Joint Syntacto-Discourse Parsing and the Syntacto-Discourse Treebank

What it is about

The paper proposes a discourse parser that builds constituency and RST discourse trees in one pass, with no pre-segmentation and no external syntactic features. The model is a span-based incremental parser with bi-directional LSTMs (building on Cross and Huang 2016). The authors also build PTB-RST, a joint treebank made by aligning RST Treebank trees with Penn Treebank trees and keeping only the common parts. It has 347 training trees (17,837 tokens) and 38 test trees (4,819 tokens), following the RST split. On the test set, end-to-end scores are: segmentation F1 95.4 (a segmentation-only system reports 95.1); tree structure 78.8 (against 72.3 for a pipeline system of Hernault et al. 2010); nuclearity 65.0 (against 59.1); relation labels 52.2 (against 47.3). The constituency F1 is 87.2.

Ideas

• I1. Align two existing treebanks by span, keeping the common parts, as the paper does. For Mova, align UD trees with EDU spans and record how much material is lost in the alignment. The loss is a number to report.
• I2. The model needs no pre-segmentation. As a deterministic baseline, cut EDUs at strong cues ("despite", "because of", as the paper's guidelines do), and measure segmentation F1 against the gold units.
• I3. Report discourse results at four levels (segmentation, structure, nuclearity, relation), as the paper does, for any Mova discourse component, so that a gain at one level cannot hide a loss at another.
• I4. Negative control: remove the syntactic tree from the training data and measure the drop. The paper's claim is that joint training helps, and this check tests that claim on Mova's own trees.

For Mova

The neural span parser is not for Mova; take the data design and the four-level report. The code is en/src/parse.rs (UD trees), with the chapter 36. Discourse relations and coherence and prag/src/roles.rs for the discourse roles. The measure is the four-level F1 on a held-out set, run twice, with the aligned-treebank loss reported.

Topics: Discourse analysis · Treebanks

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.18653/v1/d17-1225 · arXiv 1708.08484

@inproceedings{zhao2017joint,
  title = {{Joint Syntacto-Discourse Parsing and the Syntacto-Discourse Treebank}},
  author = {Zhao, Kai and Huang, Liang},
  booktitle = {Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing},
  year = {2017},
  doi = {10.18653/v1/d17-1225},
  eprint = {1708.08484},
  archivePrefix = {arXiv},
  url = {https://doi.org/10.18653/v1/d17-1225}
}

