跳到正文
研踪PaperTrail
← 研踪 PaperTrail

arXiv:2611.26565 · v1

PhyDEX: A Benchmark for Physical Reasoning in Vision-Language-Action Models via Dynamic Extrinsic Dexterity

PhyDEX:通过动态外在灵巧性评估视觉-语言-动作模型物理推理能力的基准

首次提交 2026-11-20 02:33
版本更新 2026-11-20 02:33
雷达观测 2026-11-20 12:52
arXiv ↗PDF ↗

摘要

Vision-Language-Action (VLA) models have advanced rapidly in language-conditioned robotic manipulation, yet existing benchmarks mainly evaluate gripper--object contact and largely ignore how objects interact with the environment through gravity, friction, inertia, and mass distribution. As a result, it remains unclear whether current policies truly understand physics or merely interpolate from their training data. We address this gap through the lens of dynamic extrinsic dexterity and introduce PhyDEX, a throwing benchmark in which a robot must infer an object's mass, center of mass, and friction to send it to a destination beyond its kinematic workspace. PhyDEX contains 35K trajectories (143 hours) across nine tasks and 100 scenes, covering uniform, irregular, and dynamic-mass objects with static, multiple, and moving destinations. All data are generated by a scalable, fully automatic pipeline that combines grasp generation, motion planning, and closed-loop throw optimization under randomized physical properties. We evaluate three representative VLAs, GR00T N1.7, DynamicVLA, and ForceVLA, and find that all achieve success rates below 3\%. Our analysis reveals two key bottlenecks: limited robustness of pretrained models, which causes most failures before the throw, and physical blindness, as providing ground-truth physical properties roughly doubles success. Experiments on a humanoid robot and in the real world show the same failure patterns. We hope PhyDEX will drive the development of physically grounded robot policies.

AI 论文解读

基于摘要 · 非全文审阅

作者提出一个以抛掷任务为核心的基准 PhyDEX,用于检验视觉-语言-动作模型能否推断物体质量、质心与摩擦等物理属性,并发现现有代表性模型成功率均低于3%。

正在读取账号与研究记录…

版本差异

目前只收录了一个版本。收录下一版后,可以对照标题、作者、分类和摘要。

观测到的版本

相邻记录可能跳过中间版本;首次发现 v2 及以上也会标为版本更新。

v1 · 2026-11-20 02:33 · 最新观测版本

PhyDEX: A Benchmark for Physical Reasoning in Vision-Language-Action Models via Dynamic Extrinsic Dexterity

Vision-Language-Action (VLA) models have advanced rapidly in language-conditioned robotic manipulation, yet existing benchmarks mainly evaluate gripper--object contact and largely ignore how objects interact with the environment through gravity, friction, inertia, and mass distribution. As a result, it remains unclear whether current policies truly understand physics or merely interpolate from their training data. We address this gap through the lens of dynamic extrinsic dexterity and introduce PhyDEX, a throwing benchmark in which a robot must infer an object's mass, center of mass, and friction to send it to a destination beyond its kinematic workspace. PhyDEX contains 35K trajectories (143 hours) across nine tasks and 100 scenes, covering uniform, irregular, and dynamic-mass objects with static, multiple, and moving destinations. All data are generated by a scalable, fully automatic pipeline that combines grasp generation, motion planning, and closed-loop throw optimization under randomized physical properties. We evaluate three representative VLAs, GR00T N1.7, DynamicVLA, and ForceVLA, and find that all achieve success rates below 3\%. Our analysis reveals two key bottlenecks: limited robustness of pretrained models, which causes most failures before the throw, and physical blindness, as providing ground-truth physical properties roughly doubles success. Experiments on a humanoid robot and in the real world show the same failure patterns. We hope PhyDEX will drive the development of physically grounded robot policies.

查看此版本解读 →查看此版本原文 ↗