Elamur Research⌕
☰
home / papers / 2026 / GATE: Graph-based Adaptive Tool Evolution Across Diverse Tasks

GATE: Graph-based Adaptive Tool Evolution Across Diverse Tasks

Paper · 2026 · Jianwen Luo, Yiming Huang, Jinxiang Meng et al.

Summary

LLM-based tool-making systems tend to accumulate bloated, repetitive tool libraries and are usually tied to a single task type, such as Minecraft or code generation. GATE addresses this by organizing tools as a layered Tool Graph. Human-written basic tools sit at the bottom, newly composed tools are placed one level above the deepest tool they call, and edges record call dependencies. Tools are retrieved with GraphRank, which combines embedding similarity (text-embedding-ada-002 over docstrings) with a PageRank-style random walk across the graph. Two agents work together. A Task Solver states what functionality a task needs, and a Tool Manager retrieves, composes, self-checks, merges and prunes tools. The authors evaluate the framework on open-ended play (Minecraft), agent benchmarks (TextCraft, DABench) and code generation (MATH, Date, TabMWP). GATE outperforms earlier tool-making approaches in all three settings while keeping a much smaller library. The work suggests that structured, continually maintained tool collections can carry over across domains and stay manageable as they grow.

Key points

  • Compared with Voyager, GATE completes Minecraft's technology progression 4.3× more quickly. It also discovers 3.5× more items and travels 2.7× farther.
  • On average, GATE improves over existing tool-making methods by 9.23% on code generation and 10.03% on agent tasks. It uses 86.2% fewer tools than CREATOR and CRAFT.
  • GraphRank uses a damping factor d = 0.4, slightly favoring semantic relevance over graph structure. MATH accuracy stays at about 77–79% across different top-k values but drops when d exceeds 0.7.
  • In ablations, Self-Check and Tool Merging matter most. Removing Self-Check lowers Date accuracy by 16.3%, and removing merging can make Minecraft milestones take up to six times longer.
  • For closed-ended tasks, tools are built with GPT-4o and then applied by Qwen2.5-14B-Instruct at test time. Code and data are released at github.com/ayanami2003/GATE.

Our summary of the paper, in our own words — not its text. The work belongs to its authors; read it at the links above.

 Fields, topics and concepts

📚 Shelves 5The 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 worksInformation Systems · 1,437 works

🧭 Topics 1Research subjects and languages the work is about.Code & programming languages · 129 works

Cite

Cite as Jianwen Luo, Yiming Huang, Jinxiang Meng, Fangyu Lei, Shizhu He, Xiao Liu, Shanshan Jiang, Bin Dong, Jun Zhao and Kang Liu. 2026. GATE: Graph-based Adaptive Tool Evolution Across Diverse Tasks. Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers).

The paper itself, not this page. BibTeX from the library's record of the work. DOI 10.18653/v1/2026.acl-long.87

@inproceedings{luo2026gate,
  title = {{GATE: Graph-based Adaptive Tool Evolution Across Diverse Tasks}},
  author = {Jianwen Luo, Yiming Huang, Jinxiang Meng, Fangyu Lei, Shizhu He, Xiao Liu, Shanshan Jiang, Bin Dong, Jun Zhao and Kang Liu},
  booktitle = {Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)},
  year = {2026},
  doi = {10.18653/v1/2026.acl-long.87},
  url = {https://doi.org/10.18653/v1/2026.acl-long.87}
}

References

40 references, 12 of them with a card of ours. ↗ leads to the work, «find» searches for one we have no id for yet.

