Elamur Research⌕
☰
home / papers / 2026 / From Documents to Segments: A Contextual Reformulation for Topic Assignment

From Documents to Segments: A Contextual Reformulation for Topic Assignment

Paper · 2026 · Hoonsang Yoon, Takyoung Kim, Wonkee Lee et al.

Summary

Standard topic models treat a whole document as the unit that receives topics. Many real texts, such as reviews or survey answers, mix several themes in a few sentences. This causes "topic contamination": a retrieved topic drags in unrelated content.

The authors propose segment-based topic allocation (SBTA). Topics are attached to short, coherent spans called segments, each covering one theme or a small set of tightly linked ones. They formalize the task and build SemEval-STM, a benchmark derived from the SemEval aspect-based sentiment data (Pontiki et al., 2016). In SemEval-STM, documents were split into topical segments by LLMs and then corrected by humans. They also adapt the classic word intrusion test to segments, so humans and models can judge coherence at the level where topics are actually assigned.

Experiments on laptop and restaurant reviews show that SBTA produces purer, more separable topic clusters. LLM-based methods usually beat LDA and BERTopic, although small models sometimes win in specific domains. The framework supports fine-grained analysis of multi-theme corpora.

Key points

  • SemEval-STM is released on Hugging Face (hoonst/SemEval-STM). It covers Laptop and Restaurant domains and provides segment-level topic annotations.
  • The segment intrusion evaluation has four variants: single vs. double intruder (SI, DI) at easy and hard levels (E, H). The hard variants and the two-intruder variants are harder for both models and humans.
  • Human upper bound: two annotators (paper authors) each labeled 50 samples per task, 200 annotations each. Most LLMs remained below human F1, and small models (llama-3.2-3b, qwen-2.5-7b, llama-4-maverick-17b) struggled considerably.
  • Results differed by evaluation setting:
    • Label-based: large LLMs (claude-3.7-sonnet, llama-3.3-70b, deepseek-v3) led, while LDA and BERTopic lagged.
    • Label-free: BERTopic and llama-3.2-3b did best on Laptop, and GPT variants did best on Restaurant.
  • Limitations:
    • Segmentation depends on LLMs.
    • Word co-occurrence coherence metrics fit poorly with span-level topics.
    • The dataset consists mostly of short documents.

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 Illinois Urbana-ChampaignUrbana · United States184 works · 69,372 citations🏢LG AI ResearchSeoul · South Korea7 works · 374 citations

 Fields, topics and concepts

📚 Shelves 6The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArtificial Intelligence · 16,194 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksSocial Sciences · 1,772 worksGeneral Social Sciences · 227 works

💡 Concepts 1Key ideas and methods explained on our concept pages.Topic modeling · 47 works

Cite

Cite as Hoonsang Yoon, Takyoung Kim, Wonkee Lee, Ilmin Cho, Dilek Hakkani-Tür and Stanley Jungkyu Choi. 2026. From Documents to Segments: A Contextual Reformulation for Topic Assignment. Findings of the Association for Computational Linguistics: ACL 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.findings-acl.1278

@inproceedings{yoon2026documents,
  title = {{From Documents to Segments: A Contextual Reformulation for Topic Assignment}},
  author = {Hoonsang Yoon, Takyoung Kim, Wonkee Lee, Ilmin Cho, Dilek Hakkani-Tür and Stanley Jungkyu Choi},
  booktitle = {Findings of the Association for Computational Linguistics: ACL 2026},
  year = {2026},
  doi = {10.18653/v1/2026.findings-acl.1278},
  url = {https://doi.org/10.18653/v1/2026.findings-acl.1278}
}

