---
title: "Improving Classroom Dialogue Act Recognition from Limited"
description: "Labeled K-12 classroom dialogue is scarce. 45 references, cited by 2 papers here. Among them: 1."
authors: ["Vikram Kumaran", "Jonathan Rowe", "Bradford Mott", "Snigdha Chaturvedi", "James Lester"]
year: 2023
venue: "Findings of the Association for Computational Linguistics: ACL 2023"
source: "https://aclanthology.org/2023.findings-acl.698/"
doi: "10.18653/v1/2023.findings-acl.698"
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/kumaran-2023-improving-classroom-dialogue-act-recognition/"
---
[Elamur Research](https://md.research.elamur.ai/free/en/) · [this page in HTML](https://research.elamur.ai/free/en/papers/kumaran-2023-improving-classroom-dialogue-act-recognition/)

[home](https://md.research.elamur.ai/free/en/) / [papers](https://md.research.elamur.ai/free/en/papers/) / [2023](https://md.research.elamur.ai/free/en/papers/2023/page/13/) / Improving Classroom Dialogue Act Recognition from Limited Labeled Data with Self-Supervised Contrastive Learning Classifiers

# Improving Classroom Dialogue Act Recognition from Limited Labeled Data with Self-Supervised Contrastive Learning Classifiers

**Authors:** Vikram Kumaran, [Jonathan Rowe](https://md.research.elamur.ai/free/en/authors/jonathan-rowe/), [Bradford Mott](https://md.research.elamur.ai/free/en/authors/bradford-mott/), [Snigdha Chaturvedi](https://md.research.elamur.ai/free/en/authors/snigdha-chaturvedi/) and [James Lester](https://md.research.elamur.ai/free/en/authors/james-c-lester/) · **Year:** 2023 · **Venue:** [Findings of the Association for Computational Linguistics: ACL 2023](https://md.research.elamur.ai/free/en/venues/findings/) ([Findings of the Association for Computational Linguistics 2023](https://md.research.elamur.ai/free/en/venues/findings/2023/)) · **Publisher:** [Association for Computational Linguistics](https://md.research.elamur.ai/free/en/publishers/association-for-computational-linguistics/) · **Licence:** [CC BY 4.0 (ACL Anthology)](https://creativecommons.org/licenses/by/4.0/)

**Links:** [Publisher's page](https://aclanthology.org/2023.findings-acl.698/) · [PDF](https://aclanthology.org/2023.findings-acl.698.pdf) · [DOI 10.18653/v1/2023.findings-acl.698](https://doi.org/10.18653/v1/2023.findings-acl.698) · [OpenAlex](https://openalex.org/W4385565359)

**Citations:** 1 citations (OpenAlex, 2026-10-14) · [2](https://md.research.elamur.ai/free/en/papers/kumaran-2023-improving-classroom-dialogue-act-recognition/cited-by/) in our library, of the 149,439 papers whose bibliographies we read

## Summary
Labeled K-12 classroom dialogue is scarce. Privacy rules limit sharing, research groups each use their own coding schemes, and classroom talk is tied to the subject being taught. Kumaran et al. therefore address dialogue act recognition when only a handful of annotated utterances per class exist. Their system, SSCon, works in a loop of three stages:

* **Stage 1:** a classifier uses a DialoGPT model fine-tuned on the target corpus to represent each utterance in context, alongside a sentence-BERT vector for the utterance alone.
* **Stage 2:** the confident predictions become soft labels for a twin-network contrastive encoder, which reshapes the space so that utterances with the same act sit close together.
* **Stage 3:** a second classifier works on these new embeddings, and its confident outputs feed the next round.

Evaluation covers TalkMoves (math lessons), DailyDialog and the multi-label DSTC2. With only tens of examples per act, SSCon beats strong baselines and earlier few-shot methods that need more annotation. For educational dialogue analysis, this means useful classifiers can be built without large annotation campaigns.

