MIRROR: Learning from the Other View for Multi-Modal Reasoning· MIRROR: 从另一视角学习多模态推理
Unlike large language models (LLMs) that exhibit strong reasoning capabilities, vision-language models (VLMs) struggle with visual reasoning, even on geometry problems that admit equivalent text, diagram, and combined diagram+text views. We show that these views often elicit different behaviors: a model may solve a problem from text but fail on the corresponding diagram, or succeed visually while failing textually. This inconsistency suggests that different views expose complementary reasoning paths and failure modes that standard multimodal post-training does not fully exploit. To study and exploit this phenomenon, we construct ODA-Data, a high-quality paired multimodal geometry dataset with text-dominant, image-dominant, and combined image+text views of the same problems, together with s
通过不同视角学习多模态推理，提高视觉-语言模型在几何问题上的表现。
- 核心方法
- 构建了高质量的多模态几何数据集ODA-Data，包含文本主导、图像主导及图文结合的视角，通过分析不同视角下的推理路径和失败模式来改进模型。
- 适合谁读
- 研究者、工程师
- 要解决的问题
- 现有的视觉-语言模型在解决几何问题时表现不佳，且在不同表达方式（文本、图像、图文结合）间存在推理不一致的问题。
- 关键实验
- 在ODA-Data数据集上进行了多种实验，验证了MIRROR框架在不同视角下的推理能力提升。
- 主要贡献
- 提出了MIRROR框架，利用不同视角的互补性来增强视觉-语言模型的多模态推理能力，并通过实验验证了该方法的有效性。
- 意义与局限
- 该研究揭示了多模态任务中不同表达形式的互补性，为提升多模态模型的推理能力提供了新思路，但仅限于几何问题，可能需要进一步扩展至其他领域。