Elamur Research⌕
☰
home / papers / 2020 / Integrating User History into Heterogeneous Graph for Dialogue Act Recognition

Integrating User History into Heterogeneous Graph for Dialogue Act Recognition

Paper · 2020 · Dong Wang, Ziran Li, Haitao Zheng et al.

Summary

The paper tackles multi-label Dialogue Act Recognition (DAR) in multi-party forum threads. In this setting one utterance can carry several DA labels, and different users phrase the same intention in personal, often indirect ways. The authors' insight is that a user's earlier answers, grouped by DA label, expose these habits. One example is a user who signals doubt with a question that looks polite on the surface.

Their HUH model works in two phases:

  • Phase 1: a CNN over GloVe embeddings plus a BiLSTM context encoder produces utterance vectors and a first estimate of label scores.
  • Phase 2: a heterogeneous graph connects utterance nodes with label nodes. Each label node is built from the top-K most relevant historical answers. An RGCN then passes information over this graph.

Three denoising steps keep unhelpful history from misleading the model: History Selection, Similarity Re-weighting and Edge Re-weighting.

On MSDialog and MRDA, HUH beats sequential and graph baselines such as CRNN and DialogueGCN. It roughly matches BERT while being much smaller, and it reaches state of the art when BERT replaces its encoder. The biggest gains are on subtle intentions such as Negative Feedback.

Key points

  • Task: multi-label DA tagging of forum turns. Experiments use MRDA and MSDialog, whose label set covers OQ, PA, GG and FD as well as IR, RQ, NF and JK.
  • Architecture: CNN + BiLSTM encoder with an auxiliary first-phase loss, followed by an RGCN over a graph with Utterance Nodes and Label Nodes built from user history.
  • Model size: about 20M parameters, with performance comparable to BERT (110M). Swapping in a fine-tuned BERT encoder gives the best results.
  • Per-label F1 on MSDialog:
    • Negative Feedback: HUH 41.9 vs DialogueGCN 21.8 and CRNN 28.8.
    • Repeat Question: HUH 60.8 vs 52.1 and 28.1.
  • Ablations: removing the graph, the user-history label nodes, or the denoising (replaced by random history selection) each lowers performance.

Interesting ideas

  • I1. Build one "prototype" node per label by pooling a speaker's past utterances with that label, weighted by similarity to the current context; this lets idiosyncratic phrasing be matched against the person's own earlier examples rather than a global average.
  • I2. Use a cheap first-pass prediction to gate a richer second pass: the initial label scores re-weight which history items and which graph edges matter, so the extra context cannot dominate.
  • I3. Treat added context as noisy by default and filter it in stages (top-K selection, similarity re-weighting, edge re-weighting), then verify the filtering with an ablation that substitutes random selection.
  • I4. Report per-label scores and highlight that implicit intents (Negative Feedback, Repeat Question) are where context helps most, rather than relying only on an aggregate score.

For Mova

How these ideas could serve Mova, our small open language model — its files, stages and tests:

  • I1 (frames): crate prag could keep per-speaker exemplars of past speech acts, so an indirect "are you sure?" is classified as doubt when that speaker has used it that way before, with the matched exemplar cited as the explanation.
  • I2 (frames): a first speech-act guess from the UD tree alone could decide which earlier turns or coreferent mentions are allowed to influence the final act, making the influence of each piece of context explicit and inspectable.
  • I3 (learn): evaluations could include a negative control where retrieved support (e.g. silver examples or knowledge-graph facts) is replaced by random items, proving the gain comes from relevance rather than extra input.
  • I4 (answer): test suites for readers could be broken down by how explicit the needed cue is, showing whether world/context reasoning actually fixes the hard implicit cases.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Institutions

🎓Tsinghua UniversityBeijing · China137 works · 83,262 citations🎓Sun Yat-sen UniversityGuangzhou · China23 works · 8,980 citations🎓Tsinghua Shenzhen International Graduate SchoolShenzhen · China8 works · 2,860 citations

📚 Shelves 4The field of science this work belongs to (OpenAlex fields).Computer Science · 9,086 worksArtificial Intelligence · 8,889 worksArts and Humanities · 9,296 worksLanguage and Linguistics · 9,206 works

🧭 Topics 2Research subjects and languages the work is about.Dialogue systems · 401 worksSpeech act theory · 160 works

Cite

Cite as Dong Wang, Ziran Li, Haitao Zheng and Ying Shen. 2020. Integrating User History into Heterogeneous Graph for Dialogue Act Recognition. Proceedings of the 28th International Conference on Computational Linguistics.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2020.coling-main.372

