GeoReform: Reflective Formalization Evolution for Multimodal Geometry Problem Solving· GeoReform：多模态几何问题解决的反思式形式化进化
Multimodal large language models (MLLMs) often struggle to identify and use geometric relations in diagrams. Recent methods address this challenge by converting geometric entities, relations, and constraints into explicit textual representations for the model to reason over. However, effective formalization is highly non-trivial: on Geometry3K, structure injection fixes 28 errors but introduces 13 new ones among 200 examples. Redundant relations can distract the model, while ambiguous references to diagram elements can lead it to apply constraints incorrectly. This suggests that the key challenge is not merely extracting more geometric facts, but organizing them into representations that support downstream reasoning. To fully exploit the power of formalization, we further propose GeoReform
提出 GeoReform 方法，改善多模态几何问题的形式化表示。
- 核心方法
- 通过反思式形式化进化技术，GeoReform 将几何实体、关系和约束转换为模型可以推理的显式文本表示，并优化这些表示以减少冗余和歧义，支持更有效的下游推理。
- 适合谁读
- 研究者、工程师
- 要解决的问题
- 多模态大语言模型在处理几何问题时难以准确识别和利用图中的几何关系。
- 关键实验
- 在 Geometry3K 数据集上的实验显示，GeoReform 能够减少错误数量并提高解决几何问题的性能。
- 主要贡献
- GeoReform 在几何问题解决中减少了由形式化表示不当引起的新错误，提高了模型的准确性和推理能力。
- 意义与局限
- GeoReform 为多模态大语言模型在几何问题上的应用提供了更有效的形式化表示方法，但可能存在对复杂几何关系处理不足的局限性。