Elamur Research⌕
☰
home / papers / 2025 / Beyond Outlining: Heterogeneous Recursive Planning for Adaptive Long-form Writing with Language Models

Beyond Outlining: Heterogeneous Recursive Planning for Adaptive Long-form Writing with Language Models

Paper · 2025 · Ruibin Xiong, Yimeng Chen, Dmitrii Khizbullin et al.

Summary

Most LLM writing agents first draft a fixed outline and then fill it in. This makes it hard to change course mid-writing, for example when new information or a plot problem appears. The authors present WriteHERE, which treats long-form writing as a planning problem inspired by Hierarchical Task Network planning. A writing goal is split recursively into typed subtasks of three kinds: retrieval, reasoning and composition. Splitting stops at primitive tasks, which are executed immediately. Planning and execution therefore alternate rather than being separated into an outlining stage and a writing stage. A State-based Hierarchical Task Scheduling algorithm tracks the subtasks and their dependencies as a DAG and decides what runs next. The framework was tested on fiction writing (Tell Me A Story) and on retrieval-based report generation (WildSeek). In both settings it outperformed state-of-the-art baselines such as STORM and Co-STORM on every automatic metric. This matters because a single goal-driven planner can adapt to quite different writing tasks without hand-designed workflows. The code and prompts are public.

Key points

  • Report experiments use WildSeek, which has 100 Topic-Intent samples across 24 domains. Retrieval runs through the Bing Search API. Separate search, analysis and writing agents execute the primitive tasks, and the search agent is itself made up of retrieval, reranking and summarization agents.
  • In an ablation without recursive planning ("w/o HRP"), the planner produces a flat list of subtasks. This version scores clearly lower on depth, and also loses novelty and breadth.
  • WriteHERE was tested in two configurations, one built on Claude and the other on DeepSeek-R1. With both, it beat Perplexity's commercial Deep Research (Feb. 2025) on all reported metrics.
  • The authors built a new LongReport benchmark of 12 detailed prompts for reports over 10,000 words. Five annotators compared WriteHERE with Gemini Deep-Research (2.5 Pro), and WriteHERE came out roughly on par, with a slight 7:5 preference in its favor.
  • The code is available at GitHub principia-ai/WriteHERE.

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

🎓King Abdullah University of Science and TechnologyJeddah · Saudi Arabia27 works · 1,084 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

🧭 Topics 1Research subjects and languages the work is about.Language models & pretraining · 3,320 works

Cite

Cite as Ruibin Xiong, Yimeng Chen, Dmitrii Khizbullin, Mingchen Zhuge and Jürgen Schmidhuber. 2025. Beyond Outlining: Heterogeneous Recursive Planning for Adaptive Long-form Writing with Language Models. Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing.

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2025.emnlp-main.1254

@inproceedings{xiong2025beyond,
  title = {{Beyond Outlining: Heterogeneous Recursive Planning for Adaptive Long-form Writing with Language Models}},
  author = {Ruibin Xiong, Yimeng Chen, Dmitrii Khizbullin, Mingchen Zhuge and Jürgen Schmidhuber},
  booktitle = {Proceedings of the 2025 Conference on Empirical Methods in Natural Language Processing},
  year = {2025},
  doi = {10.18653/v1/2025.emnlp-main.1254},
  url = {https://doi.org/10.18653/v1/2025.emnlp-main.1254}
}

