---
title: "Utterance Position-Aware Dialogue Act Recognition — Yano 2021"
description: "They build on the seq2seq DAR model of Colombo et al. (2020), which uses a hierarchical BiGRU encoder, a GRU label… 14 references."
authors: ["Yuki Yano", "Akihiro Tamura", "Takashi Ninomiya", "Hiroaki Obayashi"]
year: 2021
venue: "Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2021)"
source: "https://aclanthology.org/2021.ranlp-1.176/"
licence: "CC BY 4.0 (ACL Anthology)"
licence_url: "https://creativecommons.org/licenses/by/4.0/"
summary_licence: "Elamur Cosmic License"
canonical: "https://research.elamur.ai/free/en/papers/yano-2021-utterance-position-aware-dialogue-act/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/yano-2021-utterance-position-aware-dialogue-act/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2021](https://md.research.elamur.ai/free/en/papers/2021/page/8/) / Utterance Position-Aware Dialogue Act Recognition

# Utterance Position-Aware Dialogue Act Recognition

**Authors:** Yuki Yano, Akihiro Tamura, [Takashi Ninomiya](https://md.research.elamur.ai/free/en/authors/takashi-ninomiya/) and Hiroaki Obayashi · **Year:** 2021 · **Venue:** [Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2021)](https://md.research.elamur.ai/free/en/venues/ranlp/) ([RANLP 2021](https://md.research.elamur.ai/free/en/venues/ranlp/2021/)) · **Publisher:** [INCOMA Ltd.](https://md.research.elamur.ai/free/en/publishers/incoma-ltd/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2021.ranlp-1.176/) · [PDF](https://aclanthology.org/2021.ranlp-1.176.pdf) · [OpenAlex](https://openalex.org/W3213439193)

**Citations:** 1 citations (OpenAlex, 2026-10-14)

## Summary
The paper looks at dialogue act recognition (DAR), the task of labelling each utterance in a conversation with its communicative function. Existing neural DAR models use the surrounding context and the dependencies between labels, but they ignore where an utterance sits in the dialogue. The authors noticed that some acts cluster at particular points. Open-Question, for example, tends to appear early in Switchboard calls. They build on the seq2seq DAR model of Colombo et al. (2020), which uses a hierarchical BiGRU encoder, a GRU label decoder and hard guided attention. Just before label prediction, they concatenate a positional vector for each utterance. Two variants are tested: Transformer-style sinusoidal encoding of the absolute index, and a length-ratio encoding that divides the index by the dialogue length. On SwDA, both variants give small but statistically significant precision gains. The gains are largest for long conversations and for acts with a strong positional profile. The work shows that a cheap structural signal can complement contextual modelling.

## Key points
- **Data:** the Switchboard corpus (SwDA) with the standard split of Lee and Dernoncourt (2016).
  - Train: 1,003 conversations; validation: 112; test: 19.
  - 43 DA labels and a vocabulary of about 19K.
- **Overall precision:** baseline s2s scores 77.84, s2s+PE abs scores 78.15 and s2s+PE rel scores 78.06. Both gains are significant (Wilcoxon signed-rank test, p ≤ 0.01).
- **Long dialogues:** for dialogues longer than 300 utterances, precision rises from 76.39 to 77.58 (PE abs) and 77.80 (PE rel). The baseline's accuracy drops steadily as dialogues get longer, and both position-aware models soften that drop.
- **Per-label effects:**
  - Yes Answers improves by +2.88 (abs) and +3.51 (rel).
  - Wh-Question falls by 1.93 with PE abs but rises by 0.73 with PE rel.
- **Confusable label groups:** position reduces confusion between Backchannel, Agree/Accept and Yes Answer. It can increase confusion between Wh-, Open- and Rhetorical-Question, because these three share similar positional profiles.
- **Tuning:** hyperparameters were set with Optuna over 100 trials.

## Interesting ideas
- Dividing an utterance's index by the dialogue length captures its phase (opening, middle, closing) in a way that works across texts of different lengths; this helps whenever a label depends on discourse stage rather than raw offset.
- Before adding a positional feature, the authors plotted each label's distribution over absolute and relative position; this tells you in advance which classes can benefit.
- The analysis of confusable groups shows that a feature only separates similar-looking classes when their distributions on that feature differ, and that it can make confusion worse otherwise.
- Splitting the test set into length buckets revealed that the gains were concentrated on long inputs, which the overall average had hidden.

