HiFOR: Hierarchical Foresight with Action-Guided World Grounding for Vision-Language-Action Model

Rongxing Ding, Hongyu Qu, Kehong Wu +4

Vision-Language-Action (VLA) models have demonstrated promising capabilities in robotic manipulation by integrating visual perception, language understanding, and action generation…