TikArt: 通过强化学习稳定引导孔细粒度视觉推理
TikArt: Stabilizing Aperture-Guided Fine-Grained Visual Reasoning with Reinforcement Learning
AI总结：
TikArt通过强化学习引导孔机制提升细粒度视觉推理，结合语言推理与Zoom、Segment操作，实现多模态证据获取与持久记忆。
摘要：
在多模态大语言模型（MLLMs）中，细粒度视觉推理受到单次全局图像编码的限制：关键证据往往位于微小物体、杂乱区域、细微标记或密集图表中。我们提出了TikArt（Thin-k-Ape-rture），一种引导孔代理，将多模态推理建模为对感兴趣区域的顺序证据获取。TikArt遵循Think--Aperture--Observe（TAO）循环，将语言推理与两个引导孔操作交织在一起：Zoom，提取矩形裁剪；Segment，调用现成的分割器生成以对象为中心的基于掩码的视图以处理不规则目标。每次引导孔操作后必须有一个观察步骤，将局部证据写回文本，从而产生可解释的引导孔轨迹和持久的语言记忆。基于Qwen3-VL-8B构建，TikArt通过GRPO风格的强化学习在双阶段课程中进行训练。为了稳定长周期工具集成学习，我们引入了相对不确定性减少（RUR），一种由冻结评估器计算的密集奖励，倾向于证据构建轨迹并减轻退化工具使用。在高分辨率推理、通用多模态理解以及指向和推理导向分割实验中，结果显示出对基座模型的持续改进，证明了引导孔观察能够提高细粒度视觉推理，并自然转移到像素级接地。
英文摘要：
Fine-grained visual reasoning in multimodal large language models (MLLMs) is bottlenecked by single-pass global image encoding: key evidence often lies in tiny objects, cluttered regions, subtle markings, or dense charts. We present \textbf{TikArt} (\textbf{T}h\textbf{i}n\textbf{k}ing \textbf{A}pe\textbf{rt}ure), an aperture-guided agent that formulates multimodal reasoning as sequential evidence acquisition over regions of interest. TikArt follows a Think--Aperture--Observe (TAO) loop that interleaves language reasoning with two aperture actions: Zoom, which extracts rectangular crops, and Segment, which invokes an off-the-shelf segmenter to produce object-centric mask-based views for irregular targets. A mandatory Observation step after every aperture action writes local evidence back into text, yielding interpretable aperture trajectories and persistent linguistic memory. Built on Qwen3-VL-8B, TikArt is trained with GRPO-style reinforcement learning under a two-stage curriculum. To stabilize long-horizon tool-integrated learning, we introduce Relative Uncertainty Reduction (RUR), a dense reward computed by a frozen evaluator that favors evidence-building trajectories and mitigates degenerate tool use. Experiments on high-resolution reasoning, general multimodal understanding, and both referring and reasoning-oriented segmentation show consistent gains over the backbone, demonstrating that aperture-guided observation improves fine-grained visual reasoning and transfers naturally to pixel-level 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。
