Elamur Research⌕
☰
home / papers / 2023 / Task-Aware Self-Supervised Framework for Dialogue Discourse Parsing

Task-Aware Self-Supervised Framework for Dialogue Discourse Parsing

Summary

Dialogue discourse parsing (DDP) determines which earlier utterance each turn attaches to and what relation connects them. The authors identify three weaknesses in prior parsers. Sequential scanning lets mistakes accumulate. Keeping only one head per utterance distorts multi-head training labels. And a fixed relation inventory may not suit downstream applications.

They propose DialogDP, a graph-based parser. A BERT encoder with speaker tokens produces utterance representations. Two mirrored biaffine parsers then score links and relations in opposite directions: bottom-up (choosing a head for each dependent) and top-down (choosing dependents for each head). Because the two score matrices are symmetric counterparts, each parser regularizes the other through bidirectional self-supervision. A soft-window triangular (SWT) mask gently favors nearby candidate links without forbidding distant ones.

A task-aware paradigm then fine-tunes the parser on relations re-annotated for a target task, here emotion recognition in conversations (ERC). DialogDP outperforms baselines on STAC and Molweni, and task-aware parsing improves ERC. The work shows that internal structure can replace external knowledge as a regularizer.

Key points

  • Evaluated on the STAC and Molweni corpora with 16 predefined relation types (Asher et al., 2016). A root node u0 is added, and the parser outputs a dependency tree. Code is at github.com/senticnet/DialogDP.
  • Per-turn analysis on Molweni:
    • DialogDP's weakest points are 48.0% (turn 5) and 43.1% (turn 9).
    • The model of Wang et al. (2021) bottoms out at 40.0% (turn 5) and 41.5% (turn 8).
    • DialogDP leads at every turn except the 9th, while DeepSequential is lowest throughout.
  • On STAC, F1 decreases as dialogues get longer, with link-plus-relation scores dropping faster than link-only scores. Molweni dialogues (mostly 7–14 utterances) show no clear length trend.
  • In a case study, ChatGPT predicted spurious links and missed every QAP pair, whereas DialogDP's attention clearly separated correct from incorrect links.
  • Limitations: designing task-specific relations lacks theoretical grounding, and the task-aware paradigm was tested only on ERC.

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

5 papers of our library cite it (of the 140673 whose bibliographies we read); 1 of them have a page here, 4 not yet.

Cite

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2023.findings-emnlp.943

@inproceedings{li2023task,
  title = {{Task-Aware Self-Supervised Framework for Dialogue Discourse Parsing}},
  author = {Li, Wei and Zhu, Luyao and Shao, Wei and Yang, Zonglin and Cambria, Erik},
  booktitle = {Findings of the Association for Computational Linguistics: EMNLP 2023},
  year = {2023},
  doi = {10.18653/v1/2023.findings-emnlp.943},
  url = {https://doi.org/10.18653/v1/2023.findings-emnlp.943}
}

