Embodied-R1: Reinforced Embodied Reasoning for General Robotic Manipulation

ArXi:2508.13998v2 Announce Type: replace-cross Generalization in embodied AI is hindered by the "seeing-to-doing gap," which stems from data scarcity and embodiment heterogeneity. To address this, we pioneer "pointing" as a unified, embodiment-agnostic intermediate representation, defining four core embodied pointing abilities that bridge high-level vision-language comprehension with low-level action primitives. We