## Key points
- Context encoder: HuggingFace "dialoGPT-small" (12 layers), fine-tuned for four epochs on next-utterance prediction with nine preceding utterances as context. The utterance representation is the end-of-sentence hidden state from the last layer (768-d).
- The utterance-only vector comes from sentence-BERT "all-MiniLM-L6-v2" (384-d). It is concatenated with the DialoGPT state into a 1152-d input, which a five-layer MLP twin encoder maps to 384-d.
- The contrastive loss is MSE between the cosine similarity of the two embeddings and a pair target: +1 for the same act, −1 for different acts. For multi-label DSTC2 the target is instead the cosine distance between one-hot label vectors.
- Classifiers: XGBoost in Stages 1 and 3, chosen because it matched MLP accuracy and was faster. An MLP multi-label classifier is used for DSTC2. Initial supervision is roughly ten to a hundred labeled examples per class.
- SSCon is reported to beat Mi et al.'s (2021a) teacher-student self-training and Guibon et al.'s (2021) prototypical network. The authors also say it can rival some fully supervised models.
- Stated limitations: the labeled seed set was chosen by uniform sampling, and multi-label performance needs further study.

## Interesting ideas
- Two models with different objectives, a context classifier and a contrastive embedder, take turns producing confidence-filtered silver labels for each other, so neither relies only on its own judgments.
- An utterance is represented by joining a dialogue-context vector with a context-free sentence vector, so the classifier sees both how the turn reads alone and where it falls in the conversation.
- For multi-label items, the contrastive target is not a hard same/different value but the similarity between the two label sets. Partially overlapping labels then get partially close embeddings.
- Confidence filtering governs how the training set grows: only high-probability predictions are moved from the unlabeled pool into the labeled pool each round.
- The authors flag that the choice of the small seed set matters, and that uniform sampling may be a weak strategy.

## For Mova

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

- **I1** (learn): Silver data could alternate between two differently built witnesses (e.g. the perceptron parser and gparse), with each one's high-confidence output training the other.
- **I2** (frames): The speech-act recognizer in prag could combine features of the sentence itself with features of the preceding turns, keeping the two parts separable for explanation.
- **I3** (frames): Utterances carrying several speech acts or frame roles could be compared with graded targets, avoiding forced single-label decisions.
- **I4** (learn): LLM-teacher or self-play labels could enter training only above a confidence threshold, with negative controls checking that later rounds do not drift.
- **I5** (learn): When hand-labeling a few examples per class, Mova's evaluation could compare seed-selection strategies rather than assume random choice is enough.

*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

- [University of North Carolina at Chapel Hill](https://md.research.elamur.ai/free/en/universities/university-of-north-carolina-at-chapel-hill/) — University, Chapel Hill, United States · 67 works · 16,474 citations
- [North Carolina State University](https://md.research.elamur.ai/free/en/universities/north-carolina-state-university/) — University, Raleigh, United States · 13 works · 406 citations

📚 Shelves (4): [🖥️ Computer Science](https://md.research.elamur.ai/free/en/shelves/computer-science/) · 16,667 works · [🤖 Artificial Intelligence](https://md.research.elamur.ai/free/en/shelves/computer-science/artificial-intelligence/) · 16,194 works · [🎭 Arts and Humanities](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/) · 16,039 works · [🗣️ Language and Linguistics](https://md.research.elamur.ai/free/en/shelves/arts-and-humanities/language-and-linguistics/) · 15,891 works

🧭 Topics (2): [🗨️ Dialogue systems](https://md.research.elamur.ai/free/en/topics/dialogue-systems/) · 624 works · [💬 Speech act theory](https://md.research.elamur.ai/free/en/topics/speech-act-theory/) · 229 works

## Cite

Cite as: Vikram Kumaran, Jonathan Rowe, Bradford Mott, Snigdha Chaturvedi and James Lester. 2023. Improving Classroom Dialogue Act Recognition from Limited Labeled Data with Self-Supervised Contrastive Learning Classifiers. Findings of the Association for Computational Linguistics: ACL 2023.