  1. Tianle Cai, Xuezhi Wang, Tengyu Ma, Xinyun Chen, and Denny Zhou. 2023. Large language models as tool makers. arXiv preprint arXiv:2305.17126. find
  2. Wenhu Chen, Xueguang Ma, Xinyi Wang, and William W Cohen. 2022. Program of thoughts prompting: Disentangling computation from reasoning for numerical reasoning tasks. arXiv preprint arXiv:2211.12588. find
  3. Yanfei Chen, Jinsung Yoon, Devendra Singh Sachan, Qingze Wang, Vincent Cohen-Addad, Mohammadhossein Bateni, Chen-Yu Lee, and Tomas Pfister. 2024. Re-invoke: Tool invocation rewriting for zero-shot tool retrieval. find
  4. Marc-Alexandre Côté, Akos Kádár, Xingdi Yuan, Ben Kybartas, Tavian Barnes, Emery Fine, James Moore, Matthew Hausknecht, Layla El Asri, Mahmoud Adada, and 1 others. 2019. Textworld: A learning environment for text-based games. In Computer Games: 7th Workshop, CGW 2018, Held in Conjunction with the 27…
  5. Linxi Fan, Guanzhi Wang, Yunfan Jiang, Ajay Mandlekar, Yuncong Yang, Haoyi Zhu, Andrew Tang, De-An Huang, Yuke Zhu, and Anima Anandkumar. 2022. Minedojo: Building open-ended embodied agents with internet-scale knowledge. Advances in Neural Information Processing Systems, 35:18343– 18362. find
  6. Douglas J Futuyma and Gabriel Moreno. 1988. The evolution of ecological specialization. Annual review of Ecology and Systematics, pages 207–233. find
  7. Gabriel Grand, Lionel Wong, Matthew Bowers, Theo X Olausson, Muxin Liu, Joshua B Tenenbaum, and Jacob Andreas. 2023. Learning interpretable libraries by compressing and documenting code. In Intrinsically-Motivated and Open-Ended Learning Workshop@ NeurIPS2023. find
  8. Xinyu Guan, Li Lyna Zhang, Yifei Liu, Ning Shang, Youran Sun, Yi Zhu, Fan Yang, and Mao Yang. 2025. rstar-math: Small llms can master math reasoning with self-evolved deep thinking. arXiv preprint arXiv:2501.04519.
  9. Dan Hendrycks, Collin Burns, Saurav Kadavath, Akul Arora, Steven Basart, Eric Tang, Dawn Song, and Jacob Steinhardt. 2021. Measuring mathematical problem solving with the math dataset. arXiv preprint arXiv:2103.03874.
  10. Xueyu Hu, Ziyu Zhao, Shuang Wei, Ziwei Chai, Qianli Ma, Guoyin Wang, Xuwu Wang, Jing Su, Jingjing Xu, Ming Zhu, and 1 others. 2024. Infiagent-dabench: Evaluating agents on data analysis tasks. arXiv preprint arXiv:2401.05507. — [arXiv:2401.05507](https://arxiv.org/abs/2401.05507) find
  11. Xuefeng Li, Haoyang Zou, and Pengfei Liu. 2025. Torl: Scaling tool-integrated rl. arXiv preprint arXiv:2503.23383. find
  12. OpenAI. 2022. Text-embedding-ada-002: New and improved embedding model. Accessed: 2025-02-11. find
  13. PrismarineJS. 2013. Prismarinejs/mineflayer: Create minecraft bots with a powerful, stable, and high-level javascript api. https://github.com/ PrismarineJS/mineflayer. Accessed: 2025-02-15. ↗
  14. PrismarineJS. 2013. Prismarinejs/mineflayer: Create minecraft bots with a powerful, stable, and high-level javascript api. https://github.com/ PrismarineJS/mineflayer. Accessed: 2025-01-08. ↗
  15. Cheng Qian, Chi Han, Yi R Fung, Yujia Qin, Zhiyuan Liu, and Heng Ji. 2023. Creator: Tool creation for disentangling abstract and concrete reasoning of large language models. arXiv preprint arXiv:2305.14318.
  16. Changle Qu, Sunhao Dai, Xiaochi Wei, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, Jun Xu, and Ji-Rong Wen. 2024a. From exploration to mastery: Enabling llms to master tools via self-driven interactions. find
  17. Changle Qu, Sunhao Dai, Xiaochi Wei, Hengyi Cai, Shuaiqiang Wang, Dawei Yin, Jun Xu, and Ji-Rong Wen. 2024b. Towards completeness-oriented tool retrieval for large language models. In Proceedings of the 33rd ACM International Conference on Information and Knowledge Management, CIKM ’24, page 1930–19… find
  18. Toran Bruce Richards. Significant-gravitas/autogpt: An experimental open-source attempt to make gpt-4 fully autonomous., 2023. URL https://github.com/Significant-Gravitas/AutoGPT. ↗
  19. Zhengliang Shi, Yuhan Wang, Lingyong Yan, Pengjie Ren, Shuaiqiang Wang, Dawei Yin, and Zhaochun Ren. 2025. Retrieval models aren’t tool-savvy: Benchmarking tool retrieval for large language models. arXiv preprint arXiv:2503.01763. find
  20. Noah Shinn, Beck Labash, and Ashwin Gopinath. 2023. Reflexion: an autonomous agent with dynamic memory and self-reflection. arXiv preprint arXiv:2303.11366, 2(5):9.
