Embodied AI

Vision-Language-Action (VLA) Models: A Review of Recent Progress

Recent VLAs evolve from discrete to continuous, and from single-system (system 1 only) to dual-system.