Elamur Research
☰
home / papers / 2009 / Recognizing Implicit Discourse Relations in the Penn Discourse Treebank

Recognizing Implicit Discourse Relations in the Penn Discourse Treebank

Written for the Research library (claude-opus-5-5) from the paper text, in our own words: the license "CC BY-NC-SA 3.0 (ACL Anthology, before 2016)" does not allow redistributing the text itself. Original — aclanthology.org/D09-1036.pdf

Summary

Discourse relations that carry no explicit connective, such as "because" or "but", are much harder to identify automatically than those that do. Before this paper, few published studies had tried to classify them. Lin, Kan and Ng built one of the first classifiers for these implicit relations on the second release of the Penn Discourse Treebank (PDTB 2.0). The PDTB 2.0 adds discourse annotation to the Wall Street Journal corpus. The task is to assign one of 11 Level 2 relation types to each pair of adjacent arguments. The authors trained a maximum entropy model on four kinds of features: contextual features, constituent-parse features, dependency-parse features and word-pair features. The contextual features capture how a relation is linked to the relations just before and after it. Using gold-standard argument spans, the system reached 40.2% accuracy, 14.1 points above the majority-class baseline. An error analysis found four obstacles: ambiguity between relation types, the need for semantic inference, dependence on wider context, and missing world knowledge. The work set an early reference point for implicit relation recognition, which matters for summarization, question answering and textual entailment.

Key points

  • Data: implicit relations from PDTB 2.0, trained on WSJ Sections 2–21. Five rare Level 2 types were dropped, leaving 11 types and 13,366 training instances. Cause (~25.6%), Conjunction (~22.3%) and Restatement (~19.2%) are the most frequent.
  • Model: a maximum entropy classifier using binary features. Contextual features come from neighbouring relations (two dependency patterns between relations: fully embedded and shared arguments) and from the connectives of adjacent explicit relations. The other features are constituent and dependency parse features and word pairs.
  • Result: 40.2% accuracy against a majority baseline of about 26.1%, an absolute gain of 14.1 points.
  • Comparison with earlier work: the authors argue that "implicit" data made by deleting connectives from explicit relations, as in Marcu & Echihabi (2002), differs from naturally implicit relations.
  • Error analysis: Contrast is often mislabelled as Conjunction. Some relations can only be resolved with formal-semantic reasoning, long-range context or world knowledge (for example, the "Trojan Horse" metaphor).

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: Discourse analysis · Treebanks

Cited by

138 papers of our library cite it (of the 140673 whose bibliographies we read); 8 of them have a page here, 130 not yet.

Cite

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

@inproceedings{lin2009recognizing,
  title = {{Recognizing Implicit Discourse Relations in the Penn Discourse Treebank}},
  author = {Lin, Ziheng and Kan, Min-Yen and Ng, Hwee Tou},
  booktitle = {Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing},
  year = {2009},
  doi = {10.3115/1699510.1699555},
  url = {https://doi.org/10.3115/1699510.1699555}
}

