Elamur Research⌕
☰
home / papers / 2026 / UniCite: A Dataset and Unified Hierarchical Taxonomy for Multi-Dimensional Citation Analysis

UniCite: A Dataset and Unified Hierarchical Taxonomy for Multi-Dimensional Citation Analysis

Paper · 2026 · Amina Mourky, Elena Leitner, Julian Moreno-Schneider et al.

Summary

Citation Context Analysis (CCA) is fragmented. Competing taxonomies range from three to more than twelve classes and have no mappings between them. Function, sentiment and importance are usually modelled in isolation, and most benchmarks cover papers from before 2019. The authors propose UniCite, a hierarchical scheme that merges the categories of Teufel et al. (2006), Jurgens-CFC and SciCite, with compatibility to MultiCite. Its Level 1 tier holds six citation functions (Motivation, Background, Methodology, Comparison, Extension and Future Work), and beneath these sit twelve Level 2 subcategories. Sentiment and importance are annotated as separate, independent dimensions. Existing datasets were converted automatically at Level 1. A combined set of 4,017 citations, including newly collected contexts from 2018–2024 papers, received expert labels on every dimension. SciBERT-based multi-task models, especially a hierarchical variant that feeds function predictions into the subfunction classifier, beat single-task baselines on fine-grained subfunction labelling. The work offers a common reference scheme and a more up-to-date resource, and it shows that jointly modelling related citation tasks pays off.

Key points

  • The dataset has 4,017 citations, 1,547 of them newly extracted from 2018–2024 publications. It is split 70/15/15: 2,814 train, 602 validation and 601 test.
  • Direct mappings cover 89% of conversions from existing schemes. Only MultiCite's "Similarities" and "Differences" had to be merged into Comparison.
  • Multi-task learning gives a 21.1% relative gain in subfunction classification over single-task training.
  • In the ablation, hierarchical MTL with function conditioning reaches 65.3% macro F1, versus 50.9% without conditioning. Replacing focal loss with standard cross-entropy drops it to 56.6%.
  • All models use allenai/scibert_scivocab_uncased on a Tesla T4 GPU. Data and code are released at github.com/aminamourky/UniCite.

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

🔬German Research Centre for Artificial IntelligenceKaiserslautern · Germany77 works · 29,479 citations

 Fields, topics and concepts

📚 Shelves 8The field of science this work belongs to (OpenAlex fields).Computer Science · 16,667 worksArts and Humanities · 16,039 worksLanguage and Linguistics · 15,891 worksSocial Sciences · 1,772 worksInformation Systems · 1,437 worksDecision Sciences · 209 worksGeneral Social Sciences · 227 worksStatistics, Probability and Uncertainty · 61 works

Cite

Cite as Amina Mourky, Elena Leitner, Julian Moreno-Schneider, Raia Abu Ahmad, Ekaterina Borisova and Georg Rehm. 2026. UniCite: A Dataset and Unified Hierarchical Taxonomy for Multi-Dimensional Citation Analysis. Proceedings of Natural Scientific Language Processing (NSLP) @ LREC 2026.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.63317/472sud9pm8t5

@inproceedings{mourky2026unicite,
  title = {{UniCite: A Dataset and Unified Hierarchical Taxonomy for Multi-Dimensional Citation Analysis}},
  author = {Amina Mourky, Elena Leitner, Julian Moreno-Schneider, Raia Abu Ahmad, Ekaterina Borisova and Georg Rehm},
  booktitle = {Proceedings of Natural Scientific Language Processing (NSLP) @ LREC 2026},
  year = {2026},
  doi = {10.63317/472sud9pm8t5},
  url = {https://doi.org/10.63317/472sud9pm8t5}
}