@inproceedings{wang2020integrating,
  title = {{Integrating User History into Heterogeneous Graph for Dialogue Act Recognition}},
  author = {Dong Wang, Ziran Li, Haitao Zheng and Ying Shen},
  booktitle = {Proceedings of the 28th International Conference on Computational Linguistics},
  year = {2020},
  doi = {10.18653/v1/2020.coling-main.372},
  url = {https://doi.org/10.18653/v1/2020.coling-main.372}
}

References

29 references, 16 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Jeremy Ang, Yang Liu, and Elizabeth Shriberg. 2005. Automatic dialog act segmentation and classification in multiparty meetings. In 2005 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2005, pages 1061–1064. find
  2. Christophe Cerisara, Somayeh Jafaritazehjani, Adedayo Oluokun, and Hoa T. Le. 2018. Multi-task dialog act and sentiment recognition on mastodon. In Proceedings of the 27th International Conference on Computational Linguistics, COLING 2018, pages 745–754. ↗
  3. Zheqian Chen, Rongqin Yang, Zhou Zhao, Deng Cai, and Xiaofei He. 2018. Dialogue act recognition via crf-attentive structured network. In The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval, SIGIR 2018, pages 225–234.
  4. Pierre Colombo, Emile Chapuis, Matteo Manica, Emmanuel Vignon, Giovanna Varni, and Chloé Clavel. 2020. Guiding attention in sequence-to-sequence models for dialogue act prediction. In The Thirty-Fourth AAAI Conference on Artificial Intelligence, AAAI 2020, The Thirty-Second Innovative Applications of Artificial Intelligence Conference, IAAI 2020, The Tenth AAAI Symposium on Educational Advances in Artificial Intelligence, EAAI 2020, pages 7594–7601. ↗
  5. Jacob Devlin, Ming-Wei Chang, Kenton Lee, and Kristina Toutanova. 2019. BERT: pre-training of deep bidirectional transformers for language understanding. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL-HLT 2019, pages 4171–4186.
  6. Deepanway Ghosal, Navonil Majumder, Soujanya Poria, Niyati Chhaya, and Alexander F. Gelbukh. 2019. Dialoguegcn: A graph convolutional neural network for emotion recognition in conversation. In Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing, EMNLP-IJCNLP 2019, pages 154–164.
  7. Ryuichiro Higashinaka, Kenji Imamura, Toyomi Meguro, Chiaki Miyazaki, Nozomi Kobayashi, Hiroaki Sugiyama, Toru Hirano, Toshiro Makino, and Yoshihiro Matsuo. 2014. Towards an open-domain conversational system fully based on natural language processing. In COLING 2014, 25th International Conference on Computational Linguistics, Proceedings of the Conference: Technical Papers, pages 928–939. find
  8. Wenpeng Hu, Zhangming Chan, Bing Liu, Dongyan Zhao, Jinwen Ma, and Rui Yan. 2019. GSN: A graph-structured network for multi-party dialogues. In Proceedings of the Twenty-Eighth International Joint Conference on Artificial Intelligence, IJCAI 2019, pages 5010–5016. find
  9. Adam Janin, Don Baron, Jane Edwards, Dan Ellis, David Gelbart, Nelson Morgan, Barbara Peskin, Thilo Pfau, Elizabeth Shriberg, Andreas Stolcke, and Chuck Wooters. 2003. The ICSI meeting corpus. In 2003 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2003, pages 364–367. ↗
  10. Yangfeng Ji, Gholamreza Haffari, and Jacob Eisenstein. 2016. A latent variable recurrent neural network for discourse-driven language models. In NAACL HLT 2016, The 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 332–342. find
  11. Simon Keizer, Rieks op den Akker, and Anton Nijholt. 2002. Dialogue act recognition with bayesian networks for dutch dialogues. In Proceedings of the SIGDIAL 2002 Workshop, The 3rd Annual Meeting of the Special Interest Group on Discourse and Dialogue, pages 88–94.
  12. Hamed Khanpour, Nishitha Guntakandla, and Rodney D. Nielsen. 2016. Dialogue act classification in domain-independent conversations using a deep recurrent neural network. In COLING 2016, 26th International Conference on Computational Linguistics, Proceedings of the Conference: Technical Papers, pages 2012–2021.
  13. Minkyoung Kim and Harksoo Kim. 2018. Integrated neural network model for identifying speech acts, predicators, and sentiments of dialogue utterances. Pattern Recognit. Lett., 101:1–5. ↗
  14. Yoon Kim. 2014. Convolutional neural networks for sentence classification. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing, EMNLP 2014, pages 1746–1751.
