外观
论文HF Daily Papers · arXiv 2610.316706 天前
Sensorimotor Alignment：免训练离线指标预测具身策略成功率
本文由 AI 根据公开内容整理，可能有误，请以原文为准：Sensorimotor Alignment: A Platonic Proxy for Embodied Policy Success（HF Daily Papers · arXiv 2610.31670）
核心内容
针对机器人基础模型闭环评测成本高昂的痛点，研究基于柏拉图表征假说（PRH）提出Sensorimotor Alignment (S²)指标。该指标无需额外训练即可离线计算，旨在量化模型感官表征与物理运动轨迹结构的几何对齐程度。实验证实，S²分数在不同模型家族中与下游机器人任务成功率呈显著正相关，为在昂贵的真机或仿真闭环测试前筛选高潜力模型提供了可靠的低成本代理方案。
技术层面，S²将运动轨迹视为独立模态，通过度量感知空间与动作空间的统计结构一致性来评估模型能力。该方法不仅能在微调前识别出适合转化为视觉-语言-动作模型（VLA）的优质视觉-语言模型（VLM）底座，还能从表征维度解释VLA与世界动作模型（WAMs）的性能差异。这一进展将PRH理论扩展至具身智能领域，弥补了仅依赖任务成功率无法洞察模型内部表征质量的评估盲区。
关键看点
- S²指标定义：免训练离线度量模型感官表征与物理运动轨迹几何结构的对齐程度，作为柏拉图表征假说在机器人领域的延伸验证
- 性能预测有效性：跨模型家族实验显示S²分数与下游任务成功率正相关，支持在闭环评测前进行低成本模型优先级排序
- VLA底座选型：可在高成本微调前评估VLM骨干网络潜力，辅助筛选适合转化为视觉-语言-动作模型的预训练基座
- 表征级可解释性：超越单一成功率指标，从内部表征结构层面解析VLA与世界动作模型的能力来源及性能瓶颈
工程与产业观察
S²为具身大模型研发提供了关键的“预筛选”工具，解决了VLA迭代中盲测成本高、周期长的工程痛点。通过将抽象的表征质量转化为可量化的几何对齐指标，算法工程师可在数据飞轮启动前快速剔除劣质底座，显著提升从VLM到VLA的转化效率与研发ROI。
英文摘要 (Abstract)
Robot foundation models are typically evaluated through costly closed-loop rollouts, motivating low-cost offline proxies for model capability. The Platonic Representation Hypothesis (PRH) suggests that as representations improve, modalities such as vision and language converge toward a shared statistical model of reality. We ask whether motion trajectories form another such modality: as embodied models become more capable, should their sensory representations converge toward the structure of physical motion? To test this, we propose Sensorimotor Alignment (S^2), a training-free offline metric that measures how well a model's sensory representation aligns with the geometry of physical motion. We find that S^2 correlates positively with downstream robot task success across model families, making it a low-cost offline proxy for prioritizing models before closed-loop evaluation. S^2 can also identify promising vision-language model (VLM) backbones before costly fine-tuning into vision-language-action models (VLAs). Beyond this correlation, S^2 offers representation-level explanations for phenomena observed in VLAs and world action models (WAMs). With S^2, we extend PRH to robotics and make sensorimotor alignment measurable, offering a low-cost, representation-level view of robot foundation models that closed-loop success rates alone cannot reveal.
Hugging Face Daily Papers 热度：33 upvotes 作者：Yunfei Xie, Howard H. Qian, Shiyi Lan, Kaiyu Hang, Chen Wei
来源
本页事实对应的公开页面，以原文为准。