发表机构
Xiaomi Robotics（小米机器人）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对部分观测3D场景中语言指物定位的挑战，提出Finder智能体闭环查找原语，通过类型化循环状态整合规划、证据收集、候选验证与控制，在Habitat/HM3D和真实RGB-D场景中将平均1米成功率提升15.75个百分点，并可迁移至顺序定位和物体中心问答。
AI 中文摘要
在部分观测的3D场景中，找到语言所指的物体是具身智能体的核心能力。现有方法要么将物体搜索与在线探索耦合，当相关观测已被捕获时，这种做法可能代价高昂；要么以静态、一次性方式查询预构建的开放词汇地图和场景图。我们提出Finder，一种用于具身定位的智能体闭环物体查找原语。Finder不将定位视为从固定场景表示中的被动检索，而是维护一个类型化循环状态，该状态将查询条件规划、范围证据收集、候选验证以及接受/继续/中止控制联系起来。当证据不完整或模糊时，循环可以重定向后续感知和比较，而不是简单地返回检索到的顶部物体。在Habitat/HM3D中的开放词汇具身物体检索以及真实世界RGB-D场景中，Finder在平均1米成功率上比强基线提高了15.75个百分点。同一原语还可迁移到顺序物体定位和具身物体中心问答，在不改变内部定位协议的情况下改进空间和时间定位。项目页面：此https URL。
英文摘要
Finding the object referred to by language in a partially observed 3D scene is a core capability for embodied agents. Existing approaches either couple object search with online exploration, which can be costly when relevant observations have already been captured, or query pre-built open-vocabulary maps and scene graphs in a static, one-shot fashion. We present Finder, an agentic closed-loop object-finding primitive for embodied grounding. Instead of treating grounding as passive retrieval from a fixed scene representation, Finder maintains a typed loop state that links query-conditioned planning, scoped evidence gathering, candidate verification, and accept/continue/abort control. When evidence is incomplete or ambiguous, the loop can redirect subsequent perception and comparison rather than simply returning the top retrieved object. On open-vocabulary embodied Object Retrieval in Habitat/HM3D and real-world RGB-D scenes, Finder improves the averaged 1m success rate by 15.75 points over strong baselines. The same primitive also transfers to sequential object grounding and embodied object-centric question answering, improving spatial and temporal localization without changing the inner grounding protocol. Project page: https://finder-vln.github.io.