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

```bibtex
@inproceedings{kumaran2023improving,
  title = {{Improving Classroom Dialogue Act Recognition from Limited Labeled Data with Self-Supervised Contrastive Learning Classifiers}},
  author = {Vikram Kumaran, Jonathan Rowe, Bradford Mott, Snigdha Chaturvedi and James Lester},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2023},
  year = {2023},
  doi = {10.18653/v1/2023.findings-acl.698},
  url = {https://doi.org/10.18653/v1/2023.findings-acl.698}
}
```

- [← previous Improving Automated Prediction of English Lexical Blends Through the Use of Observable Linguistic Features](https://md.research.elamur.ai/free/en/papers/saunders-2023-improving-automated-prediction-english-lexical/)
- [next → IUNADI at NADI 2023 shared task: Country-level Arabic Dialect Classification in Tweets for the Shared Task NADI 2023](https://md.research.elamur.ai/free/en/papers/hatekar-2023-iunadi-nadi-2023-shared-task/)
- [↑ back to papers of 2023](https://md.research.elamur.ai/free/en/papers/2023/page/13/)

## References

45 references, 15 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

1. Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. [Language models are few-shot learners](https://md.research.elamur.ai/free/en/papers/brown-2020-language-models-few-shot-learners/). Advances in neural information processing systems, 33:1877–1901.
2. Maximillian Chen, Alexandros Papangelis, Chenyang Tao, Andy Rosenbaum, Seokhwan Kim, Yang Liu, Zhou Yu, and Dilek Hakkani-Tur. 2022. Weakly supervised data augmentation through prompting for dialogue understanding. arXiv preprint arXiv:2210.14169. [↗](https://arxiv.org/abs/2210.14169)
3. Tianqi Chen and Carlos Guestrin. 2016. XGBoost: A scalable tree boosting system. In Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, KDD ’16, pages 785–794, New York, NY, USA. ACM. [↗](https://doi.org/10.48550/arxiv.1603.02754)
4. 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 The 41st international acm sigir conference on research & development in information retrieval, pages 225–234. [↗](https://doi.org/10.1145/3209978.3209997)
5. Li Fei-Fei, Robert Fergus, and Pietro Perona. 2006. One-shot learning of object categories. IEEE transactions on pattern analysis and machine intelligence, 28(4):594–611. [↗](https://doi.org/10.1109/tpami.2006.79)
6. Chih-Wen Goo and Yun-Nung Chen. 2018. Abstractive dialogue summarization with sentence-gated modeling optimized by dialogue acts. In 2018 IEEE Spoken Language Technology Workshop (SLT), pages 735– 742. IEEE. [find](https://scholar.google.com/scholar?q=Chih%2DWen+Goo+and+Yun%2DNung+Chen%2E+2018%2E+Abstractive+dialogue+summarization+with+sentence%2Dgated+modeling+optimized+by+dialogue+acts%2E+In+2018+IEEE+Spoken+Language+Technology+Workshop+%28SLT%29%2C+pages+735%E2%80%93+742)
7. Gaël Guibon, Matthieu Labeau, Hélène Flamein, Luce Lefeuvre, and Chloé Clavel. 2021. Few-shot emotion recognition in conversation with sequential prototypical networks. In The 2021 Conference on Empirical Methods in Natural Language Processing (EMNLP 2021). [↗](https://doi.org/10.1016/j.simpa.2022.100237)
8. Raia Hadsell, Sumit Chopra, and Yann LeCun. 2006. Dimensionality reduction by learning an invariant mapping. In 2006 IEEE Computer Society Conference on Computer Vision and Pattern Recognition (CVPR’06), volume 2, pages 1735–1742. IEEE. [find](https://scholar.google.com/scholar?q=Raia+Hadsell%2C+Sumit+Chopra%2C+and+Yann+LeCun%2E+2006%2E+Dimensionality+reduction+by+learning+an+invariant+mapping%2E+In+2006+IEEE+Computer+Society+Conference+on+Computer+Vision+and+Pattern+Recognition+%28CVPR%E2%80%990)
