Elamur Research⌕
☰
home / papers / 2021 / Let's be explicit about that: Distant supervision for implicit discourse relation classification via connective prediction

Let's be explicit about that: Distant supervision for implicit discourse relation classification via connective prediction

Summary

Classifying implicit discourse relations, where adjacent sentences are linked without any overt connective, is hard. Even human annotators often disagree, so labeled data is scarce. Kurfalı and Östling avoid using any annotated implicit relations. Instead, they split the problem into language modeling and explicit relation classification, which is a much easier task. Their pipeline works in three steps:

  • It inserts each of 65 single-word PDTB 2.0 connectives between the two arguments.
  • It ranks the resulting candidates with an off-the-shelf pre-trained language model (BERT variants, RoBERTa, DistilBERT or GPT-2).
  • It labels the candidates with a classifier trained only on PDTB explicit relations. The final sense comes either from the top-ranked connective or from marginalizing over all connectives.

Despite its simplicity, this distantly supervised setup slightly beats the previous best method that also uses no implicit labels, a more complex adversarial domain-adaptation model. It is also the first such work to report 11-way results. Without any tuning on biomedical text, it generalizes well to BioDRB, surpassing the zero-shot transfer of supervised systems. The work also offers connective explicitation as a probe of how well language models grasp relations between sentences.

Key points

  • Motivation from an oracle experiment: filling in the gold implicit connectives allowed 4-way classification at an F-score of 88.74. This shows how much is gained once a relation is made explicit.
  • Connective set: of the 100 connectives in the PDTB 2.0 guidelines, only the 65 one-word items were used, because masked LMs cannot fill several tokens at once.
  • Classifier training data: 13,639 PDTB explicit relations for 4-way and 12,695 for 11-way classification. Explicit relations not in arg1–connective–arg2 order were dropped, which removed 2,558 instances (13.85%).
  • Comparisons: in-domain, the baseline is Huang and Li (2019). Cross-domain, results on BioDRB are compared with Shi and Demberg (2019b), on both the standard test split (247 instances for 4-way, 140 for 11-way) and the full corpus (3,001 and 1,755).
  • Why it transfers well: the authors argue that LMs left unfine-tuned carry knowledge from many domains. Once these models surface a fitting connective, the explicit classifier is less sensitive to the shift in domain.

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

17 papers of our library cite it (of the 140673 whose bibliographies we read); 7 of them have a page here, 10 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2021.unimplicit-1.1

@inproceedings{kurfal2021let,
  title = {{Let's be explicit about that: Distant supervision for implicit discourse relation classification via connective prediction}},
  author = {Kurfalı, Murathan and Östling, Robert},
  booktitle = {Proceedings of the 1st Workshop on Understanding Implicit and Underspecified Language},
  year = {2021},
  doi = {10.18653/v1/2021.unimplicit-1.1},
  url = {https://doi.org/10.18653/v1/2021.unimplicit-1.1}
}

