arXiv cs.AI· Baiqiao Yin, Zhongang Cai, Ruisi Wang, Qiwei Liang, Lei Yang, Ziwei Liu·· 14 小时前AI 评分31
视觉基础模型是否具备空间状态表征？四层探测给出答案
Probing Spatial State Representations in Visual Foundation Models
AI 导读
研究通过空间结构、物体持续性、预测性空间更新和可操作性四个层级，用冻结特征探针在十个表征上探测视觉基础模型的空间状态能力。结果显示，对可见场景的强读出并不能保证准确的物体持续性或未来特征检索；在真实特征上训练的空间读取器无法原样迁移到预测特征。物体持续性与可操作性依赖相机位姿读出，而特征预测依赖可见场景读出。
正文 · 原文
Abstract:An observer sees only part of the surrounding world, yet must keep track of what lies beyond the current view. Spatial perception therefore involves more than recognizing individual images: it calls for a persistent spatial state that is updated under self-motion and used to guide action. Do visual foundation models provide representations that support such a state? We study this question through four levels of spatial capability---spatial structure, object persistence, predictive spatial update, and actionability---using frozen-feature probes on ten representations under streaming egocentric observations of static indoor environments. The resulting capability profiles vary across tasks, network layers, and prefix lengths: strong readout of the visible scene does not ensure accurate object persistence or future-feature retrieval. Matched controls separate readout accuracy from reliance on visual history and motion. Native future prediction retains recoverable spatial structure, but a spatial reader trained on real features does not transfer to predicted features unchanged. Finally, goal-dependent candidate selection does not establish reliable continuous action utility: predicted actions respond to goal changes, yet supplying the correct goal does not reliably improve progress. Across the four levels, object persistence and actionability build on camera-pose readout, whereas feature prediction builds on visible-scene readout. Together, these findings motivate evaluating spatial representations not only by what can be decoded from them, but by whether that information is maintained beyond observation, updated predictively, and used for goal-directed action.
| Subjects: | Computer Vision and Pattern Recognition (cs.CV); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2610.40636 [cs.CV] |
| (or arXiv:2610.40636v1 [cs.CV] for this version) | |
| https://doi.org/10.48550/arXiv.2610.40636 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Baiqiao Yin [view email]
[v1]
Thu, 29 Oct 2026 17:58:16 UTC (2,373 KB)
来源：arXiv cs.AI · arxiv.org