References

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

  1. Stergos Afantenos, Eric Kow, Nicholas Asher, and Jérémy Perret. 2015. Discourse parsing for multi-party chat dialogues. In Conference on Empirical Methods on Natural Language Processing (EMNLP 2015), pages pp–928. find our card
  2. Nicholas Asher, Julie Hunter, Mathieu Morey, Farah Benamara, and Stergos Afantenos. 2016. Discourse structure and dialogue acts in multiparty dialogue: the stac corpus. In 10th International Conference on Language Resources and Evaluation (LREC 2016), pages 2721–2727. find our card
  3. Nicholas Asher and Alex Lascarides. 2003. Logics of conversation. Cambridge University Press. find
  4. Deng Cai and Wai Lam. 2020. Graph transformer for graph-to-sequence learning. In Proceedings of the AAAI conference on artificial intelligence, volume 34, pages 7464–7471. find
  5. Jiaao Chen and Diyi Yang. 2021. Structure-aware abstractive conversation summarization via discourse and action graphs. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 1380–1391. find
  6. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2018. Bert: Pre-training of deep bidirectional transformers for language understanding. arXiv preprint arXiv:1810.04805. find
  7. Jiaqi Li, Ming Liu, Min-Yen Kan, Zihao Zheng, Zekun Wang, Wenqiang Lei, Ting Liu, and Bing Qin. 2020. Molweni: A challenge multiparty dialogues-based machine reading comprehension dataset with discourse structure. In Proceedings of the 28th International Conference on Computational Linguistics, page… find
  8. Jiwei Li, Rumeng Li, and Eduard Hovy. 2014a. Recursive deep models for discourse parsing. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 2061–2069. find our card
  9. Sujian Li, Liang Wang, Ziqiang Cao, and Wenjie Li. 2014b. Text-level discourse dependency parsing. In Proceedings of the 52nd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 25–35. find
  10. Wei Li, Yang Li, Vlad Pandelea, Mengshi Ge, Luyao Zhu, and Erik Cambria. 2022. Ecpec: Emotion-cause pair extraction in conversations. IEEE Transactions on Affective Computing, (01):1–12. find
  11. Wei Li, Luyao Zhu, Rui Mao, and Erik Cambria. 2023. Skier: A symbolic knowledge integrated model for conversational emotion recognition. In Proceedings of the AAAI Conference on Artificial Intelligence. find
  12. Timothy Dozat and Christopher D. Manning. 2017. Deep biaffine attention for neural dependency parsing. In International Conference on Learning Representations. find our card
  13. Yanran Li, Hui Su, Xiaoyu Shen, Wenjie Li, Ziqiang Cao, and Shuzi Niu. 2017. Dailydialog: A manually labelled multi-turn dialogue dataset. In Proceedings of the Eighth International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 986–995. find our card
  14. Yaxin Fan, Peifeng Li, Fang Kong, and Qiaoming Zhu. 2022. A distance-aware multi-task framework for conversational discourse parsing. In Proceedings of the 29th International Conference on Computational Linguistics, pages 912–921. find our card
  15. Zhengyuan Liu and Nancy Chen. 2021. Improving multi-party dialogue discourse parsing via domain integration. In Proceedings of the 2nd Workshop on Computational Approaches to Discourse, pages 122–127. find our card
  16. Qingqing Gao, Biwei Cao, Xin Guan, Tianyun Gu, Xing Bao, Junyan Wu, Bo Liu, and Jiuxin Cao. 2022. Emotion recognition in conversations with emotion shift Ilya Loshchilov and Frank Hutter. 2018. Decoupled weight decay regularization. In International Conference on Learning Representations. 14171 10 find
  17. Ryan Lowe, Nissan Pow, Iulian Vlad Serban, and Joelle Pineau. 2015. The ubuntu dialogue corpus: A large dataset for research in unstructured multi-turn dialogue systems. In Proceedings of the 16th Annual Meeting of the Special Interest Group on Discourse and Dialogue, pages 285–294. find our card
  18. Xinbei Ma, Zhuosheng Zhang, and Hai Zhao. 2023. Enhanced speaker-aware multi-party multi-turn dialogue comprehension. IEEE/ACM Transactions on Audio, Speech, and Language Processing, pages 1– 16. find
  19. 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
  20. Ninareh Mehrabi, Fred Morstatter, Nripsuta Saxena, Kristina Lerman, and Aram Galstyan. 2021. A survey on bias and fairness in machine learning. ACM computing surveys (CSUR), 54(6):1–35. find our card
  21. Soujanya Poria, Devamanyu Hazarika, Navonil Majumder, Gautam Naik, Erik Cambria, and Rada Mihalcea. 2019. Meld: A multimodal multi-party dataset for emotion recognition in conversations. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 527–536. find our card
  22. 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 LREC. find our card
  23. Zhouxing Shi and Minlie Huang. 2019. A deep sequential model for discourse parsing on multi-party dialogues. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 33, pages 7007–7014. find
  24. Ante Wang, Linfeng Song, Hui Jiang, Shaopeng Lai, Junfeng Yao, Min Zhang, and Jinsong Su. 2021. A structure self-aware model for discourse parsing on multi-party dialogues. In Proceedings of the Thirtieth International Joint Conference on Artificial Intelligence, IJCAI-21, pages 3943–3949. Internati… find
  25. Nan Yu, Guohong Fu, and Min Zhang. 2022. Speaker-aware discourse parsing on multi-party dialogues. In Proceedings of the 29th International Conference on Computational Linguistics, pages 5372–5382. find our card
  26. Yu Zhang, Zhenghua Li, and Zhang Min. 2020. Efficient second-order TreeCRF for neural dependency parsing. In Proceedings of ACL, pages 3295–3305. find our card
  27. Zhuosheng Zhang and Hai Zhao. 2021. Structural pre-training for dialogue comprehension. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 5134–5145, Onli… find

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/li-2023-task-aware-self-supervised-framework/