发现视觉推理中工具使用会逐渐消失，但性能不降反升。
Diversity Over Frequency: Rethinking Tool Use in Visual Chain-of-Thought Agents
- 引入熵正则化提升推理路径多样性，缓解工具依赖退化。
- 工具使用率下降时准确率仍提升，出现工具使用崩溃现象。
- 适合研究多步视觉推理与智能体训练策略的学者参考。
视觉智能体在视觉思维链中使用外部视觉工具以获取细粒度证据。以往研究主要集中于简单的视觉搜索任务，而对复杂视觉推理任务中的工具作用关注不足。本文拓展至3D空间推理和医学视觉问答等挑战性任务，探究工具获取局部证据与全局上下文融合的问题。我们发现一种工具使用崩溃现象：模型在准确率持续提升的同时逐步停止使用工具。同时观察到显著不对称性：完全禁用工具导致性能下降，而激励工具使用仅带来微弱收益却大幅增加使用频率。分析表明，基础训练与工具鼓励均降低推理轨迹多样性，解释了为何更高工具使用未提升性能。基于此，我们引入熵正则化项以促进多样化探索，在工具使用率逐渐下降的情况下实现最优性能。整体表明，工具在训练阶段应视为支撑结构，通过扩大语言生成与工具调用的探索范围，可增强推理能力，即使出现工具使用衰减。
原文摘要 · Abstract (English)
Visual agents employ external visual tools within visual chains of thought to incorporate fine-grained evidence. While prior work has mainly studied these tools in visual search tasks, their role in more complex visual reasoning remains underexplored. In this paper, we move beyond simple visual search tasks to investigate more challenging tasks, including 3D spatial reasoning and medical visual question answering, where agents must integrate tool-acquired local evidence with the global context. We identify a {tool-use collapse phenomenon: models progressively stop using tools while still achieving higher task accuracy. Moreover, we observe a clear asymmetry: (i) completely eliminating tool use degrades performance, whereas (ii) incentivizing tool use yields only marginal gains despite substantially increasing usage. We find that vanilla training and tool-use encouragement both reduce rollout diversity, explaining why higher tool use does not yield stronger reasoning performance. Motivated by these findings, we add an entropy regularization term to encourage diverse rollout exploration, achieving the best performance despite gradually declining tool usage. Overall, our findings suggest a training-time view of tools as scaffolding, where broader exploration over language generation and visual tool invocation improves reasoning despite tool-use collapse. Project page: https://scaffolded-exploration.github.io
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。