Yang Li


Welcome to my homepage! I'm an assistant professor at the Internet of Things and Societal Cyber Physical System Lab of Tsinghua Berkeley Shenzhen Institute (TBSI).

My overall research interest is learning shared structures from uncertain and heterogeneous data. I exploit such structures in both data-driven algorithm design, such as the collective analysis of GPS trajectories, and in machine learning, such as multi-modal learning and transfer learning. Applications of my research cover a wide range of problems, including

  • image-based computer vision & semantics, medical image processing

  • mobility computing and smart transportation

  • interactions of mobile agents

My current work focuses on incorporating information theory and statistics concepts into transfer learning, where we aim to answer fundamental questions such as how transferable are features between different tasks and domains, and how to learn robust classifiers from different source domains.

I obtained my PhD. degree in Computer Science from Stanford University in 2017, advised by Leonidas Guibas. My dissertation explores how to find shared structures in large GPS trajectories under uncertainty, and how the shared structures can be used to solve challenging problems. See the Research page for details about my past projects. I was also a contributor to the KINARI project.

Curriculum Vitae






Information Building, Rm 1108A
Tsinghua Shenzhen International Graduate School
Shenzhen, Guangdong, China (518055)