Large Language Models, Multi-modal Learning, 3D Perception and Understanding, Self-supervised Learning