@inproceedings{qian-etal-2025-agentthink,
    title = "{A}gent{T}hink: A Unified Framework for Tool-Augmented Chain-of-Thought Reasoning in Vision-Language Models for Autonomous Driving",
    author = "Qian, Kangan  and
      Jiang, Sicong  and
      Zhong, Yang  and
      Luo, Ziang  and
      Huang, Zilin  and
      Zhu, Tianze  and
      Jiang, Kun  and
      Yang, Mengmeng  and
      Fu, Zheng  and
      Miao, Jinyu  and
      Shi, Yining  and
      Lim, He Zhe  and
      Liu, Li  and
      Zhou, Tianbao  and
      Wang, Hongyi  and
      Yu, Huang  and
      Hu, Yifei  and
      Li, Guang  and
      Chen, Guang  and
      Ye, Hao  and
      Sun, Lijun  and
      Yang, Diange",
    editor = "Christodoulopoulos, Christos  and
      Chakraborty, Tanmoy  and
      Rose, Carolyn  and
      Peng, Violet",
    booktitle = "Findings of the Association for Computational Linguistics: EMNLP 2025",
    month = nov,
    year = "2025",
    address = "Suzhou, China",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/author-page-yu-wang-polytechnic/2025.findings-emnlp.564/",
    doi = "10.18653/v1/2025.findings-emnlp.564",
    pages = "10663--10682",
    ISBN = "979-8-89176-335-7",
    abstract = "Vision-Language Models (VLMs) show promise for autonomous driving, yet their struggle with hallucinations, inefficient reasoning, and limited real-world validation hinders accurate perception and robust step-by-step reasoning. To overcome this, we introduce $\textbf{AgentThink}$, a pioneering unified framework that, for the first time, integrates Chain-of-Thought (CoT) reasoning with dynamic, agent-style tool invocation for autonomous driving tasks. AgentThink{'}s core innovations include: $\textbf{(i) Structured Data Generation}$, by establishing an autonomous driving tool library to automatically construct structured, self-verified reasoning data explicitly incorporating tool usage for diverse driving scenarios; $\textbf{(ii) A Two-stage Training Pipeline}$, employing Supervised Fine-Tuning (SFT) with Group Relative Policy Optimization (GRPO) to equip VLMs with the capability for autonomous tool invocation; and $\textbf{(iii) Agent-style Tool-Usage Evaluation}$, introducing a novel multi-tool assessment protocol to rigorously evaluate the model{'}s tool invocation and utilization. Experiments on the DriveLMM-o1 benchmark demonstrate AgentThink significantly boosts overall reasoning scores by $\textbf{53.91\%}$ and enhances answer accuracy by $\textbf{33.54\%}$, while markedly improving reasoning quality and consistency. Furthermore, ablation studies and robust zero-shot/few-shot generalization experiments across various benchmarks underscore its powerful capabilities. These findings highlight a promising trajectory for developing trustworthy and tool-aware autonomous driving models."
}