References

22 references, 14 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Hongxiao Bai and Hai Zhao. 2018. Deep enhanced representation for implicit discourse relation recognition. In Proceedings of the 27th International Conference on Computational Linguistics, pages 571– 583, Santa Fe, New Mexico, USA. Association for Computational Linguistics. find our card
  2. Daniel Marcu and Abdessamad Echihabi. 2002. An unsupervised approach to recognizing discourse relations. In Proceedings of the 40th annual meeting of the association for computational linguistics, pages 368–375. find our card
  3. Chloé Braud and Pascal Denis. 2014. Combining natural and artificial examples to improve implicit discourse relation identification. In Proceedings of COLING 2014, the 25th International Conference on Computational Linguistics: Technical Papers, pages 1694–1705, Dublin, Ireland. Dublin City Univers… find our card
  4. Allen Nie, Erin Bennett, and Noah Goodman. 2019. DisSent: Learning sentence representations from explicit discourse relations. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 4497–4510, Florence, Italy. Association for Computational Linguistics. find our card
  5. Emily Pitler, Annie Louis, and Ani Nenkova. 2009. Automatic sense prediction for implicit discourse relations in text. In Proceedings of the Joint Conference of the 47th Annual Meeting of the ACL and the Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. Bert: Pre-training of de… find our card
  6. Emily Pitler and Ani Nenkova. 2009. Using syntax to disambiguate explicit discourse connectives in text. In Proceedings of the ACL-IJCNLP 2009 Conference Short Papers, pages 13–16, Suntec, Singapore. Association for Computational Linguistics. find our card
  7. Caroline Sporleder and Alex Lascarides. 2008. Using automatically labelled examples to classify rhetorical relations: An assessment. Natural Language Engineering, 14(3):369. find
  8. Rashmi Prasad, Nikhil Dinesh, Alan Lee, Eleni Milt-sakaki, Livio Robaldo, Aravind K Joshi, and Bonnie L Webber. 2008. The penn discourse treebank 2.0. In In Proceedings of the 6th International Conference on Language Resources and Evaluation (LREC). find our card
  9. Evgeny Stepanov and Giuseppe Riccardi. 2014. Towards cross-domain pdtb-style discourse parsing. In Proceedings of the 5th International Workshop on Health Text Mining and Information Analysis (Louhi), pages 30–37. find our card
  10. Rashmi Prasad, Susan McRoy, Nadya Frid, Aravind Joshi, and Hong Yu. 2011. The biomedical discourse relation bank. BMC bioinformatics, 12(1):1– 18. find
  11. Thomas Wolf, Lysandre Debut, Victor Sanh, Julien Chaumond, Clement Delangue, Anthony Moi, Pier-ric Cistac, Tim Rault, Remi Louf, Morgan Funtow-icz, Joe Davison, Sam Shleifer, Patrick von Platen, Clara Ma, Yacine Jernite, Julien Plu, Canwen Xu, Teven Le Scao, Sylvain Gugger, Mariama Drame, Quentin Lh… find our card
  12. Lianhui Qin, Zhisong Zhang, Hai Zhao, Zhiting Hu, and Eric Xing. 2017. Adversarial connective-exploiting networks for implicit discourse relation classification. In Proceedings of the 55th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 1006–1017, Vanco… find our card
  13. Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, and Ilya Sutskever. 2019. Language models are unsupervised multitask learners. OpenAI blog, 1(8):9. find
  14. Changxing Wu, Xiaodong Shi, Yidong Chen, Yanzhou Huang, and Jinsong Su. 2016. Bilingually-constrained synthetic data for implicit discourse relation recognition. In Proceedings of the 2016 Conference on Empirical Methods in Natural Language Processing, pages 2306–2312, Austin, Texas. Association for… find our card
  15. Attapol Rutherford and Nianwen Xue. 2015. Improving the inference of implicit discourse relations via classifying explicit discourse connectives. In Proceedings of the 2015 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 7… find our card
  16. Yu Xu, Man Lan, Yue Lu, Zheng Yu Niu, and Chew Lim Tan. 2012. Connective prediction using machine learning for implicit discourse relation classification. In The 2012 International Joint Conference on Neural Networks (IJCNN), pages 1–8. IEEE. find
  17. Victor Sanh, Lysandre Debut, Julien Chaumond, and Thomas Wolf. 2019. Distilbert, a distilled version of bert: smaller, faster, cheaper and lighter. arXiv preprint arXiv:1910.01108. find
  18. Deniz Zeyrek and Murathan Kurfalı. 2017. TDB 1.1: Extensions on Turkish discourse bank. In Proceedings of the 11th Linguistic Annotation Workshop, pages 76–81, Valencia, Spain. Association for Computational Linguistics. find
  19. Wei Shi and Vera Demberg. 2019a. Learning to explic-itate connectives with Seq2Seq network for implicit discourse relation classification. In Proceedings of the 13th International Conference on Computational Semantics - Long Papers, pages 188–199, Gothenburg, Sweden. Association for Computational Li… find
  20. Zhi-Min Zhou, Yu Xu, Zheng-Yu Niu, Man Lan, Jian Su, and Chew Lim Tan. 2010. Predicting discourse connectives for implicit discourse relation recognition. In Proceedings of the 23rd International Conference on Computational Linguistics: Posters, pages 1507–1514. Association for Computational Linguis… find our card
  21. Wei Shi and Vera Demberg. 2019b. Next sentence prediction helps implicit discourse relation classification within and across domains. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (E… find our card
  22. Šárka Zikánová, Jiřı́ Mı́rovskỳ, and Pavlı́na Synková. 2019. Explicit and implicit discourse relations in the prague discourse treebank. In International Conference on Text, Speech, and Dialogue, pages 236–248. Springer. find

↑ back to top

This page in Markdown →