---
title: "Generation of Dialogue Breakdown Repair — Tsubokura 2026"
description: "Even capable LLM-based chat systems sometimes produce \"dialogue breakdowns,\" such as factual mistakes or contextually odd replies."
authors: ["Kazuya Tsubokura", "Yurie Iribe", "Norihide Kitaoka"]
year: 2026
venue: "Proceedings of the 27th Annual Meeting of the Special Interest Group on Discourse and Dialogue"
source: "https://aclanthology.org/2026.sigdial-1.38/"
doi: "10.18653/v1/2026.sigdial-1.38"
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/tsubokura-2026-generation-dialogue-breakdown-repair-utterance/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/tsubokura-2026-generation-dialogue-breakdown-repair-utterance/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2026](https://md.research.elamur.ai/free/en/papers/2026/page/5/) / Generation of Dialogue Breakdown Repair Utterance in Non-task Oriented Conversation

# Generation of Dialogue Breakdown Repair Utterance in Non-task Oriented Conversation

**Authors:** [Kazuya Tsubokura](https://md.research.elamur.ai/free/en/authors/kazuya-tsubokura/), [Yurie Iribe](https://md.research.elamur.ai/free/en/authors/yurie-iribe/), [Norihide Kitaoka](https://md.research.elamur.ai/free/en/authors/norihide-kitaoka/) · **Year:** 2026 · **Venue:** Proceedings of the 27th Annual Meeting of the Special Interest Group on Discourse and Dialogue · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2026.sigdial-1.38/) · [PDF](https://aclanthology.org/2026.sigdial-1.38.pdf) · [DOI 10.18653/v1/2026.sigdial-1.38](https://doi.org/10.18653/v1/2026.sigdial-1.38) · [OpenAlex](https://openalex.org/W7214026877)

**Citations:** 0 citations (openalex-snapshot · 2026-10-14)

## Summary
Even capable LLM-based chat systems sometimes produce "dialogue breakdowns," such as factual mistakes or contextually odd replies. Earlier repair methods only covered a few predefined error types, often with hand-crafted rules. This paper asks whether a general LLM can learn to recover smoothly from many kinds of breakdown in casual, open-domain Japanese conversation.

The authors use the Dialogue Breakdown Repair Corpus (DBR Corpus). In it, crowd workers who saw a dialogue ending in a system error wrote both a plausible user reaction and the repair reply they would expect from the system.

They first compared zero-shot GPT-4o-mini with two open Japanese-capable models; GPT-4o-mini scored best. They then fine-tuned GPT-4o-mini on the corpus. The fine-tuned model matched human references more closely on BLEU, BERTScore and S-BERT. Its outputs were also shorter and less verbose than the over-polite, information-heavy zero-shot replies.

The work shows that a modest, targeted dataset can push an LLM toward natural, human-like repair behaviour in casual dialogue.

## Key points
- **DBR Corpus:** Japanese, built on the Higashinaka et al. (2022) breakdown taxonomy. It covers 140 breakdown scenarios with 3,990 responses from 57 crowd workers. Data was split 8:1:1, and 399 test instances were used for evaluation.
- **Zero-shot baselines:** GPT-4o-mini outperformed llm-jp-3-3.7b-instruct and DeepSeek-R1-Distill-Qwen-7B-Japanese on all metrics. On BERTScore it reached 0.8889, versus 0.8740 and 0.8487. On S-BERT it reached 0.4795, versus 0.4034 and 0.2956.
- **Fine-tuning gains:** GPT-4o-mini was fine-tuned for 3 epochs with a learning-rate multiplier of 1.0. Compared with zero-shot prompting, BLEU rose from 6.769 to 12.843, BERTScore from 0.889 to 0.914, and S-BERT from 0.480 to 0.522.
- **Utterance length (sentences / morphemes):**
  - Fine-tuned model: 1.22 / 9.18
  - Crowd workers: 1.33 / 10.57
  - Zero-shot prompting: 2.66 / 27.68
  
  Friedman and Wilcoxon tests showed significant differences between the methods.
- **LLM-as-a-Judge and limitations:** A GPT-4o judge rated zero-shot replies as more redundant, polite and informative. Fine-tuned outputs scored closer to the human ones. Stated limitations are that the data is Japanese only and that no human evaluation was carried out.