9. Fangyu Liu, Yunlong Jiao, Jordan Massiah, Emine Yilmaz, and Serhii Havrylov. 2021a. Trans-encoder: Unsupervised sentence-pair modelling through self-and mutual-distillations. In International Conference on Learning Representations. [↗](https://doi.org/10.17863/cam.93939)
10. Tianyong Hao, Xieling Chen, and Yu Song. 2020. A topic-based bibliometric analysis of two decades of research on the application of technology in classroom dialogue. Journal of educational computing research, 58(7):1311–1341. [↗](https://doi.org/10.1177/0735633120940956)
11. Jiexi Liu, Ryuichi Takanobu, Jiaxin Wen, Dazhen Wan, Hongguang Li, Weiran Nie, Cheng Li, Wei Peng, and Minlie Huang. 2021b. Robustness testing of language understanding in task-oriented dialog. 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 2467– 2480. [↗](https://doi.org/10.48550/arxiv.2012.15262)
12. Simon Haykin. 1994. Neural networks: a comprehensive foundation. Prentice Hall PTR. [find](https://scholar.google.com/scholar?q=Simon+Haykin%2E+1994%2E+Neural+networks%3A+a+comprehensive+foundation%2E+Prentice+Hall+PTR%2E)
13. Matthew Henderson, Blaise Thomson, and Jason D Williams. 2014. The second dialog state tracking challenge. In Proceedings of the 15th annual meeting of the special interest group on discourse and dialogue (SIGDIAL), pages 263–272. [find](https://scholar.google.com/scholar?q=Matthew+Henderson%2C+Blaise+Thomson%2C+and+Jason+D+Williams%2E+2014%2E+The+second+dialog+state+tracking+challenge%2E+In+Proceedings+of+the+15th+annual+meeting+of+the+special+interest+group+on+discourse+and+dial)
14. Cindy E Hmelo-Silver. 2004. Problem-based learning: What and how do students learn? Educational psychology review, 16(3):235–266. [↗](https://doi.org/10.1023/b:edpr.0000034022.16470.f3)
15. Yangfeng Ji, Gholamreza Haffari, and Jacob Eisenstein. 2016. [A latent variable recurrent neural network for discourse relation language models](https://md.research.elamur.ai/free/en/papers/ji-2016-latent-variable-recurrent-neural-network/). In Proceedings of NAACL-HLT, pages 332–342. [↗](https://doi.org/10.18653/v1/n16-1037)
16. Nal Kalchbrenner and Phil Blunsom. 2013. [Recurrent convolutional neural networks for discourse compositionality](https://md.research.elamur.ai/free/en/papers/kalchbrenner-2013-recurrent-convolutional-neural-networks-discourse/). In Proceedings of the 2013 Workshop on Continuous Vector Space Models and their Compositionality.
17. 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 aaai conference on artificial intelligence, volume 32. [↗](https://doi.org/10.1609/aaai.v32i1.11701)
18. 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 NAACL-HLT, pages 515–520.
19. Yanran Li, Hui Su, Xiaoyu Shen, Wenjie Li, Ziqiang Cao, and Shuzi Niu. 2017. [Dailydialog: A manually labelled multi-turn dialogue dataset](https://md.research.elamur.ai/free/en/papers/li-2017-dailydialog-manually-labelled-multi-turn/). In Proceedings of the Eighth International Joint Conference on Natural Language Processing (Volume 1: Long Papers), pages 986–995. [↗](https://doi.org/10.48550/arxiv.1710.03957)
