arXiv cs.AI· Denis Grachev, Damien LaRocque, Omar Moured·· 10 小时前AI 评分48
你的 VLA 模型有多少是 kNN？研究揭示 LIBERO 高分背后的检索依赖
How Much of Your VLA Is kNN
AI 导读
在 LIBERO 仿真基准上，直接检索邻近演示并复用其动作块，无需策略训练即可达到 91.7% 成功率，而现有 VLA 模型成功率已超 95%。对九个公开 VLA 的评估显示，微调主要让视觉-语言主干成为更好的检索键，其动作仍接近演示数据；在 LIBERO-Plus 扰动下检索方法的性能下降最小。
正文 · 原文
Abstract:Vision-language-action (VLA) models now exceed 95% success on the LIBERO simulation benchmark, yet it remains unclear how much of this performance requires generalization beyond the training demonstrations. In this work, we show that retrieving nearby demonstrations and reusing their action chunks achieves a success rate of 91.7 % on LIBERO without policy training. Under LIBERO-Plus perturbations, retrieval shows the smallest drop of all evaluated policies. Across nine publicly available VLAs, fine-tuning mostly makes the vision-language backbone a better retrieval key, and the actions of the VLAs remain close to the demonstrations. In addition, we introduce Local Action Consistency (LAC), an offline metric that measures whether nearby observations in representation space are also associated with similar future actions. We show that LAC tracks the success rate of retrieval and can guide the training of a retrieval key. Our evaluation and source code are available at this https URL.
| Comments: | 21 pages, 10 figures, accepted to the CoRL 2026 G2A/TAG workshop |
| Subjects: | Robotics (cs.RO); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2611.13160 [cs.RO] |
| (or arXiv:2611.13160v1 [cs.RO] for this version) | |
| https://doi.org/10.48550/arXiv.2611.13160 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Damien LaRocque [view email]
[v1]
Mon, 9 Nov 2026 19:44:58 UTC (2,159 KB)
来源：arXiv cs.AI · arxiv.org