Vision-Language-Action Models: From CLIP to Humanoid Robots19 February 2026Robotics VLA Vision-Language-Action Foundation Model Robot Learning RT-2 OpenVLA Pi-0