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. Qiguang Chen, Libo Qin, Jiaqi Wang, Jinxuan Zhou, and Wanxiang Che. 2024a. Unlocking the boundaries of thought: A reasoning granularity framework to quantify and optimize chain-of-thought. arXiv eprints. find
  2. Weize Chen, Ziming You, Ran Li, Yitong Guan, Chen Qian, Chenyang Zhao, Cheng Yang, Ruobing Xie, Zhiyuan Liu, and Maosong Sun. 2024b. Internet of agents: Weaving a web of heterogeneous agents for collaborative intelligence. arXiv preprint arXiv:2407.07061. — [arXiv:2407.07061](https://arxiv.org/abs/2407.07061) find
  3. Ken Currie and Austin Tate. 1991. O-plan: the open planning architecture. Artificial intelligence. find
  4. DeepSeek. 2024. DeepSeek-R1: First-Rank Implementation Details of DeepSeek. Technical report, DeepSeek. Available at: github.com/deepseek-ai/DeepSeek-R1. find
  5. Kutluhan Erol, James Hendler, and Dana S Nau. 1994. Htn planning: Complexity and expressivity. In AAAI. find
  6. Tira Nur Fitria. 2023. Artificial intelligence (ai) technology in openai chatgpt application: A review of chatgpt in writing english essay. In ELT Forum: Journal of English Language Teaching. find
  7. Ilche Georgievski and Marco Aiello. 2015. HTN planning: Overview, comparison, and beyond. Artificial Intelligence. find
  8. Malik Ghallab, Dana Nau, and Paolo Traverso. 2004. Automated Planning: theory and practice. Elsevier. find
  9. Sirui Hong, Xiawu Zheng, Jonathan Chen, Yuheng Cheng, Jinlin Wang, Ceyao Zhang, Zili Wang, Steven Ka Shing Yau, Zijuan Lin, Liyang Zhou, et al. 2023. Metagpt: Meta programming for multi-agent collaborative framework. arXiv preprint arXiv:2308.00352.
  10. Fantine Huot, Reinald Kim Amplayo, Jennimaria Palomaki, Alice Shoshana Jakobovits, Elizabeth Clark, and Mirella Lapata. 2024. Agents’ room: Narrative generation through multi-step collaboration. arXiv preprint arXiv:2410.02603. — [arXiv:2410.02603](https://arxiv.org/abs/2410.02603) find
  11. Yucheng Jiang, Yijia Shao, Dekun Ma, Sina J Semnani, and Monica S Lam. 2024. Into the unknown unknowns: Engaged human learning through participation in language model agent conversations. arXiv preprint arXiv:2408.15232. find
  12. Omar Khattab, Arnav Singhvi, Paridhi Maheshwari, Zhiyuan Zhang, Keshav Santhanam, Sri Vardhamanan A, Saiful Haq, Ashutosh Sharma, Thomas T. Joshi, Hanna Moazam, Heather Miller, Matei Zaharia, and Christopher Potts. 2024. DSPy: Compiling declarative language model calls into state-of-the-art pipeline… find
  13. Jürgen Schmidhuber. 2018. One big net for everything. arXiv preprint arXiv:1802.08864. — [arXiv:1802.08864](https://arxiv.org/abs/1802.08864) find
  14. Tushar Khot, Harsh Trivedi, Matthew Finlayson, Yao Fu, Kyle Richardson, Peter Clark, and Ashish Sabharwal. 2023. Decomposed prompting: A modular approach for solving complex tasks. In The Eleventh International Conference on Learning Representations. find
  15. Alessandro Sordoni, Xingdi Yuan, Marc-Alexandre Côté, Matheus Pereira, Adam Trischler, Ziang Xiao, Arian Hosseini, Friederike Niedtner, and Nicolas Le Roux. 2023. Joint Prompt Optimization of Stacked LLMs using Variational Inference. In Thirty-seventh Conference on Neural Information Processing Syst…
  16. Seungone Kim, Juyoung Suk, Shayne Longpre, Bill Yuchen Lin, Jamin Shin, Sean Welleck, Graham Neubig, Moontae Lee, Kyungjae Lee, and Minjoon Seo. 2024. Prometheus 2: An open source language model specialized in evaluating other language models. arXiv preprint arXiv:2405.01535.
  17. Chris Lu, Cong Lu, Robert Tjarko Lange, Jakob Foerster, Jeff Clune, and David Ha. 2024. The AI Scientist: Towards fully automated open-ended scientific discovery. arXiv preprint arXiv:2408.06292.
  18. Dana Nau, Yue Cao, Amnon Lotem, and Hector Munoz-Avila. 1999. SHOP: Simple hierarchical ordered planner. In Proceedings of the 16th international joint conference on Artificial intelligence-Volume 2. find
  19. Dana S Nau. 2007. Current trends in automated planning. AI magazine. find
  20. Dana S Nau, Tsz-Chiu Au, Okhtay Ilghami, Ugur Kuter, J William Murdock, Dan Wu, and Fusun Yaman. 2003. SHOP2: An HTN planning system. Journal of artificial intelligence research. find