References

43 references, 11 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Enrique Amigó, Julio Gonzalo, Javier Artiles, and Felisa Verdejo. 2009. A comparison of extrinsic clustering evaluation metrics based on formal constraints. Inf. Retr., 12(4):461–486. find
  2. David Andrzejewski and Xiaojin Zhu. 2009. Latent Dirichlet Allocation with topic-in-set knowledge. In Proceedings of the NAACL HLT 2009 Workshop on Semi-supervised Learning for Natural Language Processing, pages 43–48, Boulder, Colorado. Association for Computational Linguistics. find
  3. Sebastian Arnold, Rudolf Schneider, Philippe Cudré-Mauroux, Felix A. Gers, and Alexander Löser. 2019. SECTOR: A neural model for coherent topic segmentation and classification. Transactions of the Association for Computational Linguistics, 7:169–184.
  4. Christian Baden, Christian Pipal, Martijn Schoonvelde, and Mariken van der Velden. 2021. Three gaps in computational text analysis methods for social sciences: A research agenda. Communication Methods and Measures, 16:1–18. find
  5. Shraey Bhatia, Jey Han Lau, and Timothy Baldwin. 2018. Topic intrusion for automatic topic model evaluation. In Proceedings of the 2018 Conference on Empirical Methods in Natural Language Processing, pages 844–849, Brussels, Belgium. Association for Computational Linguistics. find
  6. David M. Blei, Andrew Y. Ng, and Michael I. Jordan. 2003. Latent dirichlet allocation. J. Mach. Learn. Res., 3(null):993–1022.
  7. Tadeusz Caliński and Jerzy Harabasz. 1974. A dendrite method for cluster analysis. Communications in Statistics-theory and Methods, 3(1):1–27. find
  8. Jonathan Chang, Sean Gerrish, Chong Wang, Jordan Boyd-graber, and David Blei. 2009. Reading tea leaves: How humans interpret topic models. In Advances in Neural Information Processing Systems, volume 22. Curran Associates, Inc.
  9. David L. Davies and Donald W. Bouldin. 1979. A cluster separation measure. IEEE Transactions on Pattern Analysis and Machine Intelligence, PAMI-1(2):224–227. find
  10. Adji B. Dieng, Francisco J. R. Ruiz, and David M. Blei. 2020. Topic modeling in embedding spaces. Transactions of the Association for Computational Linguistics, 8:439–453.
  11. Tomoki Doi, Masaru Isonuma, and Hitomi Yanaka. 2024. Topic modeling for short texts with large language models. In Proceedings of the 62nd Annual Meeting of the Association for Computational Linguistics (Volume 4: Student Research Workshop), pages 21–33, Bangkok, Thailand. Association for Computatio… find
  12. Joseph C Dunn. 1974. Well-separated clusters and optimal fuzzy partitions. Journal of cybernetics, 4(1):95– 104. find
  13. Yaxin Fan, Feng Jiang, Peifeng Li, and Haizhou Li. 2024. Uncovering the potential of chatgpt for discourse analysis in dialogue: An empirical study. Preprint, arXiv:2305.08391. find
  14. Iacopo Ghinassi. 2021. Unsupervised text segmentation via deep sentence encoders: a first step towards a common framework for text-based segmentation, summarization and indexing of media content. find
  15. Maarten Grootendorst. 2022. Bertopic: Neural topic modeling with a class-based tf-idf procedure. Preprint, arXiv:2203.05794.
  16. Marti A. Hearst. 1994. Multi-paragraph segmentation of expository text. In Proceedings of the 32nd Annual Meeting on Association for Computational Linguistics, ACL ’94, page 9–16, USA. Association for Computational Linguistics. find
  17. Alexander Miserlis Hoyle, Rupak Sarkar, Pranav Goel, and Philip Resnik. 2022. Are neural topic models broken? In Findings of the Association for Computational Linguistics: EMNLP 2022, pages 5321–5344, Abu Dhabi, United Arab Emirates. Association for Computational Linguistics. find
  18. Lawrence Hubert and Phipps Arabie. 1985. Comparing partitions. Journal of Classification, 2(1):193–218. find
  19. Damir Korenčić, Strahil Ristov, and Jan Šnajder. 2018. Document-based topic coherence measures for news media text. Expert Systems with Applications, 114:357–373. find
  20. Diego Kozlowski, Carolina Pradier, and Pierre Benz. 2024. Generative ai for automatic topic labelling. Preprint, arXiv:2408.07003. — [arXiv:2408.07003](https://arxiv.org/abs/2408.07003) find
  21. Jey Han Lau, David Newman, and Timothy Baldwin. 2014. Machine reading tea leaves: Automatically evaluating topic coherence and topic model quality. In Proceedings of the 14th Conference of the European Chapter of the Association for Computational Linguistics, pages 530–539, Gothenburg, Sweden. Assoc…
  22. Ujjwal Maulik and Sanghamitra Bandyopadhyay. 2003. Performance evaluation of some clustering algorithms and validity indices. IEEE Transactions on pattern analysis and machine intelligence, 24(12):1650–1654. find
  23. Andrew Kachites McCallum. 2002. Mallet: A machine learning for language toolkit. Http://www.cs.umass.edu/ mccallum/mallet. find