References

31 references, 7 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Ron Artstein and Massimo Poesio. 2008. Survey article: Inter-coder agreement for computational linguistics. Computational Linguistics, 34(4):555– 596.
  2. Yasa M. Baig, Alex X. Oesterling, Rui Xin, Haoyang Yu, Angikar Ghosal, Lesia Semenova, and Cynthia Rudin. 2021. Multitask learning for citation purpose classification. In Proceedings of the Second Workshop on Scholarly Document Processing, pages 134–139, Online. Association for Computational Linguis… find
  3. Iz Beltagy, Kyle Lo, and Arman Cohan. 2019. SciB-ERT: A pretrained language model for scientific text. 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 3615–3620,… find
  4. Francisco Bolaños, Angelo Salatino, Francesco Osborne, and Enrico Motta. 2025. Modelling and classifying the components of a literature review. arXiv preprint arXiv:2508.04337. — [arXiv:2508.04337](https://arxiv.org/abs/2508.04337) find
  5. Arman Cohan, Waleed Ammar, Madeleine van Zuylen, and Field Cady. 2019. Structural scaffolds for citation intent classification in scientific publications. In Proceedings of the 2019 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies…
  6. Marilena Daquino, Silvio Peroni, and David Shotton. 2020. The opencitations data model. In The Semantic Web – ISWC 2020, volume 12507 of Lecture Notes in Computer Science, Cham. Springer. find
  7. This work was supported by the consortium NFDI for Data Science and Artificial Intelligence https://www.nfdi4datascience.de ↗
  8. Changxu Duan and Zhiyin Tan. 2026. Semantically orthogonal framework for citation classification: Disentangling intent and content. arXiv preprint. find
  9. Alex Fogelson, Ana Trišović, and Neil Thompson. 2025. Llms in citation intent classification: Progress, precision, and reproducibility challenges. In Proceedings of the 3rd ACM Conference on Reproducibility and Replicability, ACM REP ’25, page 250–253, New York, NY, USA. Association for Computing Ma… find
  10. Eugene Garfield, Mary Elizabeth Stevens, and Vincent E. Giuliano. 1964. Can citation indexing be automated? In Statistical Association Methods for Mechanical Documentation, Symposium Proceedings, volume 269, pages 189–196. National Bureau of Standards. Accessed August 4, 2025. find
  11. Myriam Hernández-Álvarez, José M. Gómez Soriano, and Patricio Martínez-Barco. 2017. Citation function, polarity and influence classification. Natural Language Engineering, 23(4):561–588. find
  12. Lasse M. Jantsch, Dong-Jae Koh, Seonghwan Yoon, Jisu Lee, Anne Lauscher, and Young-Kyoon Suh. 2025. FineCite: A novel approach for fine-grained citation context analysis. In Findings of the Association for Computational Linguistics: ACL 2025, pages 24525–24542, Vienna, Austria. Association for Compu… find
  13. X. Jiang. 2025. Ensembling approaches to citation function classification and important citation screening. Scientometrics, 130:1371–1419. find
  14. Charles Jochim and Hinrich Schütze. 2012. Towards a generic and flexible citation classifier based on a faceted classification scheme. In Proceedings of COLING 2012, pages 1343–1358, Mumbai, India. find
  15. The COLING 2012 Organizing Committee. find
  16. David Jurgens, Srijan Kumar, Raine Hoover, Dan McFarland, and Dan Jurafsky. 2018. Measuring the evolution of a scientific field through citation frames. Transactions of the Association for Computational Linguistics, 6:391–406.
  17. Jan-Christoph Klie, Michael Bugert, Beto Boullosa, Richard Eckart de Castilho, and Iryna Gurevych. 2018. The INCEpTION platform: Machine-assisted and knowledge-oriented interactive annotation. In Proceedings of the 27th International Conference on Computational Linguistics: System Demonstrations, pa…
  18. Paris Koloveas, Serafeim Chatzopoulos, Thanasis Vergoulis, and Christos Tryfonopoulos. 2025. Can llms predict citation intent? an experimental analysis of in-context learning and fine-tuning on open llms. find
  19. Suchetha N. Kunnath, David Pride, and Petr Knoth. 2023. Prompting strategies for citation classification. In Proceedings of the 32nd ACM International Conference on Information and Knowledge Management, pages 1127–1137. find
  20. Anne Lauscher, Brandon Ko, Bailey Kuehl, Sophie Johnson, Arman Cohan, David Jurgens, and Kyle Lo. 2022. MultiCite: Modeling realistic citations requires moving beyond the single-sentence single-label setting. In Proceedings of the 2022 Conference of the North American Chapter of the Association for… find
  21. Tsung-Yi Lin, Priya Goyal, Ross Girshick, Kaiming He, and Piotr Dollár. 2017. Focal loss for dense object detection. In Proceedings of the IEEE International Conference on Computer Vision (ICCV), pages 2999–3007.
  22. Michael J. Moravcsik and P. Murugesan. 1979. Citation patterns in scientific revolutions. Scientometrics, 1(2):161–169. Accessed August 5, 2025. find
  23. Roman Plaud, Matthieu Labeau, Antoine Saillenfest, and Thomas Bonald. 2024. Revisiting hierarchical text classification: Inference and metrics. In Proceedings of the 28th Conference on Computational Natural Language Learning, pages 231–242, Miami, FL, USA. Association for Computational Linguistics. find
  24. David Pride and Petr Knoth. 2020. An authoritative approach to citation classification. In Proceedings of the ACM/IEEE Joint Conference on Digital Libraries in 2020, JCDL ’20, pages 337–340, New York, NY, USA. Association for Computing Machinery. find
  25. Sebastian Ruder. 2017. An overview of multi-task learning in deep neural networks. arXiv preprint arXiv:1706.05098.
  26. Carlos N. Silla and Alex A. Freitas. 2011. A survey of hierarchical classification across different application domains. Data Mining and Knowledge Discovery, 22(1-2):31–72. find
  27. Trevor Standley, Amir R. Zamir, Dawn Chen, Leonidas J. Guibas, Jitendra Malik, and Silvio Savarese. 2019. Which tasks should be learned together in multi-task learning? arXiv preprint arXiv:1905.07553. Presented at ICML 2020. — [arXiv:1905.07553](https://arxiv.org/abs/1905.07553) find
  28. Xuan Su, Animesh Prasad, Min-Yen Kan, and Kazunari Sugiyama. 2019. Neural multi-task learning for citation function and provenance. In Proceedings of the 2019 ACM/IEEE Joint Conference on Digital Libraries (JCDL), pages 394–395. IEEE. find
  29. Category Model Architecture Base Model Max Sequence Length Hidden Size Dropout Training Configuration Batch Size Aixin Sun and Ee-Peng Lim. 2001. Hierarchical text classification and evaluation. In Proceedings of the 2001 IEEE International Conference on Data Mining (ICDM), pages 521–528. IEEE. find
  30. Learning Rate Optimizer Weight Decay Warmup Ratio John Swales. 1986. Citation analysis and discourse analysis. Applied Linguistics, 7(1):39–56. Accessed August 4, 2025. find
  31. Max Epochs Simone Teufel, Advaith Siddharthan, and Dan Tidhar. 2006. Automatic classification of citation function. In Proceedings of the 2006 Conference on Empirical Methods in Natural Language Processing, pages 103–110, Sydney, Australia. Association for Computational Linguistics.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/mourky-2026-unicite-dataset-unified-hierarchical-taxonomy/