  15. Harshit Kumar, Arvind Agarwal, Riddhiman Dasgupta, and Sachindra Joshi. 2018. Dialogue act sequence labeling using hierarchical encoder with CRF. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence, (AAAI-18), the 30th innovative Applications of Artificial Intelligence (IAAI-18), and the 8th AAAI Symposium on Educational Advances in Artificial Intelligence (EAAI-18), pages 3440–3447. ↗
  16. Ji Young Lee and Franck Dernoncourt. 2016. Sequential short-text classification with recurrent and convolutional neural networks. In NAACL HLT 2016, The 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 515–520.
  17. Ruizhe Li, Chenghua Lin, Matthew Collinson, Xiao Li, and Guanyi Chen. 2019. A dual-attention hierarchical recurrent neural network for dialogue act classification. In Proceedings of the 23rd Conference on Computational Natural Language Learning, CoNLL 2019, pages 383–392. ↗
  18. Jeffrey Pennington, Richard Socher, and Christopher D. Manning. 2014. Glove: Global vectors for word representation. In Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing, EMNLP 2014, pages 1532–1543.
  19. Libo Qin, Wanxiang Che, Yangming Li, Minheng Ni, and Ting Liu. 2020. Dcr-net: A deep co-interactive relation network for joint dialog act recognition and sentiment classification. In The Thirty-Fourth AAAI Conference on Artificial Intelligence, AAAI 2020, The Thirty-Second Innovative Applications of Artificial Intelligence Conference, IAAI 2020, The Tenth AAAI Symposium on Educational Advances in Artificial Intelligence, EAAI 2020, pages 8665–8672.
  20. Chen Qu, Liu Yang, W. Bruce Croft, Johanne R. Trippas, Yongfeng Zhang, and Minghui Qiu. 2018. Analyzing and characterizing user intent in information-seeking conversations. In The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval, SIGIR 2018, pages 989–992. find
  21. Chen Qu, Liu Yang, W. Bruce Croft, Yongfeng Zhang, Johanne R. Trippas, and Minghui Qiu. 2019. User intent prediction in information-seeking conversations. In Proceedings of the 2019 Conference on Human Information Interaction and Retrieval, CHIIR 2019, pages 25–33. find
  22. Vipul Raheja and Joel R. Tetreault. 2019. Dialogue act classification with context-aware self-attention. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL-HLT 2019, pages 3727–3733. ↗
  23. Michael Sejr Schlichtkrull, Thomas N. Kipf, Peter Bloem, Rianne van den Berg, Ivan Titov, and Max Welling. 2018. Modeling relational data with graph convolutional networks. In The Semantic Web - 15th International Conference, ESWC 2018, pages 593–607. ↗
  24. Elizabeth Shriberg, Rajdip Dhillon, Sonali Bhagat, Jeremy Ang, and Hannah Carvey. 2004. The ICSI meeting recorder dialog act (MRDA) corpus. In Proceedings of the SIGDIAL 2004 Workshop, The 5th Annual Meeting of the Special Interest Group on Discourse and Dialogue, pages 97–100. ↗
  25. Andreas Stolcke, Klaus Ries, Noah Coccaro, Elizabeth Shriberg, Rebecca Bates, Daniel Jurafsky, Paul Taylor, Rachel Martin, Carol Van Ess-Dykema, and Marie Meteer. 2000. Dialogue act modeling for automatic tagging and recognition of conversational speech. Computational Linguistics, 26(3):339–373. ↗
  26. Dinoj Surendran and Gina-Anne Levow. 2006. Dialog act tagging with support vector machines and hidden markov models. In INTERSPEECH 2006 - ICSLP, Ninth International Conference on Spoken Language Processing. find
  27. Jiahui Wen, Jingwei Ma, Yiliu Feng, and Mingyang Zhong. 2018. Hybrid attentive answer selection in CQA with deep users modelling. In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence, (AAAI-18), the 30th innovative Applications of Artificial Intelligence (IAAI-18), and the 8th AAAI Symposium on Educational Advances in Artificial Intelligence (EAAI-18), pages 2556–2563. find
  28. Liu Yang, Minghui Qiu, Chen Qu, Jiafeng Guo, Yongfeng Zhang, W. Bruce Croft, Jun Huang, and Haiqing Chen. 2018. Response ranking with deep matching networks and external knowledge in information-seeking conversation systems. In The 41st International ACM SIGIR Conference on Research & Development in Information Retrieval, SIGIR 2018, pages 245–254. find
  29. Yue Yu, Siyao Peng, and Grace Hui Yang. 2019. Modeling long-range context for concurrent dialogue acts recognition. In Proceedings of the 28th ACM International Conference on Information and Knowledge Management, CIKM 2019, pages 2277–2280.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/wang-2020-integrating-user-history-into-heterogeneous/