  21. Archiki Prasad, Alexander Koller, Mareike Hartmann, Peter Clark, Ashish Sabharwal, Mohit Bansal, and Tushar Khot. 2023. Adapt: As-needed decomposition and planning with language models. arXiv preprint arXiv:2311.05772.
  22. Earl D Sacerdoti. 1971. A structure for plans and behavior. Tech. Note 109. find
  23. Earl D Sacerdoti. 1975. The nonlinear nature of plans. In Proceedings of the 4th international joint conference on Artificial intelligence-Volume 1. find
  24. Yijia Shao, Yucheng Jiang, Theodore A Kanell, Peter Xu, Omar Khattab, and Monica S Lam. 2024. Assisting in writing wikipedia-like articles from scratch with large language models. arXiv preprint arXiv:2402.14207.
  25. Theodore Sumers, Shunyu Yao, Karthik Narasimhan, and Thomas Griffiths. 2024. Cognitive architectures for language agents. Transactions on Machine Learning Research. find
  26. Austin Tate. 1977. Generating project networks. In Proceedings of the 5th international joint conference on Artificial intelligence-Volume 2. find
  27. Austin Tate, Brian Drabble, and Richard Kirby. 1994. O-Plan2: an open architecture for command, planning and control. Intelligent scheduling. find
  28. Liang Wang, Haonan Chen, Nan Yang, Xiaolong Huang, Zhicheng Dou, and Furu Wei. 2025. Chain-of-Retrieval Augmented Generation. arXiv preprint arXiv:2501.14342. — [arXiv:2501.14342](https://arxiv.org/abs/2501.14342) find
  29. 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.
  30. David E Wilkins. 1990. Can AI planners solve practical problems? Computational intelligence, 6(4):232– 246. find
  31. Yiran Wu, Tianwei Yue, Shaokun Zhang, Chi Wang, and Qingyun Wu. 2024. Stateflow: Enhancing llm task-solving through state-driven workflows. arXiv preprint arXiv:2403.11322.
  32. J. Schmidhuber. 1992. Learning complex, extended sequences using the principle of history compression. Neural Computation. find
  33. Kevin Yang, Dan Klein, Nanyun Peng, and Yuandong Tian. 2023. DOC: Improving Long Story Coherence With Detailed Outline Control. In Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics.
  34. J. Schmidhuber and R. Wahnsiedler. 1992. Planning Simple Trajectories Using Neural Subgoal Generators. In Proc. of the 2nd International Conference on Simulation of Adaptive Behavior. MIT Press. 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. Jürgen Schmidhuber. 2015. On learning to think: Algorithmic information theory for novel combinations of reinforcement learning controllers and recurrent neural world models. arXiv preprint arXiv:1511.09249. — [arXiv:1511.09249](https://arxiv.org/abs/1511.09249) find
  37. Ling Yang, Zhaochen Yu, Bin Cui, and Mengdi Wang. 2025. Reasonflux: Hierarchical llm reasoning via scaling thought templates. arXiv preprint arXiv:2502.06772. find
  38. Lili Yao, Nanyun Peng, Ralph Weischedel, Kevin Knight, Dongyan Zhao, and Rui Yan. 2019. Plan-and-write: Towards better automatic storytelling. In Proceedings of the AAAI Conference on Artificial Intelligence. find
  39. Shunyu Yao, Dian Yu, Jeffrey Zhao, Izhak Shafran, Thomas L Griffiths, Yuan Cao, and Karthik Narasimhan. 2023. Tree of thoughts: deliberate problem solving with large language models. In Proceedings of the 37th International Conference on Neural Information Processing Systems. find
  40. Shunyu Yao, Jeffrey Zhao, Dian Yu, Nan Du, Izhak Shafran, Karthik Narasimhan, and Yuan Cao. 2022. React: Synergizing reasoning and acting in language models. arXiv preprint arXiv:2210.03629.
  41. Jiayi Zhang, Jinyu Xiang, Zhaoyang Yu, Fengwei Teng, Xionghui Chen, Jiaqi Chen, Mingchen Zhuge, Xin Cheng, Sirui Hong, Jinlin Wang, et al. 2024. Aflow: Automating agentic workflow generation. arXiv preprint arXiv:2410.10762. find
  42. Denny Zhou, Nathanael Schärli, Le Hou, Jason Wei, Nathan Scales, Xuezhi Wang, Dale Schuurmans, Claire Cui, Olivier Bousquet, Quoc Le, et al. 2022. Least-to-most prompting enables complex reasoning in large language models. arXiv preprint arXiv:2205.10625.
  43. Mingchen Zhuge, Wenyi Wang, Louis Kirsch, Francesco Faccio, Dmitrii Khizbullin, and Jürgen Schmidhuber. 2024. GPTSwarm: Language Agents as Optimizable Graphs. In Forty-first International Conference on Machine Learning. find

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/xiong-2025-beyond-outlining-heterogeneous-recursive-planning/