## For Mova

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

- **I1** (frames): the speech-act layer in prag could include a normalised position feature. Greetings, openers and closings would then be explained by a rule like "occurs in the first 5% of the dialogue".
- **I2** (learn): we could produce these histograms for speech-act or frame labels on silver data. Positional cues would then be added only where the histograms show a real skew, and the histograms would serve as documented evidence for that choice.
- **I3** (frames): when distinguishing look-alike acts such as backchannel "yeah", agreement and yes-answers, we could check whether a candidate cue actually separates them before trusting it.
- **I4** (answer): QA on fables and tales, and the school-test readers, could report accuracy per text-length bucket. This would show where long-range world tracking breaks down.

*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

- [Nara Institute of Science and Technology](https://md.research.elamur.ai/free/en/universities/nara-institute-of-science-and-technology/) — University, Ikoma, Japan · 34 works · 4,419 citations
- [Ehime University](https://md.research.elamur.ai/free/en/universities/ehime-university/) — University, Matsuyama, Japan · 3 works · 2 citations

Also on the byline: [Doshisha University](https://ror.org/01fxdkm29).

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 9,085 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 8,888 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 9,296 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 9,206 works

🧭 Topics (2): [🗨️ Dialogue systems](https://md.research.elamur.ai/free/en/topics/dialogue-systems/) · 401 works · [💬 Speech act theory](https://md.research.elamur.ai/free/en/topics/speech-act-theory/) · 160 works

## Cite

Cite as: Yuki Yano, Akihiro Tamura, Takashi Ninomiya and Hiroaki Obayashi. 2021. Utterance Position-Aware Dialogue Act Recognition. Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2021).

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

```bibtex
@inproceedings{yano2021utterance,
  title = {{Utterance Position-Aware Dialogue Act Recognition}},
  author = {Yuki Yano, Akihiro Tamura, Takashi Ninomiya and Hiroaki Obayashi},
  booktitle = {Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2021)},
  year = {2021},
  doi = {10.26615/978-954-452-072-4\_176},
  url = {https://doi.org/10.26615/978-954-452-072-4_176}
}
```

- [← previous TPT: An Empirical Term Selection for Arabic Text Categorization](https://md.research.elamur.ai/free/en/papers/abbas-2021-tpt-empirical-term-selection-arabic/)
- [next → Validation of Universal Dependencies by regeneration](https://md.research.elamur.ai/free/en/papers/lapalme-2021-validation-universal-dependencies-regeneration/)
- [↑ back to papers of 2021](https://md.research.elamur.ai/free/en/papers/2021/page/8/)

