@inproceedings{ding-etal-2025-toolcoder,
    title = "{T}ool{C}oder: A Systematic Code-Empowered Tool Learning Framework for Large Language Models",
    author = "Ding, Hanxing  and
      Tao, Shuchang  and
      Pang, Liang  and
      Wei, Zihao  and
      Gao, Jinyang  and
      Ding, Bolin  and
      Shen, Huawei  and
      Cheng, Xueqi",
    editor = "Che, Wanxiang  and
      Nabende, Joyce  and
      Shutova, Ekaterina  and
      Pilehvar, Mohammad Taher",
    booktitle = "Proceedings of the 63rd Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)",
    month = jul,
    year = "2025",
    address = "Vienna, Austria",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/ingestion-acl-25/2025.acl-long.874/",
    pages = "17876--17891",
    ISBN = "979-8-89176-251-0",
    abstract = "Tool learning has emerged as a crucial capability for large language models (LLMs) to solve complex real-world tasks through interaction with external tools. Existing approaches face significant challenges, including reliance on hand-crafted prompts, difficulty in multi-step planning, and lack of precise error diagnosis and reflection mechanisms. We propose $\texttt{ToolCoder}$, a novel framework that reformulates tool learning as a code generation task. Inspired by software engineering principles, $\texttt{ToolCoder}$ transforms natural language queries into structured Python function scaffold and systematically breaks down tasks with descriptive comments, enabling LLMs to leverage coding paradigms for complex reasoning and planning. It then generates and executes function implementations to obtain final responses. Additionally, $\texttt{ToolCoder}$ stores successfully executed functions in a repository to promote code reuse, while leveraging error traceback mechanisms for systematic debugging, optimizing both execution efficiency and robustness. Experiments demonstrate that $\texttt{ToolCoder}$ achieves superior performance in task completion accuracy and execution reliability compared to existing approaches, establishing the effectiveness of code-centric approaches in tool learning."
}