References

28 references, 3 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Ngo Xuan Bach, Nguyen Le Minh, and Akira Shimazu. 2012. A reranking model for discourse segmentation using subtree features. In Proceedings of the 13th Annual Meeting of the Special Interest Group on Discourse and Dialogue, pages 160–168. Association for Computational Linguistics. find
2. Parminder Bhatia, Yangfeng Ji, and Jacob Eisenstein. 2015. Better document-level sentiment analysis from rst discourse parsing. arXiv preprint arXiv:1509.01599. ↗
3. James Cross and Liang Huang. 2016. Incremental parsing with minimal features using bi-directional lstm. arXiv preprint arXiv:1606.06406. ↗
4. Iria da Cunha, Eric SanJuan, Juan-Manuel Torres-Moreno, M Teresa Cabré, and Gerardo Sierra. 2012. A symbolic approach for automatic detection of nuclearity and rhetorical relations among intra-sentence discourse segments in spanish. In International Conference on Intelligent Text Processing and Computational Linguistics, pages 462–474. Springer. find
5. Vanessa Wei Feng and Graeme Hirst. 2014. A linear-time bottom-up discourse parser with constraints and post-editing. In ACL (1), pages 511–521. find
6. David Ferrucci, Eric Brown, Jennifer Chu-Carroll, James Fan, David Gondek, Aditya A Kalyanpur, Adam Lally, J William Murdock, Eric Nyberg, John Prager, et al. 2010. Building watson: An overview of the deepqa project. AI magazine, 31(3):59–79. find
7. Yoav Goldberg and Joakim Nivre. 2013. Training deterministic parsers with non-deterministic oracles. Transactions of the association for Computational Linguistics, 1:403–414. find
8. Michael Heilman and Kenji Sagae. 2015. Fast rhetorical structure theory discourse parsing. arXiv preprint arXiv:1505.02425. ↗
9. Hugo Hernault, Helmut Prendinger, David A DuVerle, Mitsuru Ishizuka, and Tim Paek. 2010. Hilda: a discourse parser using support vector machine classification. Dialogue and Discourse, 1(3):1–33. find
10. Peter Jansen, Mihai Surdeanu, and Peter Clark. 2014. Discourse complements lexical semantics for non-factoid answer reranking. In ACL (1), pages 977– 986. find
11. Yangfeng Ji and Jacob Eisenstein. 2014. Representation learning for text-level discourse parsing. In ACL (1), pages 13–24. find
12. Shafiq Joty and Alessandro Moschitti. 2014. Discriminative reranking of discourse parses using tree kernels. a) A, 4(5):6. find
13. Shafiq R Joty, Giuseppe Carenini, Raymond T Ng, and Yashar Mehdad. 2013. Combining intra-and multi-sentential rhetorical parsing for document-level discourse analysis. In ACL (1), pages 486–496. find
14. Jiwei Li and Dan Jurafsky. 2016. Neural net models for open-domain discourse coherence. arXiv preprint arXiv:1606.01545. ↗
15. Jiwei Li, Rumeng Li, and Eduard H Hovy. 2014a. Recursive deep models for discourse parsing. In EMNLP, pages 2061–2069. find
16. Sujian Li, Liang Wang, Ziqiang Cao, and Wenjie Li. 2014b. Text-level discourse dependency parsing. In ACL (1), pages 25–35. find
17. Annie Louis, Aravind Joshi, and Ani Nenkova. 2010. Discourse indicators for content selection in summarization. In Proceedings of the 11th Annual Meeting of the Special Interest Group on Discourse and Dialogue, pages 147–156. Association for Computational Linguistics. find
18. William C Mann and Sandra A Thompson. 1988. Rhetorical structure theory: Toward a functional theory of text organization. Text-Interdisciplinary Journal for the Study of Discourse, 8(3):243–281. find our card
19. Daniel Marcu. 2000a. The rhetorical parsing of unrestricted texts: A surface-based approach. Computational linguistics, 26(3):395–448. find
20. Daniel Marcu. 2000b. The theory and practice of discourse parsing and summarization. MIT press. find
21. Mitchell P Marcus, Mary Ann Marcinkiewicz, and Beatrice Santorini. 1993. Building a large annotated corpus of english: The penn treebank. Computational linguistics, 19(2):313–330. find our card
22. Thiago Alexandre Salgueiro Pardo and Maria das Graças Volpe Nunes. 2008. On the development and evaluation of a brazilian portuguese discourse parser. Revista de Informática Teórica e Aplicada, 15(2):43–64. find
23. Swapna Somasundaran, Galileo Namata, Janyce Wiebe, and Lise Getoor. 2009. Supervised and unsupervised methods in employing discourse relations for improving opinion polarity classification. In Proceedings of the 2009 Conference on Empirical Methods in Natural Language Processing: Volume 1-Volume 1, pages 170–179. Association for Computational Linguistics. find
24. 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. Association for Computational Linguistics. find our card
25. ETS Tetreault et al. 2013. Holistic discourse coherence annotation for noisy essay writing. find
26. Kimberly Voll and Maite Taboada. 2007. Not all words are created equal: Extracting semantic orientation as a function of adjective relevance. In Australasian Joint Conference on Artificial Intelligence, pages 337–346. Springer. find
27. Yizhong Wang, Sujian Li, and Houfeng Wang. 2017. A two-stage parsing method for text-level discourse analysis. In Proceedings of ACL. find
28. Yasuhisa Yoshida, Jun Suzuki, Tsutomu Hirao, and Masaaki Nagata. 2014. Dependency-based discourse parser for single-document summarization. In EMNLP, pages 1834–1839. Citeseer. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/zhao-2017-joint-syntacto-discourse-parsing-syntacto/