  24. Qiaozhu Mei, Xuehua Shen, and ChengXiang Zhai. 2007. Automatic labeling of multinomial topic models. In Knowledge Discovery and Data Mining. find
  25. Stephen Merity, Nitish Shirish Keskar, and Richard Socher. 2018. Regularizing and optimizing LSTM language models. In International Conference on Learning Representations.
  26. David Mimno, Hanna Wallach, Edmund Talley, Miriam Leenders, and Andrew McCallum. 2011. Optimizing semantic coherence in topic models. In Proceedings of the 2011 Conference on Empirical Methods in Natural Language Processing, pages 262–272, Edinburgh, Scotland, UK. Association for Computational Lingu…
  27. David Newman, Youn Noh, Edmund Talley, Sarvnaz Karimi, and Timothy Baldwin. 2010. Evaluating topic models for digital libraries. In Proceedings of the 10th annual joint conference on Digital libraries, pages 215–224. find
  28. Sergey I. Nikolenko. 2016. Topic quality metrics based on distributed word representations. In Proceedings of the 39th International ACM SIGIR Conference on Research and Development in Information Retrieval, SIGIR ’16, page 1029–1032, New York, NY, USA. Association for Computing Machinery. find
  29. Chau Minh Pham, Alexander Hoyle, Simeng Sun, Philip Resnik, and Mohit Iyyer. 2024. TopicGPT: A prompt-based topic modeling framework. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Pape…
  30. Maria Pontiki, Dimitris Galanis, Haris Papageorgiou, Ion Androutsopoulos, Suresh Manandhar, Mohammad AL-Smadi, Mahmoud Al-Ayyoub, Yanyan Zhao, Bing Qin, Orphée De Clercq, Véronique Hoste, Marianna Apidianaki, Xavier Tannier, Natalia Loukachevitch, Evgeniy Kotelnikov, Nuria Bel, Salud María Jiménez-Z… — SemEval-2016 Task 5: Aspect Based Sentiment Analysis
  31. Hamed Rahimi, David Mimno, Jacob Hoover, Hubert Naacke, Camelia Constantin, and Bernd Amann. 2024. Contextualized topic coherence metrics. In Findings of the Association for Computational Linguistics: EACL 2024, pages 1760–1773, St. Julian’s, Malta. Association for Computational Linguistics. find
  32. Nitin Ramrakhiyani, Sachin Pawar, Swapnil Hingmire, and Girish Palshikar. 2017. Measuring topic coherence through optimal word buckets. In Proceedings of the 15th Conference of the European Chapter of the Association for Computational Linguistics: Volume 2, Short Papers, pages 437–442, Valencia, Spa… find
  33. Martin Riedl and Chris Biemann. 2012. How text segmentation algorithms gain from topic models. In Proceedings of the 2012 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies, pages 553–557, Montréal, Canada. Association for Computati… find
  34. Michael Röder, Andreas Both, and Alexander Hinneburg. 2015. Exploring the space of topic coherence measures. In Proceedings of the Eighth ACM International Conference on Web Search and Data Mining, WSDM ’15, page 399–408, New York, NY, USA. Association for Computing Machinery.
  35. Peter J Rousseeuw. 1987. Silhouettes: a graphical aid to the interpretation and validation of cluster analysis. Journal of computational and applied mathematics, 20:53–65. find
  36. C. E. Shannon. 1948. A mathematical theory of communication. The Bell System Technical Journal, 27(3):379–423. find
  37. Akash Srivastava and Charles Sutton. 2017. Autoencoding variational inference for topic models. In International Conference on Learning Representations. find
  38. Dominik Stammbach, Vilém Zouhar, Alexander Hoyle, Mrinmaya Sachan, and Elliott Ash. 2023. Revisiting automated topic model evaluation with large language models. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 9348–9357, Singapore. Association for Com… find
  39. Alexander Strehl and Joydeep Ghosh. 2002. Cluster ensembles - a knowledge reuse framework for combining partitionings. J Mach Learn Res, 3. find
  40. Yee Whye Teh, Michael I Jordan, Matthew J Beal, and David M Blei and. 2006. Hierarchical dirichlet processes. Journal of the American Statistical Association, 101(476):1566–1581. find
  41. Nguyen Xuan Vinh, Julien Epps, and James Bailey. 2009. Information theoretic measures for clusterings comparison: is a correction for chance necessary? In Proceedings of the 26th Annual International Conference on Machine Learning, ICML ’09, page 1073–1080, New York, NY, USA. Association for Computi… find
  42. Selma Wanna, Nicholas Solovyev, Ryan Barron, Maksim E. Eren, Manish Bhattarai, Kim Ø. Rasmussen, and Boian S. Alexandrov. 2024. Topictag: Automatic annotation of nmf topic models using chain of thought and prompt tuning with llms. In Proceedings of the ACM Symposium on Document Engineering 2024, Doc… find
  43. Xuanli Lisa Xie and Gerardo Beni. 1991. A validity measure for fuzzy clustering. IEEE Transactions on Pattern Analysis & Machine Intelligence, 13(08):841– 847. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/yoon-2026-documents-segments-contextual-reformulation-topic/