Tianhong Li

Tianhong Li is a computer science researcher known for his work in generative models and representation learning. He completed his Ph.D. at MIT under Dina Katabi and is currently a postdoctoral associate at MIT CSAIL working with Kaiming He. He is best known for his research on masked generative modeling, autoregressive image generation without vector quantization, and the JiT framework, which applies computer-vision denoising insights to robot control. He received his B.E. from the Yao Class at Tsinghua University.