Don't Just Look, Intervene: Perturbation Based Region Labeling for VQA Images

论文摘要：Vision Language Models (VLMs) should rely on visual evidence that directly determines the correct answer, but supervision for grounding visual reasoning is often expensive to ...

为什么重要

该论文来自一手研究源并通过 3qk7 Research Radar 的新鲜度、来源与研究价值筛选；后续代码、模型权重、复现和相关产品事件可继续关联到该记录。

检测到的变化

• 研究论文进入 3qk7 高价值 Research Radar

一手证据

• arXiv cs.CV · 2026-09-15T22:05:54Z