*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](https://md.research.elamur.ai/free/en/shelves/linguistics/)

## Cite

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

```bibtex
@inproceedings{tsubokura2026generation,
  title = {{Generation of Dialogue Breakdown Repair Utterance in Non-task Oriented Conversation}},
  author = {Tsubokura, Kazuya and Iribe, Yurie and Kitaoka, Norihide},
  booktitle = {Proceedings of the 27th Annual Meeting of the Special Interest Group on Discourse and Dialogue},
  year = {2026},
  doi = {10.18653/v1/2026.sigdial-1.38},
  url = {https://doi.org/10.18653/v1/2026.sigdial-1.38}
}
```

- [← previous Generating Sign Language Poses from HamNoSys and Natural Language Descriptions](https://md.research.elamur.ai/free/en/papers/maximo-2026-generating-sign-language-poses-hamnosys/)
- [next → Generation of Instruction and Preference Dataset for Improving Japanese Instruction Following in LLMs](https://md.research.elamur.ai/free/en/papers/moriyama-2026-generation-instruction-preference-dataset-improving/)
- [↑ back to papers of 2026](https://md.research.elamur.ai/free/en/papers/2026/page/5/)

## References

12 references, 2 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Timothy Bickmore, Ha Trinh, Reza Asadi, and Stefan Olafsson. 2018. Safety First: Conversational Agents for Health Care, pages 33–57. Springer International Publishing, Cham. [find](https://scholar.google.com/scholar?q=Timothy+Bickmore%2C+Ha+Trinh%2C+Reza+Asadi%2C+and+Stefan+Olafsson%2E+2018%2E+Safety+First%3A+Conversational+Agents+for+Health+Care%2C+pages+33%E2%80%9357%2E+Springer+International+Publishing%2C+Cham%2E)
2. Chayan Chakrabarti and George F. Luger. 2015. Artificial conversations for customer service chatter bots: Architecture, algorithms, and evaluation metrics. Expert Systems with Applications, 42(20):6878– 6897. [find](https://scholar.google.com/scholar?q=Chayan+Chakrabarti+and+George+F%2E+Luger%2E+2015%2E+Artificial+conversations+for+customer+service+chatter+bots%3A+Architecture%2C+algorithms%2C+and+evaluation+metrics%2E+Expert+Systems+with+Applications%2C+42%2820%29%3A687)
3. Alexis Conneau, Kartikay Khandelwal, Naman Goyal, Vishrav Chaudhary, Guillaume Wenzek, Francisco Guzmán, Edouard Grave, Myle Ott, Luke Zettlemoyer, and Veselin Stoyanov. 2019. Unsupervised cross-lingual representation learning at scale. CoRR, abs/1911.02116. [find](https://scholar.google.com/scholar?q=Alexis+Conneau%2C+Kartikay+Khandelwal%2C+Naman+Goyal%2C+Vishrav+Chaudhary%2C+Guillaume+Wenzek%2C+Francisco+Guzm%C3%A1n%2C+Edouard+Grave%2C+Myle+Ott%2C+Luke+Zettlemoyer%2C+and+Veselin+Stoyanov%2E+2019%2E+Unsupervised+cross%2Dlingu)
4. Ryuichiro Higashinaka, Masahiro Araki, Hiroshi Tsukahara, and Masahiro Mizukami. 2022. Classification of utterances that lead to dialogue breakdowns in chat-oriented dialogue systems (in japanese). Journal of Natural Language Processing, 29(2):443–466. [find](https://scholar.google.com/scholar?q=Ryuichiro+Higashinaka%2C+Masahiro+Araki%2C+Hiroshi+Tsukahara%2C+and+Masahiro+Mizukami%2E+2022%2E+Classification+of+utterances+that+lead+to+dialogue+breakdowns+in+chat%2Doriented+dialogue+systems+%28in+japanese%29%2E+Jo)
5. Michimasa Inaba and Kenichi Takahashi. 2019. Applying dialogue breakdown detection to dialogue systems (in japanese). Transactions of the Japanese Society for Artificial Intelligence, 34(3):B–I64_1–8. [find](https://scholar.google.com/scholar?q=Michimasa+Inaba+and+Kenichi+Takahashi%2E+2019%2E+Applying+dialogue+breakdown+detection+to+dialogue+systems+%28in+japanese%29%2E+Transactions+of+the+Japanese+Society+for+Artificial+Intelligence%2C+34%283%29%3AB%E2%80%93I64%5F1%E2%80%938%2E)