## References

14 references, 10 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Takuya Akiba, Shotaro Sano, Toshihiko Yanase, Takeru Ohta, and Masanori Koyama. 2019. [Optuna: A next-generation hyperparameter optimization framework](https://md.research.elamur.ai/free/en/papers/akiba-2019-optuna-next-generation-hyperparameter-optimization/). In Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining, KDD 2019, Anchorage, AK, USA, August 4-8, 2019, pages 2623–2631. ACM.
2. Chandrakant Bothe, Cornelius Weber, Sven Magg, and Stefan Wermter. 2018. [A context-based approach for dialogue act recognition using simple recurrent neural networks](https://md.research.elamur.ai/free/en/papers/bothe-2018-context-based-approach-dialogue-act/). In Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018), Miyazaki, Japan. European Language Resources Association (ELRA).
3. Zheqian Chen, Rongqin Yang, Zhou Zhao, Deng Cai, and Xiaofei He. 2018. [Dialogue act recognition via crf-attentive structured network](https://md.research.elamur.ai/free/en/papers/chen-2017-dialogue-act-recognition-crf-attentive/). In Proceedings of the 41st International ACM SIGIR Conference on Research & Development in Information Retrieval, SIGIR 2018, Ann Arbor, MI, USA, July 08-12, 2018, pages 225–234. ACM.
4. Kyunghyun Cho, Bart van Merriënboer, Dzmitry Bahdanau, and Yoshua Bengio. 2014. [On the properties of neural machine translation: Encoder–decoder approaches](https://md.research.elamur.ai/free/en/papers/cho-2014-properties-neural-machine-translation-encoder/). In Proceedings of SSST-8, Eighth Workshop on Syntax, Semantics and Structure in Statistical Translation, pages 103–111, Doha, Qatar. Association for Computational Linguistics.
5. 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](https://md.research.elamur.ai/free/en/papers/colombo-2020-guiding-attention-sequence-sequence-models/). In Proceedings of the Thirty-Fourth AAAI Conference on Artificial Intelligence, pages 7594–7601. AAAI Press. [↗](https://doi.org/10.48550/arxiv.2002.08801)
6. Diederik P. Kingma and Jimmy Ba. 2015. Adam: A method for stochastic optimization. In Proceedings of the 3rd International Conference on Learning Representations, ICLR 2015, San Diego, CA, USA, May 7-9, 2015, Conference Track Proceedings. [↗](https://doi.org/10.48550/arxiv.1412.6980)
7. Harshit Kumar, Arvind Agarwal, Riddhiman Dasgupta, and Sachindra Joshi. 2018. [Dialogue act sequence labeling using hierarchical encoder with CRF](https://md.research.elamur.ai/free/en/papers/kumar-2017-dialogue-act-sequence-labeling-using/). In Proceedings of the Thirty-Second AAAI Conference on Artificial Intelligence, pages 3440–3447. AAAI Press. [↗](https://doi.org/10.1609/aaai.v32i1.11701)
8. Ji Young Lee and Franck Dernoncourt. 2016. [Sequential short-text classification with recurrent and convolutional neural networks](https://md.research.elamur.ai/free/en/papers/lee-2016-sequential-short-text-classification-recurrent/). In Proceedings of the 2016 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 515–520, San Diego, California. Association for Computational Linguistics.
9. 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), pages 383–392, Hong Kong, China. Association for Computational Linguistics. [↗](https://doi.org/10.48550/arxiv.1810.09154)
10. Vipul Raheja and Joel Tetreault. 2019. [Dialogue Act Classification with Context-Aware Self-Attention](https://md.research.elamur.ai/free/en/papers/raheja-2019-dialogue-act-classification-context-aware/). In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, Volume 1 (Long and Short Papers), pages 3727–3733, Minneapolis, Minnesota. Association for Computational Linguistics. [↗](https://doi.org/10.18653/v1/n19-1373)
11. Nitish Srivastava, Geoffrey E. Hinton, Alex Krizhevsky, Ilya Sutskever, and Ruslan Salakhutdinov. 2014. Dropout: a simple way to prevent neural networks from overfitting. Journal of Machine Learning Research 15, 15(1):1929–1958. [find](https://scholar.google.com/scholar?q=Nitish+Srivastava%2C+Geoffrey+E%2E+Hinton%2C+Alex+Krizhevsky%2C+Ilya+Sutskever%2C+and+Ruslan+Salakhutdinov%2E+2014%2E+Dropout%3A+a+simple+way+to+prevent+neural+networks+from+overfitting%2E+Journal+of+Machine+Learning+R)
12. 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](https://md.research.elamur.ai/free/en/papers/stolcke-2000-dialogue-act-modeling-automatic-tagging/). Computational Linguistics, 26(3):339–374. [↗](https://doi.org/10.1162/089120100561737)
13. Sho Takase and Naoaki Okazaki. 2019. Positional encoding to control output sequence length. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, NAACL-HLT 2019, Minneapolis, MN, USA, June 2-7, 2019, Volume 1 (Long and Short Papers), pages 3999–4004. Association for Computational Linguistics. [↗](https://doi.org/10.48550/arxiv.1904.07418)
14. Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N. Gomez, Lukasz Kaiser, and Illia Polosukhin. 2017. [Attention is all you need](https://md.research.elamur.ai/free/en/papers/vaswani-2017-attention-all-need/). In Proceedings of the Advances in Neural Information Processing Systems 30: Annual Conference on Neural Information Processing Systems 2017, 4-9 December 2017, Long Beach, CA, USA, pages 5998–6008.

*Updated 24 October 2026*