让模型在推理时精准定位决策点，提升多模态几何题解答能力。
Learning Where Outcomes Change:Credit-Addressable Reasoning for Multimodal Geometry
- 用可定位的代码表达视觉关系，将推理拆解为带类型的事件
- 在9个几何基准上达到76.04%准确率，领先基线超8个百分点
- 适合需要长链条逻辑推理的多模态任务研究者
多模态几何推理要求视觉语言模型精确提取视觉关系，并在多步推导中保持其一致性。现有自由形式的推理轨迹掩盖了决定答案的关键决策，而轨迹级强化学习将单一终止信号分布于整个响应。本文提出可溯源推理机制，使推理过程中暴露的语义单元同时定义学习时比较选项与分配信用的位置。我们通过Code-CoT实现该原则：保留原始图示，将视觉关系表示为行可定位的可执行代码，并将推理组织为有类型事件；并提出CE-GRPO方法，利用结构先验和类型归一化熵确定事件边界，从共享前缀采样完整延续路径，并将结果差异转化为局部优势。在九个几何基准上，CE-GRPO平均准确率达76.04%，优于Qwen3-VL-8B和轨迹级GRPO分别8.09和3.43个百分点。其相对优势随中间事件数量增加而提升，证明了表示与优化协同设计对长依赖多模态推理的价值。
原文摘要 · Abstract (English)
Multimodal geometry reasoning requires VLMs to extract precise visual relations and preserve them through multi-step deduction. Existing free-form traces obscure the decisions that determine the answer, and trajectory-level reinforcement learning distributes a single terminal signal across the entire response. We introduce credit-addressable reasoning, in which the semantic units exposed during inference also define where learning compares alternatives and assigns credit. We instantiate this principle with Code-CoT, which retains the diagram, represents visual relations as line-addressable executable code, and organizes reasoning into typed events, and CE-GRPO, which selects event boundaries using structural priors and type-normalized entropy, samples complete continuations from shared prefixes, and converts outcome differences into localized advantages. Across nine geometry benchmarks, CE-GRPO achieves an average accuracy of 76.04, outperforming Qwen3-VL-8B and trajectory-level GRPO by $8.09$ and 3.43 points, respectively. Its relative advantage increases with the number of intermediate events, demonstrating the value of representation--optimization co-design for long, dependency-heavy multimodal reasoning.
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。