Ground-R1: 通过强化学习激励基于地面的视觉推理
Ground-R1: Incentivizing Grounded Visual Reasoning via Reinforcement Learning
AI总结：
Ground-R1 通过 Scale Relative Policy Optimization 方法，解决 LVLM 在视觉证据 grounding 方面的偏差问题，提升推理准确性和证据 grounding 能力。
摘要：
大型视觉-语言模型（LVLMs）已成为强大的通用助手，但其预测往往缺乏可靠性和可解释性，这是因为其在视觉证据上的 grounding 不足。新兴的“图像思考”范式试图通过显式地将推理锚定在图像区域来解决这一问题。然而，我们实证发现，大多数现有方法存在系统性的规模驱动偏差，在优化过程中训练奖励被大型视觉区域主导，抑制了对小但语义关键的证据的学习，导致推理时出现虚假的 grounding。为了解决这一限制，我们提出了 Ground-R1，一个通过新颖的 Scale Relative Policy Optimization（SRPO）目标训练的去偏差“图像思考”框架，该目标取代了标准的 GRPO。具体而言，我们的 SRPO 通过规模感知的分桶和跨/内分桶比较，重新校准不同大小证据区域的奖励学习，从而在训练过程中实现平衡的信用分配。在通用 LVLM、高分辨率和视觉 grounding 评估基准上的实验结果验证了 Ground-R1 的有效性，并显示 SRPO 在响应准确性和证据 grounding 方面均优于标准 GRPO。
英文摘要：
Large Vision-Language Models (LVLMs) have become powerful general-purpose assistants, yet their predictions often lack reliability and interpretability due to insufficient grounding in visual evidence. The emerging thinking-with-images paradigm seeks to address this issue by explicitly anchoring reasoning to image regions. However, we empirically find that most existing methods suffer from a systematic scale-driven bias in optimization, where training rewards are dominated by large visual regions, suppressing learning from small but semantically critical evidence and leading to spurious grounding at inference time. To address this limitation, we propose Ground-R1, a de-biased thinking-with-images framework trained via a novel Scale Relative Policy Optimization (SRPO) objective that replaces standard GRPO. Specifically, our SRPO recalibrates reward learning across evidence regions of different sizes through scale-aware binning and intra-/inter-bin comparisons, enabling balanced credit assignment during training. Experimental results on general LVLM, high-resolution, and visual grounding benchmarks validate the effectiveness of Ground-R1 and show that SRPO yields consistent gains over standard GRPO in both response accuracy and evidence grounding.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
