Elamur Research
☰
home / papers / 2012 / Text-level Discourse Parsing with Rich Linguistic Features

Text-level Discourse Parsing with Rich Linguistic Features

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/P12-1007.pdf

Summary

Most RST-style discourse parsers work within single sentences, but coherence across a whole document cannot be inferred from locally coherent sentences alone. The authors therefore build a text-level parser on top of HILDA, a bottom-up system. In HILDA, a binary Structure classifier decides which adjacent spans to merge, and a Relation classifier then labels the merged span. The authors keep HILDA's segmentation and focus on tree building. They enrich its feature set with their own linguistic features and with features adapted from Lin et al. (2009), applying feature selection to keep the dimensionality manageable. Experiments on RST-DT compare three conditions: within-sentence, cross-sentence and combined spans. The enriched features clearly outperform HILDA's for both classifiers, and contextual features prove especially helpful. The same features, however, perform consistently worse across sentence boundaries than inside sentences. This shows that a sentence-level parser cannot simply be scaled up to whole texts, and it identifies cross-sentence relations as the main obstacle for document-level discourse analysis.

Key points

  • Data: RST Discourse Treebank (RST-DT), 385 Wall Street Journal documents split 347/38 for training and testing. Relation labeling uses its 18 coarse relation classes, ignoring nuclearity.
  • Structure classification on all spans reaches F1 89.51% with the full feature set and 69.84% without contextual features (NC), against 67.13% for HILDA's features. The cross-sentence F1 is far lower, 55.87% versus 38.56% for HILDA. Gains over HILDA are significant (Wilcoxon, p < .01).
  • Relation classification uses LibSVM with a linear kernel and is evaluated with macro-averaged F-score (MAFS) to avoid bias toward frequent classes such as ELABORATION:
    • All spans: 0.440 (Full) vs 0.379 (HILDA).
    • Within-sentence: 0.490 vs 0.446.
    • Cross-sentence: 0.194 vs 0.127.
  • Contextual features help under every condition, but they are reliable only when the surrounding structure is correct. The authors therefore propose iterative re-estimation of the tree and its labels as a direction for future work.
  • HILDA's existing EDU segmentation (F-score 93.8%) is treated as essentially solved, so the work concentrates entirely on the tree-building stage.

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

Cited by

65 papers of our library cite it (of the 140673 whose bibliographies we read); 4 of them have a page here, 61 not yet.

Cite

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

@inproceedings{feng2012text,
  title = {{Text-level Discourse Parsing with Rich Linguistic Features}},
  author = {Feng, Vanessa Wei and Hirst, Graeme},
  booktitle = {Proceedings of the 50th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2012},
  url = {https://aclanthology.org/P12-1007/}
}

References

21 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Jason Baldridge and Alex Lascarides. 2005. Probabilistic head-driven parsing for discourse structure. In Proceedings of the Ninth Conference on Computational Natural Language Learning, pages 96–103. find
  2. Steven Bird, Ewan Klein, and Edward Loper. 2009. Natural Language Processing with Python — Analyzing Text with the Natural Language Toolkit. O’Reilly. find
  3. Lynn Carlson, Daniel Marcu, and Mary Ellen Okurowski. 2001. Building a discourse-tagged corpus in the framework of Rhetorical Structure Theory. In Proceedings of Second SIGdial Workshop on Discourse and Dialogue, pages 1–10. find our card
  4. Chih-Chung Chang and Chih-Jen Lin. 2011. LIBSVM: A library for support vector machines. ACM Transactions on Intelligent Systems and Technology, 2:1–27. find
  5. David A. duVerle and Helmut Prendinger. 2009. A novel discourse parser based on Support Vector Machine classification. In Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the 4th International Joint Conference on Natural Language Processing of the AFNLP, Volume 2, ACL ’09, pages 665–673, Stroudsburg, PA, USA. Association for Computational Linguistics. find
  6. Hugo Hernault, Danushka Bollegala, and Mitsuru Ishizuka. 2010a. A semi-supervised approach to improve classification of infrequent discourse relations using feature vector extension. In Proceedings of the 2010 Conference on Empirical Methods in Natural Language Processing, pages 399–409, Cambridge, MA, October. Association for Computational Linguistics. find
  7. Hugo Hernault, Helmut Prendinger, David A. duVerle, and Mitsuru Ishizuka. 2010b. HILDA: A discourse parser using support vector machine classification. Dialogue and Discourse, 1(3):1–33. find our card
  8. Thorsten Joachims. 2005. A support vector method for multivariate performance measures. In International Conference on Machine Learning (ICML), pages 377– 384. find
  9. Alistair Knott and Robert Dale. 1994. Using linguistic phenomena to motivate a set of coherence relations. Discourse Processes, 18(1). find
  10. Huong LeThanh, Geetha Abeysinghe, and Christian Huyck. 2004. Generating discourse structures for written texts. In Proceedings of the 20th International Conference on Computational Linguistics, pages 329– 335. find
  11. Ziheng Lin, Min-Yen Kan, and Hwee Tou Ng. 2009. Recognizing implicit discourse relations in the Penn Discourse Treebank. In Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing, Volume 1, EMNLP ’09, pages 343–351. find our card
  12. Ziheng Lin, Hwee Tou Ng, and Min-Yen Kan. 2010. A PDTB-styled end-to-end discourse parser. Technical report, School of Computing, National University of Singapore. find
  13. William Mann and Sandra Thompson. 1988. Rhetorical structure theory: Toward a functional theory of text organization. Text, 8(3):243–281. find our card
  14. Daniel Marcu and Abdessamad Echihabi. 2002. An unsupervised approach to recognizing discourse relations. In Proceedings of 40th Annual Meeting of the Association for Computational Linguistics, pages 368–375, Philadelphia, Pennsylvania, USA, July. Association for Computational Linguistics. find our card
  15. Daniel Marcu. 1997. The rhetorical parsing of natural language texts. In Proceedings of the 35th Annual Meeting of the Association for Computational Linguistics, pages 96–103. find
  16. 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
  17. David Reitter. 2003. Simple signals for complex rhetorics: On rhetorical analysis with rich-feature support vector models. LDV Forum, 18(1/2):38–52. find
  18. Kenji Sagae. 2009. Analysis of discourse structure with syntactic dependencies and data-driven shift-reduce parsing. In Proceedings of the 11th International Conference on Parsing Technologies, pages 81–84. find
  19. Radu Soricut and Daniel Marcu. 2003. Sentence level discourse parsing using syntactic and lexical information. In Proceedings of the 2003 Conference of the North American Chapter of the Association for Computational Linguistics on Human Language Technology, Volume 1, pages 149–156. find our card
  20. Rajen Subba and Barbara Di Eugenio. 2009. An effective discourse parser that uses rich linguistic information. In Proceedings of Human Language Technologies: The 2009 Annual Conference of the North American Chapter of the Association for Computational Linguistics, pages 566–574. find
  21. Bonnie Webber. 2004. D-LTAG: Extending lexicalized TAG to discourse. Cognitive Science, 28(5):751–779. find