Yangyan Li (李扬彦)

I was a professor in Shandong University, then a Staff Algorithm Engineer of Alibaba Group. Before that, I was a postdoctoral research fellow in Stanford University and Tel Aviv University, supervised by Prof. Leonidas J. Guibas and Daniel Cohen-Or, respectively. I received my PhD degree from University of the Chinese Academy of Sciences, supervised by Prof. Baoquan Chen, and B.S. degree from Sichuan University.

Research Interests:

(Note: This summary was generated by an LLM based on my publications as of November 2025.)

My current research focuses on the convergence of Computer Graphics, Computer Vision, and Artificial Intelligence, aiming towards the realization of general-purpose Spatial Intelligence. While my background is rooted in 3D geometry processing and visual perception, my recent work explores how Large Language Models (LLMs) and Transformer architectures can be adapted to reason about the physical world. This involves two parallel tracks: advancing 3D vision through generative techniques that reduce reliance on explicit geometric priors (such as novel view synthesis and Gaussian splatting), and simultaneously refining the underlying architecture of Foundation Models to ensure they are robust, efficient, and safe enough to serve as the cognitive engines for spatial agents. By integrating the semantic reasoning capabilities of LLMs with the precise structural understanding of 3D vision, I aim to build embodied systems capable of perceiving, understanding, and interacting with complex environments in a physically grounded manner.