用视觉信息增益优化多模态推理，让每句思考都紧扣图像。
VIG: Visual Information Gain as a Reward Signal for Multimodal Chain-of-Thought Compression
- 基于信息论设计奖励机制，判断每个推理词是否真正依赖图像。
- 在6个基准上提升准确率与效率平衡，8B模型效果更优。
- 无需标注或额外模型，适合追求高效多模态推理的研究者。
多模态大模型的链式思维（CoT）常包含大量重复的视觉描述、自我反思等与图像无关的内容，增加推理开销却不提升答案质量。现有压缩方法仅关注长度，未评估推理内容是否真实基于图像。本文提出VIG（视觉信息增益），一种基于信息论的GRPO奖励机制，通过两次前向传播（有/无图像）在线计算每个推理词的图像信息贡献度。无需参考链、外部标注或辅助奖励模型。在六个主流多模态推理基准及三个Qwen3-VL-Thinking模型规模（2B/4B/8B）上，以及额外的R1-Onevision-Bench测试中，VIG持续改善准确率-效率权衡。结果支持核心观点：高效的多模态推理源于提升视觉信息密度，即每个推理词必须锚定图像，而非单纯限制长度。代码已开源。
原文摘要 · Abstract (English)
Multimodal large reasoning models often rely on long Chain-of-Thought (CoT) traces in which a substantial fraction of tokens, such as repeated visual descriptions, self-reflection, and other visually-disengaged filler, inflate inference cost without contributing to the answer. Existing CoT compression methods optimize output length but never measure whether a reasoning token is actually grounded in the image. We propose \textbf{VIG} (Visual Information Gain), an information-theoretic GRPO reward that scores each reasoning token by how much the image reduces its predictive uncertainty. VIG is computed online from two forward passes of the same policy, one with and one without the image, so no reference chains, external annotations, or auxiliary reward models are needed. Across six main multimodal reasoning benchmarks and three Qwen3-VL-Thinking model sizes (2B/4B/8B), plus an additional R1-Onevision-Bench evaluation on 8B, VIG consistently improves the accuracy--efficiency trade-off, supporting our central claim: \emph{efficient multimodal reasoning emerges from raising visual information density, where every reasoning token earns its place by anchoring to the image, rather than from imposing a length budget.} Our source code is available at https://github.com/chaser682/vig.
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。