Elamur Research⌕
☰
home / papers / 2019 / EusDisParser: improving an under-resourced discourse parser with cross-lingual data

EusDisParser: improving an under-resourced discourse parser with cross-lingual data

Summary

Iruskieta and Braud build the first RST-based discourse parser for Basque. Its training material, the Basque RST DT, has fewer than a hundred documents. They reuse the transition-based shift-reduce parser of Braud et al. (2017) and compare two kinds of systems:

  • Monolingual systems use Basque data only, with hyper-parameters tuned by cross-validation and with random or pre-trained Basque embeddings.
  • Cross-lingual systems learn from the English, Spanish and Portuguese RST treebanks, either alone ("Src Only") or combined with Basque data ("Src+Tgt"). Each language is aligned to Basque with bilingual embeddings built through VecMap.

The best monolingual model scores 34.78 F1 on the full discourse structure. This beats the earlier 29.5 obtained by Braud et al. (2017), which relied solely on other languages' data. Adding foreign treebanks did not surpass the Basque-only parser.

A qualitative analysis shows that the parser leans on a few general, uninformative relations such as ELABORATION. Human annotators, by contrast, agree on more specific relations. The authors conclude that annotating more Basque data is essential. They see the tool as a basis for summarisation, sentiment analysis and pre-annotation.

Key points

  • Data: The Basque RST DT contains specialised abstracts (medicine, terminology, science) and opinion texts labelled with 31 relations; 38 documents form the test set. The cross-lingual sources are the English RST DT (385 docs), the Spanish RST DT and a merged Portuguese corpus (329 docs).
  • Embeddings: The monolingual systems use word2vec skip-gram vectors (350 dimensions) trained on the Elhuyar Web Corpus, as well as FastText vectors (300 dimensions). Cross-lingual mappings were built with VecMap, using identical words as a seed dictionary.
  • Result: The best score is 34.78 F1 for the full structure, ahead of the prior 29.5. The cross-lingual setups fell short of the monolingual system.
  • Error analysis: JOINT was predicted 192 times with only one correct match, and LIST was never matched. The cross-lingual model trained on Portuguese got essentially just ELABORATION right.
  • Code: The code is released at gitlab.inria.fr/andiamo/eusdisparser.

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

5 papers of our library cite it (of the 140673 whose bibliographies we read); 3 of them have a page here, 2 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/w19-2709

@inproceedings{iruskieta2019eusdisparser,
  title = {{EusDisParser: improving an under-resourced discourse parser with cross-lingual data}},
  author = {Iruskieta, Mikel and Braud, Chloé},
  booktitle = {Proceedings of the Workshop on Discourse Relation Parsing and Treebanking 2019},
  year = {2019},
  doi = {10.18653/v1/w19-2709},
  url = {https://doi.org/10.18653/v1/w19-2709}
}

