arXiv Robotics — research abstracts· Nikolaos Tsagkas, Andreas Sochopoulos, Chris Xiaoxuan Lu, Oisin Mac Aodha, Alexandros Kouris·· 15 hours agoEditorial score50
Separating Knowledge Types in Vision-Language-Action Models
Decoupling the Declarative from the Procedural in Vision-Language-Action Models
Summary
This study introduces w$^{2}$VLA, a model that decouples declarative and procedural knowledge in VLA systems, enabling better zero-shot skill transfer to new objects. The approach improves interpretability and robustness in robotic tasks.
Source: arXiv Robotics — research abstracts · Read original article ↗
Loading article text…
Source：arXiv Robotics — research abstracts · arxiv.org