SciDTB: Discourse Dependency TreeBank for Scientific Abstracts

What it is about

The paper builds SciDTB, a discourse treebank of 798 ACL Anthology abstracts, annotated with dependency trees instead of RST constituency trees. Each elementary discourse unit (EDU) has one head and a labelled relation to its head, so crossing (non-projective) dependencies are possible. The relation set has 17 coarse and 26 fine types, and clauses are the default EDUs. EDUs were segmented by an automatic segmenter and then checked by hand; a second annotator changed only 28 of 2,772 EDUs. Between annotator pairs, the unlabelled attachment score (UAS) is about 0.75 to 0.81 and the labelled score (LAS) about 0.55 to 0.64. There are 39 non-projective trees (about 3%). Most relations (61.6%) join neighbouring EDUs. The paper's text gives 18,978 relations, while its size table lists 17,158. Three baseline parsers were trained on 492 abstracts. On the 152-abstract test set they score UAS 0.702 and LAS 0.535 (transition-based), and UAS 0.576 and LAS 0.425 (graph-based). Human agreement on the same test set is UAS 0.802 and LAS 0.622.

Ideas

• I1. Use the attachment-distance table (61.6% adjacent, about 75% within one intermediate EDU) as a prior in a rule-based discourse attacher for Mova, and measure the gain on open texts.
• I2. Use the 17 coarse and 26 fine relation types as a checklist for 36. Discourse relations and coherence. List the relations Mova cannot yet express, and mark the scientific ones (Background subtypes, Evidence, Goal) first.
• I3. Report discourse parsing as UAS and LAS, with the human agreement pair kept in the same table, so that the ceiling is visible.
• I4. Treat a unit split by a relative clause as one EDU with a same-unit link. This rule can be tested on the clause segmenter output of en/src/parse.rs.

For Mova

The chapters are 36. Discourse relations and coherence and 37. Register and variation (abstracts are a register). The code is en/src/parse.rs. The measure is UAS and LAS on a held-out set of open abstracts, run twice. The data are CC BY 4.0, so they can be used for building and testing.

Topics: Discourse analysis · Treebanks

Cite

The paper itself, not this card. BibTeX from the library's record of the work. DOI 10.18653/v1/p18-2071

@inproceedings{yang2018scidtb,
  title = {{SciDTB: Discourse Dependency TreeBank for Scientific Abstracts}},
  author = {Yang, An and Li, Sujian},
  booktitle = {Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers)},
  year = {2018},
  doi = {10.18653/v1/p18-2071},
  url = {https://doi.org/10.18653/v1/p18-2071}
}

References

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

1. 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
2. Steven Bird and Edward Loper. 2004. Nltk: the natural language toolkit. In Proceedings of the ACL 2004 on Interactive poster and demonstration sessions. Association for Computational Linguistics, page 31. find
3. Mathieu Morey, Philippe Muller, and Nicholas Asher. 2017. How much progress have we made on rst discourse parsing? a replication study of recent results on the rst-dt. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing. pages 1319–1324. find our card
4. Alena Böhmová, Jan Hajič, Eva Hajičová, and Barbora Hladká. 2003. The prague dependency treebank. In Treebanks, Springer, pages 103–127. find
5. Harry Bunt and Rashmi Prasad. 2016. Iso dr-core (iso 24617-8): Core concepts for the annotation of discourse relations. In Proceedings 12th Joint ACL-ISO Workshop on Interoperable Semantic Annotation (ISA-12). pages 45–54. find
6. Joakim Nivre. 2003. An efficient algorithm for projective dependency parsing. In Proceedings of the 8th International Workshop on Parsing Technologies (IWPT. Citeseer. find our card
7. Lynn Carlson and Daniel Marcu. 2001. Discourse tagging reference manual. ISI Technical Report ISI-TR-545 54:56. find
8. Joakim Nivre, Johan Hall, and Jens Nilsson. 2004. Memory-based dependency parsing. In Proceedings of the Eighth Conference on Computational Natural Language Learning (CoNLL-2004) at HLT-NAACL 2004. find
9. Lynn Carlson, Daniel Marcu, and Mary Ellen Okurowski. 2003. Building a discourse-tagged corpus in the framework of rhetorical structure theory. In Current and new directions in discourse and dialogue, Springer, pages 85–112. find our card
10. Livia Polanyi. 1988. A formal model of the structure of discourse. Journal of pragmatics 12(5-6):601–638. find
11. Rashmi Prasad, Susan McRoy, Nadya Frid, Aravind Joshi, and Hong Yu. 2011. The biomedical discourse relation bank. BMC bioinformatics 12(1):188. find
12. Eugene Charniak. 2000. A maximum-entropy-inspired parser. In Proceedings of the 1st North American chapter of the Association for Computational Linguistics conference. Association for Computational Linguistics, pages 132–139. find
13. Rashmi Prasad, Eleni Miltsakaki, Nikhil Dinesh, Alan Lee, Aravind Joshi, Livio Robaldo, and Bonnie L Webber. 2007. The penn discourse treebank 2.0 annotation manual . find our card
14. Shima Gerani, Yashar Mehdad, Giuseppe Carenini, Raymond T Ng, and Bita Nejat. 2014. Abstractive summarization of product reviews using discourse structure. In EMNLP. volume 14, pages 1602–1613. find
15. 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. Association for Computational Linguistics, pages 149–156. find our card
16. Francisco Guzmán, Shafiq Joty, Lluı́s Màrquez, and Preslav Nakov. 2014. Using discourse structure improves machine translation evaluation. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). volume 1, pages 687–698. find
17. Richard A Hudson. 1984. Word grammar. Blackwell Oxford. find
18. Manfred Stede, Stergos D Afantenos, Andreas Peldszus, Nicholas Asher, and Jérémy Perret. 2016. Parallel discourse annotations on a corpus of short texts. In LREC. find
19. Matthias Irmer. 2011. Bridging inferences: Constraining and resolving underspecification in discourse interpretation, volume 11. Walter de Gruyter. find
20. Yizhong Wang, Sujian Li, and Houfeng Wang. 2017. A two-stage parsing method for text-level discourse analysis. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 2: Short Papers). volume 2, pages 184–188. find
21. Shafiq Joty, Francisco Guzmán, Lluı́s Màrquez, and Preslav Nakov. 2014. Discotk: Using discourse structure for machine translation evaluation. In Proceedings of the Ninth Workshop on Statistical Machine Translation. pages 402–408. find
22. Florian Wolf and Edward Gibson. 2005. Representing discourse coherence: A corpus-based study. Computational Linguistics 31(2):249–287. find
23. Alex Lascarides and Nicholas Asher. 2008. Segmented discourse representation theory: Dynamic semantics with discourse structure. In Computing meaning, Springer, pages 87–124. find
24. Yasuhisa Yoshida, Jun Suzuki, Tsutomu Hirao, and Masaaki Nagata. 2014. Dependency-based discourse parser for single-document summarization. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP). pages 1834–1839. find
25. Sujian Li, Liang Wang, Ziqiang Cao, and Wenjie Li. 2014. Text-level discourse dependency parsing. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). volume 1, pages 25–35. find

This page in Markdown → https://md.mova.elamur.ai/free/en/docs/papers/yang-2018-scidtb-discourse-dependency-treebank-scientific/