20. Zhaojiang Lin, Bing Liu, Seungwhan Moon, Paul Crook, Zhenpeng Zhou, Zhiguang Wang, Zhou Yu, Andrea Madotto, Eunjoon Cho, and Rajen Subba. 2021. Leveraging slot descriptions for zero-shot cross-domain dialogue StateTracking. In Proceedings of the 2021 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 5640–5648, Online. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Zhaojiang+Lin%2C+Bing+Liu%2C+Seungwhan+Moon%2C+Paul+Crook%2C+Zhenpeng+Zhou%2C+Zhiguang+Wang%2C+Zhou+Yu%2C+Andrea+Madotto%2C+Eunjoon+Cho%2C+and+Rajen+Subba%2E+2021%2E+Leveraging+slot+descriptions+for+zero%2Dshot+cross%2Ddomain+)
21. Yang Liu, Kun Han, Zhao Tan, and Yun Lei. 2017. [Using context information for dialog act classification in dnn framework](https://md.research.elamur.ai/free/en/papers/liu-2017-context-information-dialog-act-classification/). In Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing, pages 2170–2178.
22. Ekaterina Loginova and Dries Benoit. 2022. Structural information in mathematical formulas for exercise difficulty prediction: a comparison of NLP representations. In Proceedings of the 17th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2022), pages 101–106, Seattle, Washington. Association for Computational Linguistics. [↗](https://doi.org/10.48448/h3gp-dx88)
23. Neil Mercer. 2010. The analysis of classroom talk: Methods and methodologies. British journal of educational psychology, 80(1):1–14. [↗](https://doi.org/10.1348/000709909x479853)
24. Neil Mercer, Sara Hennessy, and Paul Warwick. 2019. Dialogue, thinking together and digital technology in the classroom: Some educational implications of a continuing line of inquiry. International Journal of Educational Research, 97:187–199. [↗](https://doi.org/10.4324/9780429400759-17)
25. Fei Mi, Yasheng Wang, and Yitong Li. 2022. Cins: Comprehensive instruction for few-shot learning in task-oriented dialog systems. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, pages 11076–11084. [find](https://scholar.google.com/scholar?q=Fei+Mi%2C+Yasheng+Wang%2C+and+Yitong+Li%2E+2022%2E+Cins%3A+Comprehensive+instruction+for+few%2Dshot+learning+in+task%2Doriented+dialog+systems%2E+In+Proceedings+of+the+AAAI+Conference+on+Artificial+Intelligence%2C+volu)
26. Fei Mi, Wanhao Zhou, Lingjing Kong, Fengyu Cai, Minlie Huang, and Boi Faltings. 2021a. Self-training improves pre-training for few-shot learning in task-oriented dialog systems. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 1887–1898. [find](https://scholar.google.com/scholar?q=Fei+Mi%2C+Wanhao+Zhou%2C+Lingjing+Kong%2C+Fengyu+Cai%2C+Minlie+Huang%2C+and+Boi+Faltings%2E+2021a%2E+Self%2Dtraining+improves+pre%2Dtraining+for+few%2Dshot+learning+in+task%2Doriented+dialog+systems%2E+In+Proceedings+of+the+)
27. Fei Mi, Wanhao Zhou, Lingjing Kong, Fengyu Cai, Minlie Huang, and Boi Faltings. 2021b. Self-training improves pre-training for few-shot learning in task-oriented dialog systems. In Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 1887–1898, Online and Punta Cana, Dominican Republic. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Fei+Mi%2C+Wanhao+Zhou%2C+Lingjing+Kong%2C+Fengyu+Cai%2C+Minlie+Huang%2C+and+Boi+Faltings%2E+2021b%2E+Self%2Dtraining+improves+pre%2Dtraining+for+few%2Dshot+learning+in+task%2Doriented+dialog+systems%2E+In+Proceedings+of+the+)
28. Erik G Miller, Nicholas E Matsakis, and Paul A Viola. 2000. Learning from one example through shared densities on transforms. In Proceedings IEEE Conference on Computer Vision and Pattern Recognition. CVPR 2000 (Cat. No. PR00662), volume 1, pages 464–471. IEEE. [find](https://scholar.google.com/scholar?q=Erik+G+Miller%2C+Nicholas+E+Matsakis%2C+and+Paul+A+Viola%2E+2000%2E+Learning+from+one+example+through+shared+densities+on+transforms%2E+In+Proceedings+IEEE+Conference+on+Computer+Vision+and+Pattern+Recognition%2E)
