ReaGEN：面向高效多模态推理的自适应结构化思维链生成
ReaGEN: Adaptive Generation of Structured Chains-of-Thought for Efficient Multimodal Reasoning
对做多模态推理的视觉研究者，这篇给出了一个可复用的思路：用 VLM 内部注意力信号衡量推理各阶段的重要性，并训练一个轻量生成器动态产出问题自适应的思维链结构，从而以单路径开销获得多路径搜索的灵活性；可直接嵌入自托管 LVLM 的推理流程以替代昂贵的测试时搜索。
原文依据 leveraging attention-derived stage importance to capture how information flows across reasoning stages
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论大型视觉语言模型（LVLM）具备强大的感知与语言能力，但在需要结构化、组合式与自适应推理的复杂视觉推理任务上仍显不足。现有方法要么依赖代价高昂的推理时探索——例如多路径或基于树结构的思维链（CoT）搜索——要么依赖使用大规模精心构建的 CoT 数据集进行昂贵的后训练。我们提出 ReaGEN，一个用于自适应生成结构化推理链的轻量框架，它在不修改底层视觉语言模型（VLM）的前提下增强推理能力。ReaGEN 首先采用教师引导的进化搜索来收集针对具体样本的 CoT 结构，并利用由注意力导出的阶段重要性来刻画信息在各推理阶段之间的流动方式。随后，这些自适应 CoT 结构被用于训练一个紧凑的生成器（GEN），该生成器通过反思推理过程中的注意力反馈来学习精炼与改进 CoT 结构。在推理时，GEN 动态地产出问题自适应的结构化 CoT，并可根据 VLM 的内部状态被迭代调用以对其加以精炼——从而以单路径的效率实现深度搜索的灵活性。在多种多模态推理基准上，ReaGEN 相较测试时扩展方法最高取得 +26 个准确率点的提升，同时将平均推理时词元用量降低 79%，使其成为一种可扩展的结构化推理生成方法，适用于能够访问内部注意力信号的自托管或开源 VLM。
英文原文摘要 1540 字符
Large Vision Language Models (LVLMs) exhibit strong perceptual and linguistic capabilities yet struggle with complex visual reasoning tasks that require structured, compositional, and adaptive inference. Existing approaches either rely on costly inference-time exploration--such as multi-path or tree-based Chain-of-Thought (CoT) search--or on expensive post-training with large curated CoT datasets. We propose ReaGEN, a lightweight framework for the adaptive generation of structured reasoning chains that enhances reasoning without modifying the underlying vision-language model (VLM). ReaGEN first employs a teacher-guided evolutionary search to collect sample specific CoT structure, leveraging attention-derived stage importance to capture how information flows across reasoning stages. These adaptive CoT structures are then used to train a compact generator (GEN) that learns to refine and improve CoT structures by reflecting on attention feedback from the reasoning process. At inference, the GEN dynamically produces question-adaptive structured CoTs, and can be iteratively invoked to refine them based on the VLM's internal state--achieving the flexibility of deep search with single-path efficiency. Across diverse multimodal reasoning benchmarks, ReaGEN achieves up to +26 accuracy points over test-time scaling methods while reducing the average inference-time token usage by 79%, making it a scalable approach for structured reasoning generation in self-hosted or open-source VLMs with access to internal attention signals.