外观
论文arXiv 2610.1470418 天前
PRISM：面向通用视觉-语言-动作策略的精度强制
原文链接：PRISM: Precision Forcing for Generalist Vision-Language-Action Policies（arXiv 2610.14704 · 发布日期：2026-10-10）
PRISM: Precision Forcing for Generalist Vision-Language-Action Policies
针对VLA精密操作不可靠，PRISM按精度需求重加权监督并内化表征，在仿真和真机插接、拧紧任务验证。
核心要点
- 精密操作中，现有VLA在亚毫米级插入等任务上易失效。
- 作者将原因归结为跨上下文与流时间的监督不均衡，即Flow-Time Starvation。
- PRISM离线评估精度需求，强化关键误差监督，并将需求内化到VLM条件表征。
- 仿真和真机实验中，平均端到端成功率由38.3%提升至86.7%，且无需新传感器或观测模态。
验证范围与局限：验证范围为仿真和真机的驱动连接器插入、双臂连接器配合、螺丝拧紧任务。
英文摘要 (Abstract)
Coupling a pretrained vision-language model (VLM) with a diffusion or flow-matching action expert has become the mainstream vision-language-action (VLA) paradigm for embodied policies, yet they remain unreliable once a task demands precise manipulation, e.g., submillimeter cable insertion. We attribute this weakness to a training objective that averages flow-matching errors across contexts and flow times, allocating little supervision to the few decisive moments determining task success. Along the generative path, inference integrates a locally supervised velocity field, so equally penalized errors at different flow times can affect the executed action unequally. Consequently, precision-task performance deteriorates as physical tolerance tightens by orders of magnitude. We call this supervision imbalance across contexts and flow times Flow-Time Starvation. The remedy is to jointly align supervision allocation and representation learning with context-dependent precision demand, enabling the policy to recognize when greater precision is needed. We instantiate this precision-forcing strategy as Precision-aware Reweighting and Internalization for Skilled Manipulation (PRISM), which scores precision demand offline, strengthens flow supervision where errors are consequential and action signals are visible, and internalizes that demand in the VLM condition representation. Simulation and real-robot experiments show that PRISM preserves or improves general-task performance while raising mean end-to-end success from 38.3% to 86.7% across drive connector insertion, bimanual connector mating, and screw tightening, substantially extending the precision boundary of existing embodied policies without new sensors or observation modalities.
作者：Chenghao Yue, Peiyan Li, Hao Wang, Lei Lv, Jiahao Chen, Huaping Liu, Yu Luo, Yuanlong Zhang
来源
本页事实对应的公开页面，以原文为准。