29. Eda Okur, Saurav Sahay, Roddy Fuentes Alba, and Lama Nachman. 2022. End-to-end evaluation of a spoken dialogue system for learning basic mathematics. In Proceedings of the 1st Workshop on Mathematical Natural Language Processing (MathNLP), pages 51–64, Abu Dhabi, United Arab Emirates (Hybrid). Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Eda+Okur%2C+Saurav+Sahay%2C+Roddy+Fuentes+Alba%2C+and+Lama+Nachman%2E+2022%2E+End%2Dto%2Dend+evaluation+of+a+spoken+dialogue+system+for+learning+basic+mathematics%2E+In+Proceedings+of+the+1st+Workshop+on+Mathematical)
30. Libo Qin, Wanxiang Che, Yangming Li, Mingheng Ni, and Ting Liu. 2020. [Dcr-net: A deep co-interactive relation network for joint dialog act recognition and sentiment classification](https://md.research.elamur.ai/free/en/papers/qin-2020-dcr-net-deep-co-interactive/). In Proceedings of the AAAI conference on artificial intelligence, volume 34, pages 8665–8672.
31. Libo Qin, Zhouyang Li, Wanxiang Che, Minheng Ni, and Ting Liu. 2021. [Co-gat: A co-interactive graph attention network for joint dialog act recognition and sentiment classification](https://md.research.elamur.ai/free/en/papers/qin-2020-co-gat-co-interactive-graph/). In Proceedings of the AAAI Conference on Artificial Intelligence, volume 35, pages 13709–13717. Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. Language models are unsupervised multitask learners.
32. 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. [↗](https://doi.org/10.18653/v1/n19-1373)
33. Nils Reimers and Iryna Gurevych. 2019. [Sentence-bert: Sentence embeddings using siamese bert-networks](https://md.research.elamur.ai/free/en/papers/reimers-2019-sentence-bert-sentence-embeddings-siamese/). 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. [↗](https://doi.org/10.18653/v1/d19-1410)
34. Pengzhen Ren, Yun Xiao, Xiaojun Chang, Po-Yao Huang, Zhihui Li, Brij B Gupta, Xiaojiang Chen, and Xin Wang. 2021. A survey of deep active learning. ACM computing surveys (CSUR), 54(9):1–40. [↗](https://doi.org/10.1145/3472291)
35. Lauren B Resnick, Sarah Michaels, and Catherine O’Connor. 2010. How (well structured) talk builds the mind. Innovations in educational psychology: Perspectives on learning, teaching and human development, pages 163–194. [find](https://scholar.google.com/scholar?q=Lauren+B+Resnick%2C+Sarah+Michaels%2C+and+Catherine+O%E2%80%99Connor%2E+2010%2E+How+%28well+structured%29+talk+builds+the+mind%2E+Innovations+in+educational+psychology%3A+Perspectives+on+learning%2C+teaching+and+human+developm)
36. Jia Tracy Shen, Michiharu Yamashita, Ethan Prihar, Neil Heffernan, Xintao Wu, Sean McGrew, and Dongwon Lee. 2021. Classifying math knowledge components via task-adaptive pre-trained bert. In Artificial Intelligence in Education: 22nd International Conference, AIED 2021, Utrecht, The Netherlands, June 14–18, 2021, Proceedings, Part I 22, pages 408–419. Springer. [find](https://scholar.google.com/scholar?q=Jia+Tracy+Shen%2C+Michiharu+Yamashita%2C+Ethan+Prihar%2C+Neil+Heffernan%2C+Xintao+Wu%2C+Sean+McGrew%2C+and+Dongwon+Lee%2E+2021%2E+Classifying+math+knowledge+components+via+task%2Dadaptive+pre%2Dtrained+bert%2E+In+Artificia)