References

19 references, 6 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Marie-Catherine de Marneffe, Bill MacCartney, and Christopher D. Manning. 2006. Generating typed dependency parses from phrase structure parses. In Proceedings of the Fifth International Conference on Language Resources and Evaluation (LREC 2006), pages 449–454. find our card
  2. Jerry R. Hobbs. 1990. Literature and cognition. In CSLI Lecture Notes Number 21. CSLI Publications. find
  3. Zheng Ping Jiang and Hwee Tou Ng. 2006. Semantic role labeling of NomBank: A maximum entropy approach. In Proceedings of the 2006 Conference on Empirical Methods in Natural Language Processing (EMNLP 2006), pages 138–145, Sydney, Australia. find
  4. Alistair Knott and Ted Sanders. 1998. The classification of coherence relations and their linguistic markers: An exploration of two languages. Journal of Pragmatics, 30(2):135–175. find
  5. Alex Lascarides and Nicholas Asher. 1993. Temporal interpretation, discourse relations and commonsense entailment. Linguistics and Philosophy, 16(5):437– 493. find
  6. Alan Lee, Rashmi Prasad, Aravind Joshi, Nikhil Dinesh, and Bonnie Webber. 2006. Complexity of dependencies in discourse: Are dependencies in discourse more complex than in syntax? In Proceedings of the 5th International Workshop on Treebanks and Linguistic Theories, Prague, Czech Republic, December. find
  7. William C. Mann and Sandra A. Thompson. 1988. Rhetorical Structure Theory: Toward a functional theory of text organization. Text, 8(3):243–281. ↗ our card
  8. Daniel Marcu and Abdessamad Echihabi. 2002. An unsupervised approach to recognizing discourse relations. In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics (ACL 2002), pages 368–375, Morristown, NJ, USA. find our card
  9. Daniel Marcu. 1998. A surface-based approach to identifying discourse markers and elementary textual units in unrestricted texts. In Proceedings of the COLING-ACL 1998 Workshop on Discourse Relations and Discourse Markers, pages 1–7, Montreal, Canada, August. find
  10. Eleni Miltsakaki, Nikhil Dinesh, Rashmi Prasad, Aravind Joshi, and Bonnie Webber. 2005. Experiments on sense annotations and sense disambiguation of discourse connectives. In Proceedings of the Fourth Workshop on Treebanks and Linguistic Theories (TLT2005), Barcelona, Spain, December. find
  11. Alessandro Moschitti. 2004. A study on convolution kernels for shallow semantic parsing. In Proceedings of the 42nd Annual Meeting of the Association for Computational Linguistics (ACL 2004), Barcelona, Spain. find
  12. PDTB-Group, 2007. The Penn Discourse Treebank 2.0 Annotation Manual. The PDTB Research Group, December. find our card
  13. Emily Pitler, Mridhula Raghupathy, Hena Mehta, Ani Nenkova, Alan Lee, and Aravind Joshi. 2008. Easily identifiable discourse relations. In Proceedings of the 22nd International Conference on Computational Linguistics (COLING 2008), Manchester, UK, August. find
  14. Emily Pitler, Annie Louis, and Ani Nenkova. 2009. Automatic sense prediction for implicit discourse relations in text. To appear in Proceedings of the Joint Conference of the 47th Annual Meeting of the Association for Computational Linguistics and the 4th International Joint Conference on Natural Language Processing of the Asian Federation of Natural Language Processing (ACL-IJCNLP 2009). find our card
  15. Rashmi Prasad, Nikhil Dinesh, Alan Lee, Eleni Miltsakaki, Livio Robaldo, Aravind Joshi, and Bonnie Webber. 2008. The Penn Discourse Treebank 2.0. In Proceedings of the 6th International Conference on Language Resources and Evaluation (LREC 2008). find our card
  16. Manami Saito, Kazuhide Yamamoto, and Satoshi Sekine. 2006. Using phrasal patterns to identify discourse relations. In Proceedings of the Human Language Technology Conference of the North American Chapter of the Association for Computational Linguistics (HLT-NAACL 2006), pages 133– 136, New York, USA, June. find
  17. Bonnie Webber. 2004. D-LTAG: Extending lexicalized TAG to discourse. Cognitive Science, 28(5):751–779, September. find
  18. Ben Wellner, James Pustejovsky, Catherine Havasi, Anna Rumshisky, and Roser Sauri. 2006. Classification of discourse coherence relations: An exploratory study using multiple knowledge sources. In Proceedings of the 7th SIGdial Workshop on Discourse and Dialogue, Sydney, Australia, July. find
  19. Florian Wolf and Edward Gibson. 2005. Representing discourse coherence: a corpus-based analysis. In Proceedings of the 20th International Conference on Computational Linguistics (COLING 2004), pages 134–140, Morristown, NJ, USA. find