​I aim to push the computational- and data-efficiency of generative models such as Diffusion Models, Large-language Models and Vision-Language Models, which apply to image generation, video generation, 3D generation and textual generation (e.g., QA and caption etc).