37. Yu Song, Tianyong Hao, Zhinan Liu, and Zixin Lan. 2019. A systematic review of frameworks for coding towards classroom dialogue. In International Symposium on Emerging Technologies for Education, pages 226–236. Springer. [find](https://scholar.google.com/scholar?q=Yu+Song%2C+Tianyong+Hao%2C+Zhinan+Liu%2C+and+Zixin+Lan%2E+2019%2E+A+systematic+review+of+frameworks+for+coding+towards+classroom+dialogue%2E+In+International+Symposium+on+Emerging+Technologies+for+Education%2C+page)
38. Yu Song, Shunwei Lei, Tianyong Hao, Zixin Lan, and Ying Ding. 2021. Automatic classification of semantic content of classroom dialogue. Journal of Educational Computing Research, 59(3):496–521. [↗](https://doi.org/10.1177/0735633120968554)
39. Abhijit Suresh, Jennifer Jacobs, Charis Harty, Margaret Perkoff, James H. Martin, and Tamara Sumner. 2022a. The TalkMoves dataset: K-12 mathematics lesson transcripts annotated for teacher and student discursive moves. In Proceedings of the Thirteenth Language Resources and Evaluation Conference, pages 4654–4662, Marseille, France. European Language Resources Association. [find](https://scholar.google.com/scholar?q=Abhijit+Suresh%2C+Jennifer+Jacobs%2C+Charis+Harty%2C+Margaret+Perkoff%2C+James+H%2E+Martin%2C+and+Tamara+Sumner%2E+2022a%2E+The+TalkMoves+dataset%3A+K%2D12+mathematics+lesson+transcripts+annotated+for+teacher+and+student)
40. Abhijit Suresh, Jennifer Jacobs, Margaret Perkoff, James H. Martin, and Tamara Sumner. 2022b. Fine-tuning transformers with additional context to classify discursive moves in mathematics classrooms. In Proceedings of the 17th Workshop on Innovative Use of NLP for Building Educational Applications (BEA 2022), pages 71–81, Seattle, Washington. Association for Computational Linguistics. [find](https://scholar.google.com/scholar?q=Abhijit+Suresh%2C+Jennifer+Jacobs%2C+Margaret+Perkoff%2C+James+H%2E+Martin%2C+and+Tamara+Sumner%2E+2022b%2E+Fine%2Dtuning+transformers+with+additional+context+to+classify+discursive+moves+in+mathematics+classrooms%2E+I)
41. Lewis Tunstall, Nils Reimers, Unso Eun Seo Jo, Luke Bates, Daniel Korat, Moshe Wasserblat, and Oren Pereg. 2022. Efficient few-shot learning without prompts. arXiv preprint arXiv:2209.11055. [↗](https://arxiv.org/abs/2209.11055)
42. Dong Wang, Ziran Li, Haitao Zheng, and Ying Shen. 2020. [Integrating user history into heterogeneous graph for dialogue act recognition](https://md.research.elamur.ai/free/en/papers/wang-2020-integrating-user-history-into-heterogeneous/). In Proceedings of the 28th International Conference on Computational Linguistics, pages 4211–4221.
43. Viktor Wendel and Johannes Konert. 2016. Multiplayer serious games. In Serious Games, pages 211–241. Springer. [↗](https://doi.org/10.1007/978-3-319-40612-1_8)
44. Chien-Sheng Wu, Steven CH Hoi, Richard Socher, and Caiming Xiong. 2020. [Tod-bert: Pre-trained natural language understanding for task-oriented dialogue](https://md.research.elamur.ai/free/en/papers/wu-2020-tod-bert-pre-trained-natural/). In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 917–929.
45. Yizhe Zhang, Siqi Sun, Michel Galley, Yen-Chun Chen, Chris Brockett, Xiang Gao, Jianfeng Gao, Jingjing Liu, and William B Dolan. 2020. [Dialogpt: Large-scale generative pre-training for conversational response generation](https://md.research.elamur.ai/free/en/papers/zhang-2019-dialogpt-large-scale-generative-pre/). In Proceedings of the 58th Annual Meeting of the Association for Computational Linguistics: System Demonstrations, pages 270–278. [↗](https://doi.org/10.48550/arxiv.1911.00536)

*Updated 1 November 2026*