6. Bilyana Martinovsky and David Traum. 2003. The error is the clue: breakdown in human-machine interaction. In Error Handling in Spoken Dialogue Systems, pages 11–16. [find](https://scholar.google.com/scholar?q=Bilyana+Martinovsky+and+David+Traum%2E+2003%2E+The+error+is+the+clue%3A+breakdown+in+human%2Dmachine+interaction%2E+In+Error+Handling+in+Spoken+Dialogue+Systems%2C+pages+11%E2%80%9316%2E)
7. Kishore Papineni, Salim Roukos, Todd Ward, and Wei-Jing Zhu. 2002. Bleu: a method for automatic evaluation of machine translation. In Proceedings of the 40th Annual Meeting of the Association for Computational Linguistics, pages 311–318, Philadelphia, Pennsylvania, USA. Association for Computational… [find](https://scholar.google.com/scholar?q=Kishore+Papineni%2C+Salim+Roukos%2C+Todd+Ward%2C+and+Wei%2DJing+Zhu%2E+2002%2E+Bleu%3A+a+method+for+automatic+evaluation+of+machine+translation%2E+In+Proceedings+of+the+40th+Annual+Meeting+of+the+Association+for+Comp) [our card](https://md.research.elamur.ai/free/en/papers/papineni-2002-bleu-method-automatic-evaluation-machine/)
8. Nils Reimers and Iryna Gurevych. 2019. Sentence-BERT: Sentence embeddings using Siamese BERT-networks. 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), pages 3982–3992,… [find](https://scholar.google.com/scholar?q=Nils+Reimers+and+Iryna+Gurevych%2E+2019%2E+Sentence%2DBERT%3A+Sentence+embeddings+using+Siamese+BERT%2Dnetworks%2E+In+Proceedings+of+the+2019+Conference+on+Empirical+Methods+in+Natural+Language+Processing+and+the)
9. Hiroaki Sugiyama, Toyomi Meguro, Yuichiro Yoshikawa, and Junji Yamato. 2018. Avoiding breakdown of conversational dialogue through inter-robot coordination. In Proceedings of the 17th International Conference on Autonomous Agents and MultiAgent Systems, AAMAS ’18, page 2256–2258, Richland, SC. Inter… [find](https://scholar.google.com/scholar?q=Hiroaki+Sugiyama%2C+Toyomi+Meguro%2C+Yuichiro+Yoshikawa%2C+and+Junji+Yamato%2E+2018%2E+Avoiding+breakdown+of+conversational+dialogue+through+inter%2Drobot+coordination%2E+In+Proceedings+of+the+17th+International+Co)
10. Kazuya Tsubokura, Yurie Iribe, and Norihide Kitaoka. 2026. A corpus for personalized dialogue breakdown repair in japanese open-domain conversations. In Proceedings of the Fifteenth Language Resources and Evaluation Conference (LREC 2026), pages 2899–2912, Palma, Mallorca, Spain. European Language R… [find](https://scholar.google.com/scholar?q=Kazuya+Tsubokura%2C+Yurie+Iribe%2C+and+Norihide+Kitaoka%2E+2026%2E+A+corpus+for+personalized+dialogue+breakdown+repair+in+japanese+open%2Ddomain+conversations%2E+In+Proceedings+of+the+Fifteenth+Language+Resources) [our card](https://md.research.elamur.ai/free/en/papers/tsubokura-2026-corpus-personalized-dialogue-breakdown-repair/)
11. Takahisa Uchida, Takashi Minato, Tora Koyama, and Hiroshi Ishiguro. 2019. Who is responsible for a dialogue breakdown? an error recovery strategy that promotes cooperative intentions from humans by mutual attribution of responsibility in human-robot dialogues. Frontiers in Robotics and AI, 6. [find](https://scholar.google.com/scholar?q=Takahisa+Uchida%2C+Takashi+Minato%2C+Tora+Koyama%2C+and+Hiroshi+Ishiguro%2E+2019%2E+Who+is+responsible+for+a+dialogue+breakdown%3F+an+error+recovery+strategy+that+promotes+cooperative+intentions+from+humans+by+mu)
12. Tianyi Zhang, Varsha Kishore, Felix Wu, Kilian Q. Weinberger, and Yoav Artzi. 2020. Bertscore: Evaluating text generation with bert. In International Conference on Learning Representations. [find](https://scholar.google.com/scholar?q=Tianyi+Zhang%2C+Varsha+Kishore%2C+Felix+Wu%2C+Kilian+Q%2E+Weinberger%2C+and+Yoav+Artzi%2E+2020%2E+Bertscore%3A+Evaluating+text+generation+with+bert%2E+In+International+Conference+on+Learning+Representations%2E)