  21. Kumar Shridhar, Koustuv Sinha, Andrew Cohen, Tianlu Wang, Ping Yu, Ram Pasunuru, Mrinmaya Sachan, Jason Weston, and Asli Celikyilmaz. 2023. The art of llm refinement: Ask, refine, and trust. arXiv preprint arXiv:2311.07961. find
  22. Temple F Smith, Michael S Waterman, and 1 others. 1981. Identification of common molecular subsequences. Journal of molecular biology, 147(1):195– 197. find
  23. Aarohi Srivastava, Abhinav Rastogi, Abhishek Rao, Abu Awal Md Shoeb, Abubakar Abid, Adam Fisch, Adam R Brown, Adam Santoro, Aditya Gupta, Adrià Garriga-Alonso, and 1 others. 2022. Beyond the imitation game: Quantifying and extrapolating the capabilities of language models. arXiv preprint arXiv:2206.…
  24. Elias Stengel-Eskin, Archiki Prasad, and Mohit Bansal. 2024. Regal: Refactoring programs to discover generalizable abstractions. arXiv preprint arXiv:2401.16467. find
  25. Guanzhi Wang, Yuqi Xie, Yunfan Jiang, Ajay Mandlekar, Chaowei Xiao, Yuke Zhu, Linxi Fan, and Anima Anandkumar. 2023. Voyager: An open-ended embodied agent with large language models. arXiv preprint arXiv:2305.16291. find
  26. Renxi Wang, Xudong Han, Lei Ji, Shu Wang, Timothy Baldwin, and Haonan Li. 2024a. Toolgen: Unified tool retrieval and calling via generation. arXiv preprint arXiv:2410.03439. — [arXiv:2410.03439](https://arxiv.org/abs/2410.03439) find
  27. Xingyao Wang, Yangyi Chen, Lifan Yuan, Yizhe Zhang, Yunzhu Li, Hao Peng, and Heng Ji. 2024b. Executable code actions elicit better llm agents. arXiv preprint arXiv:2402.01030.
  28. Xingyao Wang, Boxuan Li, Yufan Song, Frank F. Xu, Xiangru Tang, Mingchen Zhuge, Jiayi Pan, Yueqi Song, Bowen Li, Jaskirat Singh, Hoang H. Tran, Fuqiang Li, Ren Ma, Mingzhang Zheng, Bill Qian, Yanjun Shao, Niklas Muennighoff, Yizhe Zhang, Binyuan Hui, and 5 others. 2025a. Openhands: An open platform…
  29. Yaoxiang Wang, Haoling Li, Xin Zhang, Jie Wu, Xiao Liu, Wenxiang Hu, Zhongxin Guo, Yangyu Huang, Ying Xin, Yujiu Yang, and 1 others. 2025b. Epicoder: Encompassing diversity and complexity in code generation. arXiv preprint arXiv:2501.04694. — [arXiv:2501.04694](https://arxiv.org/abs/2501.04694) find
  30. Zhiruo Wang, Daniel Fried, and Graham Neubig. 2024c. Trove: Inducing verifiable and efficient toolboxes for solving programmatic tasks. arXiv preprint arXiv:2401.12869. find
  31. Shirley Wu, Shiyu Zhao, Qian Huang, Kexin Huang, Michihiro Yasunaga, Kaidi Cao, Vassilis N. Ioannidis, Karthik Subbian, Jure Leskovec, and James Zou. 2024. Avatar: Optimizing llm agents for tool usage via contrastive reasoning. find
  32. Xianjie Wu, Jian Yang, Linzheng Chai, Ge Zhang, Jiaheng Liu, Xeron Du, Di Liang, Daixin Shu, Xianfu Cheng, Tianzhen Sun, and 1 others. 2025. Tablebench: A comprehensive and complex benchmark for table question answering. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 39, pa… find
  33. Wenpu Xing and Ali Ghorbani. 2004. Weighted pagerank algorithm. In Proceedings. Second Annual Conference on Communication Networks and Services Research, 2004., pages 305–314. IEEE. find
  34. An Yang, Anfeng Li, Baosong Yang, Beichen Zhang, Binyuan Hui, Bo Zheng, Bowen Yu, Chang Gao, Chengen Huang, Chenxu Lv, and 1 others. 2025. Qwen3 technical report. arXiv preprint arXiv:2505.09388.
  35. John Yang, Akshara Prabhakar, Karthik Narasimhan, and Shunyu Yao. 2023. Intercode: Standardizing and benchmarking interactive coding with execution feedback.
  36. 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.
  37. Lifan Yuan, Yangyi Chen, Xingyao Wang, Yi R Fung, Hao Peng, and Heng Ji. 2023. Craft: Customizing llms by creating and retrieving from specialized toolsets. arXiv preprint arXiv:2309.17428. — [arXiv:2309.17428](https://arxiv.org/abs/2309.17428) find
  38. Huaye Zeng, Dongfu Jiang, Haozhe Wang, Ping Nie, Xiaotong Chen, and Wenhu Chen. 2025. Acecoder: Acing coder rl via automated test-case synthesis. arXiv preprint arXiv:2502.01718. find
  39. Yuanhang Zheng, Peng Li, Wei Liu, Yang Liu, Jian Luan, and Bin Wang. 2024. Toolrerank: Adaptive and hierarchy-aware reranking for tool retrieval. find
  40. Aojun Zhou, Ke Wang, Zimu Lu, Weikang Shi, Sichun Luo, Zipeng Qin, Shaoqing Lu, Anya Jia, Linqi Song, Mingjie Zhan, and 1 others. 2023. Solving challenging math word problems using gpt-4 code interpreter with code-based self-verification. arXiv preprint arXiv:2308.07921.

Updated

↑ back to top

This page in Markdown → https://md.research.elamur.ai/free/en/papers/luo-2026-gate-graph-based-adaptive-tool/