CodeMark | Uni-3DAR Unifies Micro and Macro 3D Worlds Through Autoregression, Outperforming Diffusion Models by 256% with 21.8x Faster Inference
From the molecular and material structures of the microscopic world to the geometry and spatial intelligence of the macroscopic world, creating and understanding 3D structures is a foundational pillar for advancing scientific research. 3D structures not only carry rich physical and chemical information, but also provide scientists with essential tools for deconstructing complex systems, conducting simulation-based predictions, and driving cross-disciplinary innovation. How to accurately and efficiently build 3D models, and how to understand and generate 3D worlds, is becoming a critical frontier for AGI and AI for Science.
