotation Guidelines for Events. Version 5.4.3 2005.07.01. find
  13. Linguistic Data Consortium. 2016. Rich ERE Annotation Guidelines Overview. V4.2. find
  14. Zheng Chen and Heng Ji. 2009. Graph-based event coreference resolution. In Proceedings of the 2009 Workshop on Graph-based Methods for NLP, pages 54–57. ACL. find our card
  15. Jing Lu and Vincent Ng. 2017. Learning antecedent structures for event coreference resolution. In Machine Learning and Applications (ICMLA), 2017 coherence model. In Proceedings of the 2015 Conference of NAACL: Human Language Technologies, pages 11–20. 16th IEEE International Conference on, pages 11… find
  16. Xiaoqiang Luo. 2005. On coreference resolution performance metrics. In Proceedings of the conference on Human Language Technology and EMNLP, pages 25–32. ACL. find our card
  17. Travis Wolfe, Benjamin Van Durme, Mark Dredze, Nicholas Andrews, Charley Beller, Chris Callison-Burch, Jay DeYoung, Justin Snyder, Jonathan Weese, Tan Xu, et al. 2013. PARMA: A Predicate Argument Aligner. In ACL (2), pages 63–68. find
  18. Xiaoqiang Luo, Sameer Pradhan, Marta Recasens, and Eduard H Hovy. 2014. An Extension of BLANC to System Mentions. In ACL (2), pages 24–29. find
  19. Bishan Yang, Claire Cardie, and Peter Frazier. 2015. A Hierarchical Distance-dependent Bayesian Model for Event Coreference Resolution. Transactions of the Association for Computational Linguistics, 3:517–528. find
  20. Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg S Corrado, and Jeff Dean. 2013. Distributed representations of words and phrases and their compositionality. In Advances in neural information processing systems, pages 3111–3119. find
  21. Sameer Pradhan, Xiaoqiang Luo, Marta Recasens, Eduard H Hovy, Vincent Ng, and Michael Strube. 2014. Scoring coreference partitions of predicted mentions: A reference implementation. In ACL (2), pages 30–35. find our card
  22. James Pustejovsky, José M Castano, Robert Ingria, Roser Sauri, Robert J Gaizauskas, Andrea Setzer, Graham Katz, and Dragomir R Radev. 2003. TimeML: Robust specification of event and temporal expressions in text. New directions in question answering, 3:28–34. find
  23. Michael Roth and Anette Frank. 2012. Aligning predicate argument structures in monolingual comparable texts: A new corpus for a new task. In Proceedings of the First Joint Conference on Lexical and Computational Semantics-Volume 1: Proceedings of the main conference and the shared task, and Volume 2… find
  24. Rui Sun, Yue Zhang, Meishan Zhang, and Donghong Ji. 2015. Event-driven headline generation. In Proceedings of ACL, pages 462–472. find
  25. Shyam Upadhyay, Nitish Gupta, Christos Christodoulopoulos, and Dan Roth. 2016. Revisiting the Evaluation for Cross Document Event Coreference. In COLING. find our card
  26. Lucy Vanderwende, Michele Banko, and Arul Menezes. 2004. Event-centric summary generation. Working notes of DUC, pages 127–132. find
  27. Marc Vilain, John Burger, John Aberdeen, Dennis Connolly, and Lynette Hirschman. 1995. A model-theoretic coreference scoring scheme. In Proceedings of the 6th conference on Message understanding, pages 45–52. ACL. find our card
  28. Piek Vossen and Agata Cybulska. 2017. Identity and granularity of events in text. arXiv preprint arXiv:1704.04259. — [arXiv:1704.04259](https://arxiv.org/abs/1704.04259) find
  29. Travis Wolfe, Mark Dredze, and Benjamin Van Durme. 2015. Predicate argument alignment using a global find

↑ back to top

This page in Markdown →