Elamur Research⌕
☰
home / papers / 2025 / Debate-to-Write: A Persona-Driven Multi-Agent Framework for Diverse Argument Generation

Debate-to-Write: A Persona-Driven Multi-Agent Framework for Diverse Argument Generation

Paper · 2025 · Zhe Hu, Hou Pong Chan, Jing Li et al.

Summary

Large language models tend to write arguments that drift toward one averaged viewpoint. They also generate text token by token without an explicit plan, which limits both the range of perspectives and the logical structure of argumentative essays. The authors address this with a framework that runs entirely through prompting and is modelled on human debate.

First, an LLM builds a pool of topic-specific personas, each a short description paired with a claim. It then selects three personas and justifies each choice. These persona agents argue over several rounds with a critic agent that takes the opposing side and probes weak points. Once they reach agreement, the discussion is condensed into a high-level plan, and a writing module turns the plan into an essay. Because agents can revisit and revise earlier points, planning is nonlinear rather than a fixed sequential outline.

On counter-argument essays for controversial propositions, the framework produced more diverse and more persuasive output than the baselines, according to both automatic and human evaluation. The paper also introduces a metric for perspective diversity in long-form text.

Key points

  • The pipeline has three stages: persona assignment (N = 3 agents chosen from an LLM-generated pool), debate-based planning with an additional critic agent, and argument writing. No model training is required.
  • Test data consists of 64 propositions randomly sampled from idebate.net and reddit/CMV, covering domains such as politics, culture and education. For each, the model must write an essay refuting the proposition.
  • The backbone is ChatGPT (gpt-3.5-turbo-0301) at temperature 1.0. Baselines are E2E prompting, CoT-Plan and AMERICANO, plus ablations without persona assignment.
  • A new automatic perspective-diversity metric extracts key ideas, scores how unique each one is, and aggregates the scores. The metric is needed because existing lexical and semantic diversity measures do not capture viewpoint variety.
  • Code and data are available at github.com/Derekkk/LLM4ArgGen. The authors list three limitations: a small test set, no evidence retrieval, and testing on GPT-3.5 only. Exact scores were not in the supplied excerpt.

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

🏢Alibaba GroupHangzhou · China168 works · 15,386 citations🎓Hong Kong Polytechnic UniversityHong Kong · China84 works · 4,956 citations🏢Alibaba DAMO AcademyHangzhou · China39 works · 6,378 citations🏢Alibaba Group (Cayman Islands)George Town · Cayman Islands32 works · 3,389 citations🎓Case Western Reserve UniversityCleveland · United States5 works · 280 citations

 Fields, topics and concepts

📚 Shelves 4The 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 works

Cite

Cite as Zhe Hu, Hou Pong Chan, Jing Li and Yu Yin. 2025. Debate-to-Write: A Persona-Driven Multi-Agent Framework for Diverse Argument Generation. Proceedings of the 31st International Conference on Computational Linguistics.

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

@inproceedings{hu2025debate,
  title = {{Debate-to-Write: A Persona-Driven Multi-Agent Framework for Diverse Argument Generation}},
  author = {Zhe Hu, Hou Pong Chan, Jing Li and Yu Yin},
  booktitle = {Proceedings of the 31st International Conference on Computational Linguistics},
  year = {2025},
  url = {https://aclanthology.org/2025.coling-main.314/}
}

