Elamur Research⌕
☰
home / papers / 2017 / Cross-lingual RST Discourse Parsing

Cross-lingual RST Discourse Parsing

Summary

RST discourse parsing has been studied almost entirely on the English RST Discourse Treebank. Treebanks for other languages are small and have mostly been handled with rule-based or sentence-internal systems. The authors build a neural parser that predicts trees with a stack. Instead of encoding every word of each discourse unit, it uses fewer features and draws on wider structural context, namely the left and right children of complex units on the stack. They also harmonize six RST corpora (English, Brazilian Portuguese, Spanish, German, Dutch, Basque) into one binarized, cleaned format. This lets models be trained on some languages and applied to others, with lexical information transferred through cross-lingual word embeddings or bilingual dictionaries. For English, the parser matches or beats the state of the art on Span and Nuclearity. The work also delivers the first end-to-end statistical RST parsers for the five non-English languages and the first cross-lingual results. These show that a usable parser can be built for a language with little or no annotated data, which is directly relevant to low-resource settings.

Key points

  • Harmonized corpora:
    • En-DT: RST-DT, 385 WSJ documents.
    • Pt-DT: merges CST-News, Summ-it, Rhetalho and CorpusTCC.
    • Es-DT: the Spanish RST DT.
    • De-DT: the MAZ part of the Potsdam Commentary Corpus.
    • Nl-DT and Eu-DT: 80 and 88 documents, too small for a training split.
  • English monolingual results: 85.0 Span, 72.3 Nuclearity and 60.1 Relation. This compares with 85.0, 70.8 and 58.6 for Li et al. (2014), and 82.1, 71.1 and 61.6 for DPLP.
  • Source plus target data: adding target-language data (development set) to the other-language training data gives the best English system, at 85.1, 73.1 and 61.4.
  • Training only on other languages: unlabeled structure remains reasonable (about 70% or higher), but relation scores collapse. Part of the cause is inconsistent label sets; for example, only En-DT and Pt-DT use ATTRIBUTION, and English F1 for that relation drops from 93% to 30%.
  • Dutch and Basque, no training data: cross-lingual parsers beat the right-branching, most-frequent-relation baseline by about 11–13 points on Span, 8–15 on Nuclearity and 6–11 on Relation.
  • Word embeddings: monolingual embeddings lower Relation by 1.1–4.2 points, and bilingual dictionaries usually outperform embeddings for transfer.
  • Availability: code and preprocessing scripts are published at bitbucket.org/chloebt/discourse.

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

47 papers of our library cite it (of the 140673 whose bibliographies we read); 20 of them have a page here, 27 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. arXiv 1701.02946

@inproceedings{braud2017cross,
  title = {{Cross-lingual RST Discourse Parsing}},
  author = {Braud, Chloé and Coavoux, Maximin and Søgaard, Anders},
  booktitle = {Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 1, Long Papers},
  year = {2017},
  eprint = {1701.02946},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/1701.02946}
}

