Coreference Resolution in a Modular, Entity-Centered Model

Summary

The paper addresses coreference resolution, focusing on a common source of errors: weak modeling of whether two mentions are semantically compatible. Earlier systems either listed compatible headword pairs or used coarse NER labels. Haghighi and Klein instead propose a generative, mostly unsupervised model with three separate modules. A semantic module draws latent entity types, such as PERSON, ORG, WEAPON or VEHICLE. Each type generates specific entities, and each entity holds word lists for properties like heads, modifiers and governors. A discourse module chooses antecedents with a log-linear, feature-based distance-dependent Chinese Restaurant Process. A mention module then builds each mention from its entity and type. The only supervision is a short list of prototype words per type. Lexical statistics are learned from large amounts of unlabeled text and shared at the type level. On ACE data with automatically detected mentions, the model beat strong supervised and deterministic systems on every metric reported. This shows that rich unsupervised semantic knowledge can lift end-to-end coreference performance.

Key points

The model has three layers (types → entities → mentions). Each type stores a word distribution plus a "fertility" distribution over list lengths for each property, so knowledge is shared at both the entity level and the type level.
Seed types are the ACE categories (PERS, ORG, GPE, WEA, VEH, LOC, FAC) plus an added GROUP type. Each gets up to four prototype heads taken from A04CU and A05ST. An equal number of unconstrained "other" types is induced automatically.
Evaluation used MUC, B³ All, B³ None and Pairwise F1 on ACE sets including A04CU, A05ST and A05RA. The model outperformed Stoyanov et al. (2009), Rahman and Ng (2009) and Haghighi and Klein (2009). Scores on A05RA were lower because its blog and forum text hurts the parser.
In the oracle-filtered mention setting on A04CU, the model reaches 80.3 B³, against 78.4 reported by Bengston and Roth (2008).
Used on its own as an NER classifier, the type component reaches 83.2 accuracy on the A05ST dev set. Most of its errors confuse GPE with ORG.

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

76 papers of our library cite it (of the 140673 whose bibliographies we read); 40 of them have a page here, 36 not yet.

Cite

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

@inproceedings{haghighi2010coreference,
  title = {{Coreference Resolution in a Modular, Entity-Centered Model}},
  author = {Haghighi, Aria and Klein, Dan},
  booktitle = {Human Language Technologies: The 2010 Annual Conference of the North American Chapter of the Association for Computational Linguistics},
  year = {2010},
  url = {https://aclanthology.org/N10-1061/}
}

References

26 references, 15 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

A Bagga and B Baldwin. 1998. Algorithms for scoring coreference chains. In Linguistic Coreference Workshop (LREC). find our card
Eric Bengston and Dan Roth. 2008. Understanding the Value of Features for Corefernce Resolution. In Empirical Methods in Natural Language Processing (EMNLP). find
David Blei and Peter I. Frazier. 2009. Distance Dependent Chinese Restaurant Processes. http://arxiv.org/abs/0910.1022/. — [arXiv:0910.1022](https://arxiv.org/abs/0910.1022) ↗
Eugene Charniak. 2000. Maximum Entropy Inspired Parser. In North American Chapter of the Association of Computational Linguistics (NAACL). find
Michael Collins and Yoram Singer. 1999. Unsupervised Models for Named Entity Classification. In Empirical Methods in Natural Language Processing (EMNLP). find
Mike Collins. 1999. Head-Driven Statistical Models for Natural Language Parsing. Ph.D. thesis, University of Pennsylvania. find
A Culotta, M Wick, R Hall, and A McCallum. 2007. First-order Probabilistic Models for Coreference Resolution. In Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing (NAACL-HLT). find our card
M. C. de Marneffe, B. Maccartney, and C. D. Manning. 2006. Generating Typed Dependency Parses from Phrase Structure Parses. In LREC. find our card
M Elsner, E Charniak, and M Johnson. 2009. Structured generative models for unsupervised named-entity clustering. In Proceedings of Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, pages 164–172. find
Barbara J. Grosz, Aravind K. Joshi, and Scott Weinstein. 1995. Centering: A Framework for Modeling the Local Coherence of Discourse. Computational Linguistics, 21(2):203–225. find our card
Aria Haghighi and Dan Klein. 2006. Prototype-Driven Learning for Sequence Models. In HLT-NAACL. Association for Computational Linguistics. find
Aria Haghighi and Dan Klein. 2007. Unsupervised Coreference Resolution in a Nonparametric Bayesian Model. In Proceedings of the 45th Annual Meeting of the Association of Computational Linguistics. Association for Computational Linguistics. find our card
Aria Haghighi and Dan Klein. 2009. Simple Coreference Resolution with Rich Syntactic and Semantic Features. In Proceedings of the 2009 Conference on Empirical Conference in Natural Language Processing. find our card
J. R. Hobbs. 1978. Resolving Pronoun References. Lingua, 44. find our card
J. R. Hobbs. 1979. Coherence and Coreference. Cognitive Science, 3:67–90. find
Andrew Kehler, Laura Kertz, Hannah Rohde, and Jeffrey Elman. 2008. Coherence and Coreference Revisited. find
Vincent Ng and Claire Cardie. 2002. Improving Machine Learning Approaches to Coreference Resolution. In Association of Computational Linguists (ACL). find our card
Vincent Ng. 2005. Machine Learning for Coreference Resolution: From Local Classification to Global Ranking. In Association of Computational Linguists (ACL). find our card
Vincent Ng. 2007. Shallow semantics for coreference resolution. In IJCAI’07: Proceedings of the 20th international joint conference on Artifical intelligence, pages 1689–1694. find
Slav Petrov, Leon Barrett, Romain Thibaux, and Dan Klein. 2006. Learning Accurate, Compact, and Interpretable Tree Annotation. In Proceedings of the 21st International Conference on Computational Linguistics and 44th Annual Meeting of the Association for Computational Linguistics, pages 433–440, Syd… find our card
J. Pitman. 2002. Combinatorial Stochastic Processes. In Lecture Notes for St. Flour Summer School. find
A Rahman and V Ng. 2009. Supervised models for coreference resolution. In Proceedings of the 2009 Conference on Empirical Conference in Natural Language Processing. find our card
W.H. Soon, H. T. Ng, and D. C. Y. Lim. 1999. A Machine Learning Approach to Coreference Resolution of Noun Phrases. find our card
V Stoyanov, N Gilbert, C Cardie, and E Riloff. 2009. Conundrums in Noun Phrase Coreference Resolution: Making Sense of the State-of-the-art. In Associate of Computational Linguistics (ACL). find our card
Marc Vilain, John Burger, John Aberdeen, Dennis Connolly, and Lynette Hirschman. 1995. A model-theoretic coreference scoring scheme. In MUC-6. find our card
X Yang, J Su, and CL Tan. 2005. Improving pronoun resolution using statistics-based semantic compatibility information. In Association of Computational Linguists (ACL). find our card

↑ back to top

📄 Coreference Resolution in a Modular, Entity-Centered Model