References

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

  1. Josh Achiam, Steven Adler, Sandhini Agarwal, Lama Ahmad, Ilge Akkaya, Florencia Leoni Aleman, Diogo Almeida, Janko Altenschmidt, Sam Altman, Shyamal Anadkat, et al. 2023. Gpt-4 technical report. arXiv preprint arXiv:2303.08774.
  2. Jianzhu Bao, Yasheng Wang, Yitong Li, Fei Mi, and Ruifeng Xu. 2022. AEG: Argumentative essay generation via a dual-decoder model with content planning. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 5134–5148, Abu Dhabi, United Arab Emirates. Associa… find
  3. Sébastien Bubeck, Varun Chandrasekaran, Ronen Eldan, Johannes Gehrke, Eric Horvitz, Ece Kamar, Peter Lee, Yin Tat Lee, Yuanzhi Li, Scott Lundberg, et al. 2023. Sparks of artificial general intelligence: Early experiments with gpt-4. arXiv preprint arXiv:2303.12712.
  4. Asli Celikyilmaz, Elizabeth Clark, and Jianfeng Gao. 2020. Evaluation of text generation: A survey. arXiv preprint arXiv:2006.14799. find
  5. Hou Pong Chan, Lu Wang, and Irwin King. 2021. Controllable summarization with constrained markov decision process. Trans. Assoc. Comput. Linguistics, 9:1213–1232.
  6. Yew Ken Chia, Pengfei Hong, Lidong Bing, and Soujanya Poria. 2023. Instructeval: Towards holistic evaluation of instruction-tuned large language models. arXiv preprint arXiv:2306.04757.
  7. Yilun Du, Shuang Li, Antonio Torralba, Joshua B Tenenbaum, and Igor Mordatch. 2023. Improving factuality and reasoning in language models through multiagent debate. arXiv preprint arXiv:2305.14325.
  8. Linda Flower and John R Hayes. 1981. A cognitive process theory of writing. College composition and communication, 32(4):365–387. find
  9. Seraphina Goldfarb-Tarrant, Tuhin Chakrabarty, Ralph Weischedel, and Nanyun Peng. 2020. Content planning for neural story generation with aristotelian rescoring. In Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing (EMNLP), pages 4319–4338, Online. Association fo… find
  10. Shirley Anugrah Hayati, Minhwa Lee, Dheeraj Rajagopal, and Dongyeop Kang. 2023. How far can we extract diverse perspectives from large language models? criteria-based diversity prompting! arXiv preprint arXiv:2311.09799. find
  11. Yuhang He, Jianzhu Bao, Yang Sun, Bin Liang, Min Yang, Bing Qin, and Ruifeng Xu. 2024. Decomposing argumentative essay generation via dialectical planning of complex reasoning. In Findings of the Association for Computational Linguistics ACL 2024, pages 12305–12322, Bangkok, Thailand and virtual mee… find
  12. Zhe Hu, Zhiwei Cao, Hou Pong Chan, Jiachen Liu, Xinyan Xiao, Jinsong Su, and Hua Wu. 2023. Controllable dialogue generation with disentangled multi-grained style specification and attribute consistency reward. IEEE ACM Trans. Audio Speech Lang. Process., 31:188–199. find
  13. Zhe Hu, Hou Pong Chan, and Lifu Huang. 2022a. MOCHA: A multi-task training approach for coherent text generation from cognitive perspective. In Proceedings of the 2022 Conference on Empirical Methods in Natural Language Processing, pages 10324–10334, Abu Dhabi, United Arab Emirates. Association for… find
  14. Zhe Hu, Hou Pong Chan, Jiachen Liu, Xinyan Xiao, Hua Wu, and Lifu Huang. 2022b. PLANET: Dynamic content planning in autoregressive transformers for long-form text generation. In Proceedings of the 60th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 228… find
  15. Zhe Hu, Hou Pong Chan, and Yu Yin. 2024. AMERI-CANO: Argument generation with discourse-driven decomposition and agent interaction. In Proceedings of the 17th International Natural Language Generation Conference, pages 82–102, Tokyo, Japan. Association for Computational Linguistics. find
  16. Xinyu Hua, Zhe Hu, and Lu Wang. 2019. Argument generation with retrieval, planning, and realization. In Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, pages 2661– 2672, Florence, Italy. Association for Computational Linguistics.
  17. Xinyu Hua and Lu Wang. 2018. Neural argument generation augmented with externally retrieved evidence. In Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers), pages 219–230, Melbourne, Australia. Association for Computational Linguistics. find
  18. Hang Jiang, Xiajie Zhang, Xubo Cao, and Jad Kabbara. 2023. Personallm: Investigating the ability of large language models to express big five personality traits. arXiv preprint arXiv:2305.02547. find
  19. Ralph Henry Johnson and J Anthony Blair. 2006. Logical self-defense. Idea. find
  20. Kyungha Kim, Sangyun Lee, Kung-Hsiang Huang, Hou Pong Chan, Manling Li, and Heng Ji. 2024. Can llms produce faithful explanations for fact-checking? towards faithful explainable fact-checking via multi-agent debate. CoRR, abs/2402.07401. — [arXiv:2402.07401](https://arxiv.org/abs/2402.07401) find