References

56 references, 17 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Farah Benamara and Maite Taboada. 2015. Mapping different rhetorical relation annotations: A proposal. In Proceedings of Starsem. find
  2. Parminder Bhatia, Yangfeng Ji, and Jacob Eisenstein. 2015. Better document-level sentiment analysis from RST discourse parsing. In Proceedings of EMNLP. find our card
  3. Chloé Braud and Pascal Denis. 2015. Comparing word representations for implicit discourse relation classification. In Proceedings of EMNLP. find our card
  4. Jill Burstein, Daniel Marcu, and Kevin Knight. 2003. Finding the write stuff: automatic identification of discourse structure in student essays. IEEE Intelligent Systems: Special Issue on Advances in Natural Language Processing, 18. \#\# Conclusion Paula C.F. find
  5. Cardoso, Erick G. Maziero, Mara Luca Castro Jorge, Eloize R.M. Seno, Ariani Di Felippo, Lucia Helena Machado Rino, Maria das Gracas Volpe Nunes, and Thiago A. S. Pardo. 2011. CSTNews - a discourse-annotated corpus for single and multi-document summarization of news texts in Brazilian Portuguese. In… find
  6. We introduced a new discourse parser that obtains state-of-the-art performance for English. We harmonized discourse treebanks for several languages, enabling us to present results for five other languages for which available corpora are smaller, including the first cross-lingual discourse parsing Ly… find
  7. Derrick Higgins, Jill Burstein, Daniel Marcu, and Claudia Gentile. 2004. Evaluating multiple aspects of coherence in student essays. In Proceedings of HLT-NAACL. find
  8. Lynn Carlson, Daniel Marcu, and Mary Ellen Okurowski. 2001. Building a discourse-tagged corpus in the framework of Rhetorical Structure Theory. In Proceedings of the Second SIGdial Workshop on Discourse and Dialogue. find our card
  9. Sepp Hochreiter and Jürgen Schmidhuber. 1997. Long short-term memory. Neural Computation, 9(8):1735–1780. find
  10. Mikel Iruskieta, Marı́a J. Aranzabe, Arantza Diaz de Ilarraza, Itziar Gonzalez-Dios, Mikel Lersundi, and Oier Lopez de la Calle. 2013. The RST Basque Treebank: an online search interface to check rhetorical relations. In Proceedings of the 4th Workshop RST and Discourse Studies. find
  11. Danqi Chen and Christopher D Manning. 2014. A fast and accurate dependency parser using neural networks. In Empirical Methods in Natural Language Processing (EMNLP). find our card
  12. Maximin Coavoux and Benoit Crabbé. 2016. Neural greedy constituent parsing with dynamic oracles. In Proceedings of the 54th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 172–182, Berlin, Germany, August. Association for Computational Linguistics. find
  13. Mikel Iruskieta, Iria da Cunha, and Maite Taboada. 2015. A qualitative comparison method for rhetorical structures: identifying different discourse structures in multilingual corpora. In Proceedings of LREC. find
  14. Yangfeng Ji and Jacob Eisenstein. 2014. Representation learning for text-level discourse parsing. In Proceedings of ACL. find our card
  15. Sandra Collovini, Thiago I Carbonel, Juliana Thiesen Fuchs, Jorge César Coelho, Lúcia Rino, and Renata Vieira. 2007. Summ-it: Um corpus anotado com informaçoes discursivas visandoa sumarizaçao au-tomática. Proceedings of TIL. find
  16. Shafiq R. Joty, Giuseppe Carenini, and Raymond T. Ng. 2012. A novel discriminative framework for sentence-level discourse analysis. In Proceedings of EMNLP. find
  17. Iria da Cunha, Eric SanJuan, Juan-Manuel Torres-Moreno, Marina Lloberas, and Irene Castellón. 2010. DiSeg: Un segmentador discursivo au-tomático para el español. Procesamiento del lenguaje natural, 45:145–152. find
  18. 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 Proceedings of ACL. find
  19. Iria da Cunha, Juan-Manuel Torres-Moreno, and Gerardo Sierra. 2011. On the development of the RST Spanish Treebank. In Proceedings of the Fifth Linguistic Annotation Workshop, LAW. find
  20. Omer Levy, Anders Søgaard, and Yoav Goldberg. 2017. A strong baseline for learning cross-lingual word embeddings from sentence alignments. In Proceedings of EACL. find
  21. Iria da Cunha, Eric SanJuan, Juan-Manuel Torres-Moreno, Marina Lloberes, and Irene Castellón. 2012. DiSeg 1.0: The first system for Spanish discourse segmentation. Expert Syst. Appl., 39(2):1671–1678. find
  22. Jiwei Li, Rumeng Li, and Eduard H. Hovy. 2014. Recursive deep models for discourse parsing. In Proceedings of EMNLP. find our card
  23. Hal Daumé III and Daniel Marcu. 2009. A noisy-channel model for document compression. In Proceedings of ACL. find
  24. Annie Louis, Aravind Joshi, and Ani Nenkova. 2010. Discourse indicators for content selection in summarization. In Proceedings of SIGDIAL. find
  25. Vanessa Wei Feng and Graeme Hirst. 2012. Text-level discourse parsing with rich linguistic features. In Proceedings of ACL. find our card
  26. William C. Mann and Sandra A. Thompson. 1988. Rhetorical Structure Theory: Toward a functional theory of text organization. Text, 8:243–281. find our card
  27. Vanessa Wei Feng and Graeme Hirst. 2014. A linear-time bottom-up discourse parser with constraints and post-editing. In Proceedings of ACL. find our card
