Welcome to Jiajun Xu’s homepage

Welcome to my homepage!

I am currently a senior student in Tsinghua University, Computer Science and Technology Department.

My research interests lay on embodied AI, multi-modal and computer vision.

Specifically, I’m passionate about dexterous hand and humanoid. Besides, I’m interested in world model and VLA.

My specific research interests are:

• Using video generation models as world model to provide an environment for robot reinforcement learning.
• Exploring the latent space of transformers and fine-tuning attention to reduce multi-modal hallucinations.
• Exploring the understanding ability of diffusion models.

I have been looking forward to communicating and collaborating with others who share my interests, hoping to work together on interesting projects.

Meanwhile, I expect to graduate in 2026. If you are interested in me, you can check my cv.

Feel free to contact me!