  21. Ruosen Li, Teerth Patel, and Xinya Du. 2023. Prd: Peer rank and discussion improve large language model based evaluations. arXiv preprint arXiv:2307.02762. find
  22. Tian Liang, Zhiwei He, Wenxiang Jiao, Xing Wang, Yan Wang, Rui Wang, Yujiu Yang, Zhaopeng Tu, and Shuming Shi. 2023. Encouraging divergent thinking in large language models through multi-agent debate. arXiv preprint arXiv:2305.19118.
  23. Xuechen Liang, Meiling Tao, Tianyu Shi, and Yiting Xie. 2024. CMAT: A multi-agent collaboration tuning framework for enhancing small language models. CoRR, abs/2404.01663. — [arXiv:2404.01663](https://arxiv.org/abs/2404.01663) find
  24. Benedetta Muscato, Chandana Sree Mala, Marta Marchiori Manerba, Gizem Gezici, and Fosca Giannotti. 2024. An overview of recent approaches to enable diversity in large language models through aligning with human perspectives. In Proceedings of the 3rd Workshop on Perspectivist Approaches to NLP (NLPe… find
  25. Long Ouyang, Jeffrey Wu, Xu Jiang, Diogo Almeida, Carroll Wainwright, Pamela Mishkin, Chong Zhang, Sandhini Agarwal, Katarina Slama, Alex Ray, et al. 2022. Training language models to follow instructions with human feedback. Advances in neural information processing systems, 35:27730–27744.
  26. Vishakh Padmakumar and He He. 2024. Does writing with language models reduce content diversity? In The Twelfth International Conference on Learning Representations. find
  27. Taylor Sorensen, Liwei Jiang, Jena D Hwang, Sydney Levine, Valentina Pyatkin, Peter West, Nouha Dziri, Ximing Lu, Kavel Rao, Chandra Bhagavatula, et al. 2024. Value kaleidoscope: Engaging ai with pluralistic human values, rights, and duties. In Proceedings of the AAAI Conference on Artificial Intell…
  28. Yongqi Tong, Yifan Wang, Dawei Li, Sizhe Wang, Zi Lin, Simeng Han, and Jingbo Shang. 2023. Eliminating reasoning via inferring with planning: A new framework to guide llms’ non-linear thinking. arXiv preprint arXiv:2310.12342. — [arXiv:2310.12342](https://arxiv.org/abs/2310.12342) find
  29. Hugo Touvron, Louis Martin, Kevin Stone, Peter Albert, Amjad Almahairi, Yasmine Babaei, Nikolay Bashlykov, Soumya Batra, Prajjwal Bhargava, Shruti Bhosale, et al. 2023. Llama 2: Open foundation and fine-tuned chat models. arXiv preprint arXiv:2307.09288.
  30. Rose E Wang, Esin Durmus, Noah Goodman, and Tatsunori Hashimoto. 2022. Language modeling via stochastic processes. arXiv preprint arXiv:2203.11370. — [arXiv:2203.11370](https://arxiv.org/abs/2203.11370) find
  31. Xiaoou Wang, Elena Cabrio, and Serena Villata. 2023a. Argument and counter-argument generation: a critical survey. In International Conference on Applications of Natural Language to Information Systems, pages 500–510. Springer. find
  32. Zhenhailong Wang, Shaoguang Mao, Wenshan Wu, Tao Ge, Furu Wei, and Heng Ji. 2023b. Unleashing the emergent cognitive synergy in large language models: A task-solving agent through multi-persona self-collaboration. arXiv preprint arXiv:2307.05300.
  33. Jason Wei, Xuezhi Wang, Dale Schuurmans, Maarten Bosma, Fei Xia, Ed Chi, Quoc V Le, Denny Zhou, et al. 2022. Chain-of-thought prompting elicits reasoning in large language models. Advances in neural information processing systems, 35:24824–24837.
  34. Benfeng Xu, An Yang, Junyang Lin, Quan Wang, Chang Zhou, Yongdong Zhang, and Zhendong Mao. 2023. Expertprompting: Instructing large language models to be distinguished experts. arXiv preprint arXiv:2305.14688. find
  35. Kevin Yang, Yuandong Tian, Nanyun Peng, and Dan Klein. 2022. Re3: Generating longer stories with recursive reprompting and revision. arXiv preprint arXiv:2210.06774. find
  36. Lianmin Zheng, Wei-Lin Chiang, Ying Sheng, Siyuan Zhuang, Zhanghao Wu, Yonghao Zhuang, Zi Lin, Zhuohan Li, Dacheng Li, Eric Xing, Hao Zhang, Joseph E Gonzalez, and Ion Stoica. 2023. Judging llm-as-a-judge with mt-bench and chatbot arena. In Advances in Neural Information Processing Systems, volume 3…
  37. Wangchunshu Zhou, Yuchen Eleanor Jiang, Peng Cui, Tiannan Wang, Zhenxin Xiao, Yifan Hou, Ryan Cotterell, and Mrinmaya Sachan. 2023. Recurrentgpt: Interactive generation of (arbitrarily) long text. arXiv preprint arXiv:2305.13304.
  38. Yaoming Zhu, Sidi Lu, Lei Zheng, Jiaxian Guo, Weinan Zhang, Jun Wang, and Yong Yu. 2018. Texygen: A benchmarking platform for text generation models. In The 41st international ACM SIGIR conference on research & development in information retrieval, pages 1097–1100.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/hu-2025-debate-write-persona-driven-multi/