Elamur Research⌕
☰
home / papers / 2026 / POINTERS at UZH Shared Task 2026: Reasoning Probes for Argumentation Mining in UN Resolutions

POINTERS at UZH Shared Task 2026: Reasoning Probes for Argumentation Mining in UN Resolutions

Paper · 2026 · Sohom Sen, Avina Nakarmi, Xun Song et al.

Summary

The paper presents team POINTERS' entry to the reasoning-trace track of the UZH ArgMining 2026 Shared Task. The task is to recover argument structure in UNESCO ICPE resolutions, which come in French with English translations. Systems must label each paragraph as preambular or operative, assign education-dimension tags, and predict directed relations between paragraphs. The authors treat each resolution as evidence (preambular clauses) backing claims (operative clauses), drawing on Toulmin's model.

The system runs in two stages on a single consumer GPU with no cloud calls. First, a local Qwen3-8B-GGUF model classifies paragraphs and must cite quoted phrases to justify each decision. Second, sentence-transformer retrieval proposes candidate pairs. The model then labels each pair, using four reasoning strategies (causal, corroboration, contrastive, triangulation) as a diagnostic scaffold.

No gold test labels were available, so Claude Sonnet 4.6 judged trace quality, giving roughly 77/100 on test. The work shows that small open models can produce grounded explanations of diplomatic text. However, fine-grained relation labels, especially complemental versus modifying, remain error-prone.

Key points

  • Data: 2,694 training resolutions spanning ICPE proceedings from 1934 onward, plus 89 unannotated test resolutions.
  • Tags: a 126-code tag vocabulary (e.g., ACT_IO, LAW_INTER, POL_EQUIT). Tagging was the weakest sub-component.
  • Retrieval and prompting:
    • all-MiniLM-L6-v2 retrieves the top-15 neighbours for each paragraph (cosine ≥ 0.03).
    • Step 1 uses a context window of about 21 paragraphs.
    • Each relation trace must name one rejected relation type. This curbed the model's tendency to over-predict complemental.
  • Models: Qwen3-8B-GGUF (Q8_0 via llama-cpp-python) and Llama-3.1-8B-Instruct, both run on an RTX 4080 (16 GB) with temperature 0.1.
  • Judging: Specificity was the strongest judged criterion, and Depth & Significance varied most. The judge scores reflect trace quality, not official F1.

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

🎓New Jersey Institute of TechnologyNewark · United States15 works · 1,333 citations

 Fields, topics and concepts

📚 Shelves 7The 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 worksPhilosophy · 284 worksPolitical Science and International Relations · 182 works

🧭 Topics 1Research subjects and languages the work is about.Reasoning · 1,426 works

Cite

Cite as Sohom Sen, Avina Nakarmi, Xun Song and Aritra Dasgupta. 2026. POINTERS at UZH Shared Task 2026: Reasoning Probes for Argumentation Mining in UN Resolutions. Proceedings of the 13th Workshop on Argument Mining and Reasoning.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.argmining-1.16

@inproceedings{sen2026pointers,
  title = {{POINTERS at UZH Shared Task 2026: Reasoning Probes for Argumentation Mining in UN Resolutions}},
  author = {Sohom Sen, Avina Nakarmi, Xun Song and Aritra Dasgupta},
  booktitle = {Proceedings of the 13th Workshop on Argument Mining and Reasoning},
  year = {2026},
  doi = {10.18653/v1/2026.argmining-1.16},
  url = {https://doi.org/10.18653/v1/2026.argmining-1.16}
}