  28. Daniel Marcu. 1997. From discourse structures to text summaries. In Proceedings of the ACL Workshop on Intelligent Scalable Text Summarization, pages 82– 88. find
  29. Jiang Guo, Wanxiang Che, David Yarowsky, Haifeng Wang, and Ting Liu. 2015. Cross-lingual dependency parsing based on distributed representations. In Proceedings of ACL-IJCNLP. find
  30. Daniel Marcu. 2000a. The rhetorical parsing of unrestricted texts: A surface-based approach. Computational Linguistics. find our card
  31. Hugo Hernault, Helmut Prendinger, David A. duVerle, and Mitsuru Ishizuka. 2010. HILDA: A discourse parser using support vector machine classification. Dialogue and Discourse, 1:1–33. find our card
  32. Daniel Marcu. 2000b. The Theory and Practice of Discourse Parsing and Summarization. MIT Press, Cambridge, MA, USA. cientı́ficos em Português do Brasil e sua marcação retórica. Technical report, Technical Report. find
  33. Erick G. Maziero, Thiago A. S. Pardo, Iria da Cunha, Juan-Manuel Torres-Moreno, and Eric SanJuan. 2011. DiZer 2.0-an adaptable on-line discourse parser. In Proceedings of 3rd RST Brazilian Meeting, pages 1–17. find
  34. Thiago A. S. Pardo and Maria das Graças Volpe Nunes. 2004. Relações retóricas e seus marcadores superfi-ciais: Análise de um corpus de textos cientı́ficos em Português do Brasil. Relatório Técnico NILC. find
  35. Erick G. Maziero, Graeme Hirst, and Thiago A. S. Pardo. 2015. Adaptation of discourse parsing models for Portuguese language. In Proceedings of the Brazilian Conference on Intelligent Systems (BRACIS). find
  36. Thiago A. S. 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
  37. Ryan McDonald, Slav Petrov, and Keith Hall. 2011. Multi-source transfer of delexicalized dependency parsers. In Proceedings of EMNLP. find our card
  38. Thiago A. S. Pardo and Eloize R. M. Seno. 2005. Rhetalho: Um corpus de referłncia anotado retori-camente. In Proceedings of Encontro de Corpora. find
  39. Joakim Nivre, Željko Agić, Lars Ahrenberg, Maria Jesus Aranzabe, Masayuki Asahara, Aitziber Atutxa, Miguel Ballesteros, John Bauer, Kepa Bengoetxea, Yevgeni Berzak, Riyaz Ahmad Bhat, Cristina Bosco, Gosse Bouma, Sam Bowman, Gülşen Cebirolu Eryiit, Giuseppe G. A. Celano, Çar Çöltekin, Miriam C… find
  40. Emily Pitler, Annie Louis, and Ani Nenkova. 2009. Automatic sense prediction for implicit discourse relations in text. In Proceedings of ACL-IJCNLP. find our card
  41. Boris T. Polyak and Anatoli B. Juditsky. 1992. Acceleration of stochastic approximation by averaging. SIAM J. Control Optim., 30(4):838–855, July. find
  42. Gisela Redeker, Ildik Berzlnovich, Nynke van der Vliet, Gosse Bouma, and Markus Egg. 2012. Multilayer discourse annotation of a dutch text corpus. In Proceedings of LREC. find
  43. Charlotte Roze. 2013. Vers une algèbre des relations de discours. Ph.D. thesis, Université Paris-Diderot. find
  44. Attapol Rutherford and Nianwen Xue. 2014. Discovering implicit discourse relations through brown cluster pair representation and coreference patterns. In Proceedings of EACL. find our card
  45. Kenji Sagae and Alon Lavie. 2005. A classifier-based parser with linear run-time complexity. In Proceedings of the Ninth International Workshop on Parsing Technology, pages 125–132. Association for Computational Linguistics. find our card
  46. Kenji Sagae. 2009. Analysis of discourse structure with syntactic dependencies and data-driven shift-reduce parsing. In Proceedings of IWPT 2009. find
  47. Radu Soricut and Daniel Marcu. 2003. Sentence level discourse parsing using syntactic and lexical information. In Proceedings of NAACL. find our card
  48. Caroline Sporleder and Mirella Lapata. 2005. Discourse chunking and its application to sentence compression. In Proceedings of HLT/EMNLP. find
  49. Manfred Stede and Arne Neumann. 2014. Potsdam commentary corpus 2.0: Annotation for discourse research. In Proceedings of LREC. find
  50. Manfred Stede. 2004. The potsdam commentary corpus. In Proceedings of the ACL Workshop on Discourse Annotation. find
  51. Thiago A. S. Pardo and Maria das Graças Volpe Nunes. 2003. A construção de um corpus de textos C N Subalalitha and Ranjani Parthasarathi. 2012. An approach to discourse parsing using sangati and Rhetorical Structure Theory. In Proceedings of the Workshop on Machine Translation and Parsing in Indi… find
  52. Rajen Subba and Barbara Di Eugenio. 2009. An effective discourse parser that uses rich linguistic information. In Proceedings of ACL-HLT. find our card
  53. Maite Taboada and William C. Mann. 2006. Applications of rhetorical structure theory. Discourse Studies, 8:567–588. find
  54. Gian Lorenzo Thione, Martin Van den Berg, Livia Polanyi, and Chris Culy. 2004. Hybrid text summarization: Combining external relevance measures with structural analysis. In Proceedings of the ACL Workshop Text Summarization Branches Out. find
  55. Nynke Van Der Vliet, Ildikó Berzlnovich, Gosse Bouma, Markus Egg, and Gisela Redeker. 2011. Building a discourse-annotated Dutch text corpus. In S. Dipper and H. Zinsmeister (Eds.), Beyond Semantics, Bochumer Linguistische Arbeitsberichte 3, pages 157–171. find
  56. Yunfang Wu, Fuqiang Wan, Yifeng Xu, and Xueqiang Lü. 2016. A new ranking method for Chinese discourse tree building. Acta Scientiarum Naturalium Universitatis Pekinensis, 52(1):65–74. find

↑ back to top

This page in Markdown →