1 paper
Zijian Zhang, Zhen Zeng, Zhongshu Gu +1
Learning-based models (e.g., visuomotor and Vision-Language-Action (VLA)) are increasingly explored for industrial robotic manipulation, where model predictions are directly transl…