References

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

  1. Bonnie J. Breitmayer, Lioness Ayres, and Kathleen A. Knafl. 1993. Triangulation in qualitative research: Evaluation of completeness and confirmation purposes. Image: The Journal of Nursing Scholarship, 25(3):237–243. find
  2. Sarah K. Brem and Lance J. Rips. 2000. Explanation and evidence in informal argument. Cognitive Science, 24(4):573–604. find
  3. Elena Cabrio and Serena Villata. 2018. Five years of argument mining: A data-driven analysis. In IJCAI, volume 18, pages 5427–5433. find
  4. Giuseppina Scotto Di Carlo. 2013. Vagueness as a political strategy: Weasel words in security council resolutions relating to the second gulf war. Cambridge Scholars Publishing. find
  5. Yingqiang Gao, Fabian Winiger, Patrick Montjourides, Anastassia Shaitarova, Nianlong Gu, Simon Peng-Keller, and Gerold Schneider. 2025. SpiritRAG: A Q&A System for Religion and Spirituality in the United Nations Archive. In Proceedings of the 2025 Conference on Empirical Methods in Natural Language… find
  6. David Godden. 2019. Corroboration: Sensitivity, safety, and explanation. Acta Analytica, 34(1):15–38. find
  7. Aaron Grattafiori, Abhimanyu Dubey, Abhinav Jauhri, Abhinav Pandey, Abhishek Kadian, Ahmad Al-Dahle, Aiesha Letman, Akhil Mathur, Alan Schelten, Alex Vaughan, and 1 others. 2024. The llama 3 herd of models. arXiv preprint arXiv:2407.21783.
  8. Ivan Habernal and Iryna Gurevych. 2017. Argumentation mining in user-generated web discourse. Computational linguistics, 43(1):125–179.
  9. Marco Lippi and Paolo Torroni. 2016. Argumentation mining: State of the art and emerging trends. ACM Transactions on Internet Technology (TOIT), 16(2):1– 25. find
  10. Peter Lipton. 2013. Inference to the best explanation. In The Routledge Companion to Philosophy of Science, pages 225–234. Routledge. find
  11. Yang Liu, Dan Iter, Yichong Xu, Shuohang Wang, Ruochen Xu, and Chenguang Zhu. 2023. G-eval: NLG evaluation using gpt-4 with better human alignment. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing, pages 2511–2522, Singapore. Association for Computational Lin…
  12. Joshua Maynez, Shashi Narayan, Bernd Bohnet, and Ryan McDonald. 2020. On faithfulness and factuality in abstractive summarization. In Proceedings of the 58th annual meeting of the association for computational linguistics, pages 1906–1919.
  13. Raquel Mochales Palau and Marie-Francine Moens. 2009. Argumentation mining: the detection, classification and structure of arguments in text. In Proceedings of the 12th international conference on artificial intelligence and law, pages 98–107. find
  14. Judea Pearl. 2009. Causality. Cambridge University Press. find
  15. 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.
  16. Christian Stab and Iryna Gurevych. 2014. Annotating argument components and relations in persuasive essays. In Proceedings of COLING 2014, the 25th international conference on computational linguistics: Technical papers, pages 1501–1510.
  17. Manfred Stede, Jodi Schneider, and Graeme Hirst. 2019. Argumentation mining. Springer. find
  18. Qwen Team. 2025. Qwen3 technical report. Preprint, arXiv:2505.09388.
  19. Stephen E. Toulmin. 2003. The Uses of Argument, updated edition. Cambridge University Press. find
  20. Sahaj Vaidya and Aritra Dasgupta. 2020. Knowing what to look for: A fact-evidence reasoning framework for decoding communicative visualization. In 2020 IEEE Visualization Conference (VIS), pages 231–235. IEEE. find
  21. Henning Wachsmuth, Nona Naderi, Yufang Hou, Yonatan Bilu, Vinodkumar Prabhakaran, Tim Alberd-ingk Thijm, Graeme Hirst, and Benno Stein. 2017. Computational argumentation quality assessment in natural language. In Proceedings of the 15th Conference of the European Chapter of the Association for Compu…
  22. Douglas Walton and Chris A Reed. 2005. Argumentation schemes and enthymemes. Synthese, 145(3):339–370. find
  23. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, and 1 others. 2022. Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35:24824– 24837.
  24. Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric Xing, and 1 others. 2023. Judging llm-as-a-judge with mt-bench and chatbot arena. Advances in neural information processing systems, 36:46595–46623.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/sen-2026-pointers-uzh-shared-task-2026/