AI RESEARCH
GEM: Generative Supervision Helps Embodied Intelligence
arXiv CS.CV
•
ArXi:2605.28548v1 Announce Type: new Embodied Vision-Language Models (VLMs) have nstrated impressive performance and generalization in robotics, particularly within Vision-Language-Action frameworks. However, a significant gap remains between the high-level semantic focus of standard text-guided pre-