TIM-PRM：利用工具集成PRM验证多模态推理
TIM-PRM: Verifying multimodal reasoning with Tool-Integrated PRM
AI总结：
TIM-PRM通过工具集成PRM验证多模态推理，有效解决视觉幻觉和逻辑不一致问题，实验表明其在性能和可解释性上优于现有模型。
摘要：
多模态大语言模型（MLLMs）在数学推理方面取得了显著成就，但它们仍然容易受到视觉幻觉和逻辑不一致的影响，而标准的结果导向监督无法有效缓解这些问题。尽管过程奖励模型（PRMs）承诺逐步验证，但当前方法通常作为标量评分者或生成性批评者，存在阿谀奉承的问题，盲目验证错误假设而非基于视觉现实。为弥合这一差距，我们引入TIM-PRM（工具集成多模态PRM），一种新的代理框架，将验证从被动分类任务转变为积极的工具增强调查。TIM-PRM被训练以明确规划验证策略，并利用独立问题提问机制通过外部工具查询证据，有效将验证与推理上下文解耦，以消除确认偏见。我们通过编纂高质量的工具集成验证轨迹数据集来实现这一方法。在VisualProcessBench上的广泛实验表明，我们的8B参数模型超越了现有的开源多模态PRMs，显著优于像Qwen2.5-72B和InternVL-78B这样的更大模型，同时提供了对验证过程的可解释见解。
英文摘要：
Multimodal Large Language Models (MLLMs) have achieved impressive performances in mathematical reasoning, yet they remain vulnerable to visual hallucinations and logical inconsistencies that standard outcome-based supervision fails to mitigate. While Process Reward Models (PRMs) promise step-by-step verification, current approaches typically operate as scalar scorers or generative critics that suffer from sycophancy, blindly validating the flawed hypotheses rather than grounding them in visual reality. To bridge this gap, we introduce TIM-PRM (Tool-Integrated Multimodal PRM), a novel agentic framework that transforms verification from a passive classification task into an active, tool-augmented investigation. TIM-PRM is trained to explicitly plan verification strategies and utilizes a mechanism of Independent Question Asking to query evidence via external tools, effectively decoupling verification from the reasoning context to eliminate confirmation bias. We instantiate this method by curating a high-quality dataset of tool-integrated verification trajectories. Extensive experiments on VisualProcessBench demonstrate that our 8B parameter model surpasses existing open-source multimodal PRMs, significantly outperforming much larger models like Qwen2.5-72B and InternVL-78B, while offering interpretable insights into the verification process.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