References

48 references, 13 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Stergos Afantenos, Pascal Denis, Philippe Muller, and Laurence Danlos. 2010. Learning recursive segments for discourse parsing. arXiv preprint arXiv:1003.5372. find our card
  2. Stergos Afantenos, Eric Kow, Nicholas Asher, and Jérémy Perret. 2015. Discourse parsing for multi-party chat dialogues. Association for Computational Linguistics (ACL). find our card
  3. Jon Alkorta, Koldo Gojenola, Mikel Iruskieta, and Maite Taboada. 2017. Using lexical level information in discourse structures for Basque sentiment analysis. In Proceedings of the 6th Workshop on Recent Advances in RST and Related Formalisms, pages 39–47. find
  4. Mikel Artetxe, Gorka Labaka, and Eneko Agirre. 2018. A robust self-learning method for fully unsupervised cross-lingual mappings of word embeddings. In other annotation campaigns) ATTRIBUTION is not considered a rhetorical relation. find our card
  5. Iria da Cunha and Mikel Iruskieta. 2010. Comparing rhetorical structures in different languages: The influence of translation strategies. Discourse Studies, 12(5):563–598. find
  6. Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers). find
  7. Nicholas Asher and Alex Lascarides. 2003. Logics of conversation. Cambridge University Press. find
  8. Iria da Cunha, Eric San Juan, Juan-Manuel Torres-Moreno, Marina Lloberese, and Irene Castelln. 2012. DiSeg 1.0: The first system for Spanish discourse segmentation. Expert Systems with Applications, 39(2):1671–1678. find
  9. Unai Atutxa, Mikel Iruskieta, Olatz Ansa, and Alejandro Molina. 2017. COMPRESS-EUS: I(ra)kasleen laburpenak lortzeko tresna. In EUDIA: Euskararen bariazioa eta bariazioaren irakaskuntza-III, pages 87–98. find
  10. Debopam Das, Manfred Stede, and Maite Taboada. 2017. The good, the bad, and the disagreement: Complex ground truth in rhetorical structure analysis. In Proceedings of the 6th Workshop on Recent Advances in RST and Related Formalisms, pages 11–19. find our card
  11. Kepa Bengoetxea, Aitziber Atutxa, and Mikel Iruskieta. 2017. Un detector de la unidad central de un texto basado en técnicas de aprendizaje automático en textos cientı́ficos para el euskera. Procesamiento del Lenguaje Natural, 58:37–44. find
  12. Edouard Grave, Piotr Bojanowski, Prakhar Gupta, Armand Joulin, and Tomas Mikolov. 2018. Learning word vectors for 157 languages. In Proceedings of the International Conference on Language Resources and Evaluation (LREC 2018). find our card
  13. Kepa Bengoetxea and Mikel Iruskieta. 2017. A Supervised Central Unit Detector for Spanish. Procesamiento del Lenguaje Natural, 60:29–36. find
  14. Chloé Braud, Maximin Coavoux, and Anders Søgaard. 2017. Cross-lingual RST Discourse Parsing. In Proceedings of the European Chapter of the Association for Computational Linguistics. find our card
  15. Jet Hoek and Sandrine Zufferey. 2015. Factors influencing the implicitation of discourse relations across languages. In Proceedings of the 11th Joint ACL-ISO Workshop on Interoperable Semantic Annotation (ISA-11). find our card
  16. Paula C.F. 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 Port… find
  17. Eduard Hovy. 2010. Annotation. a tutorial. In 48th Annual Meeting of the Association for Computational Linguistics. find
  18. 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 Workshop RST and Discourse Studies. find
  19. 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
  20. Mikel Iruskieta and Zapirain Beñat. 2015. EusE-duSeg: a Dependency-Based EDU Segmentation for Basque. In Procesamiento del Lenguaje Natural, 55 41-48. Consultado en http://rua.ua.es/dspace/handle/10045/49274 ISBN (edicion digital): 978-84-608-1989-9. ↗
  21. Maximin Coavoux and Benoit Crabbé. 2016. Neural greedy constituent parsing with dynamic oracles. In Proceedings of the Annual Meeting of the Association for Computational Linguistics. find
  22. 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
  23. Mikel Iruskieta, Iria da Cunha, and Taboada Maite. 2015. A qualitative comparison method for rhetorical structures: Identifying different discourse structures in multilingual corpora. Language Resources and Evaluation vol. 49: 263-309. find
  24. Iria da Cunha, Juan Manuel Torres Moreno, and Gerardo Sierra. 2011. On the development of the RST Spanish Treebank. In Linguistic Annotation Workshop. find
  25. Mikel Iruskieta, Arantza Dı́az de Ilarraza, and Mikel Lersundi. 2014. The annotation of the central unit in rhetorical structure trees: A key step in annotating rhetorical relations. In Proceedings of COLING 2014, the 25th International Conference on Computational Linguistics: Technical Papers, page… find
  26. 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 Co… find
  27. Bernd Kortmann. 1991. Free adjuncts and absolutes in English: Problems of control and interpretation. Psychology Press, New York. find
  28. Thiago A. S. Pardo and Maria das Graças Volpe Nunes. 2003. A construção de um corpus de textos cientı́ficos em Português do Brasil e sua marcação retórica. Technical report, Technical Report. find
  29. Igor Leturia. 2012. Evaluating different methods for automatically collecting large general corpora for Basque from the web. In 24th International Conference on Computational Linguistics (COLING 2012), pages 1553–1570, Mumbai, India. find
  30. 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
  31. 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
  32. 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
  33. Ziheng Lin, Hwee Tou Ng, and Min-Yen Kan. 2014. A PDTB-styled end-to-end discourse parser. Natural Language Engineering, 20(02):151–184. find
  34. 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
  35. Yang Liu and Mirella Lapata. 2017. Learning contextually informed representations for linear-time discourse parsing. In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing. find our card
  36. Boris T. Polyak and Anatoli B. Juditsky. 1992. Acceleration of stochastic approximation by averaging. SIAM J. Control Optim., 30(4):838–855. find
  37. Willian 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
  38. Rashmi Prasad, Nikhil Dinesh, Alan Lee, Eleni Milt-sakaki, Livio Robaldo, Aravind Joshi, and Bonnie Webber. 2008. The penn discourse treebank 2.0. In Proceedings of LREC. find our card
  39. Daniel Marcu. 2000. The rhetorical parsing of unrestricted texts: A surface-based approach. Computational Linguistics, 26(3):395–448. find our card
  40. Radim Řehůřek and Petr Sojka. 2010. Software Framework for Topic Modelling with Large Corpora. In Proceedings of the LREC 2010 Workshop on New Challenges for NLP Frameworks, pages 45–50, Valletta, Malta. ELRA. find
  41. 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. find
  42. Kenji Sagae. 2009. Analysis of discourse structure with syntactic dependencies and data-driven shift-reduce parsing. In Proceedings of IWPT 2009. find
  43. Nynke van der Vliet. 2010. Inter annotator agreement in discourse analysis. http://www.let.rug.nl/ nerbonne/teach/rema-stats-meth-seminar/. ↗
  44. Erick Galani Maziero, Thiago Alexandre Salgueiro Pardo, and Núcleo Interinstitucional de Lingüı́stica Computacional. 2009. Automatização de um método de avaliação de estruturas retóricas. In Proceedings of the RST Brazilian Meeting, pages 1–9. find
  45. Nan Yu, Meishan Zhang, and Guohong Fu. 2018. Transition-based neural rst parsing with implicit syntax features. In Proceedings of the 27th International Conference on Computational Linguistics. find
  46. Tomas Mikolov, Ilya Sutskever, Kai Chen, Greg S Corrado, and Jeff Dean. 2013. Distributed representations of words and phrases and their compositionality. In C. J. C. Burges, L. Bottou, M. Welling, Z. Ghahramani, and K. Q. Weinberger, editors, Advances in Neural Information Processing Systems 26, pa… find
  47. Elena Mitocariu, Daniel Alexandru Anechitei, and Dan Cristea. 2013. Comparing discourse tree structures. In International Conference on Intelligent Text Processing and Computational Linguistics, pages 513– 522. Springer. find
  48. 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 Empirical Methods on Natural Language Processing (EMNLP). find our card

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/iruskieta-2019-eusdisparser-improving-under-resourced-discourse/