Department of Automation, Tsinghua University
I am an undergraduate student in the Department of Automation at Tsinghua University. My current interests center on embodied intelligence, world models, world action models, visual-language-action systems, robotic foundation models, humanoid robotics, and whole-body control.
I am a member of Tsinghua's class of 2024 in Automation, with a minor in Statistics. I enjoy problems where perception, control, and learning meet: how embodied agents build usable representations, how world/action models support planning, and how humanoid systems can turn multimodal inputs into robust whole-body behavior.
My project experience spans discrete vision tokenizers, embodied benchmarks, active learning for compositional robotic tasks, tactile representation for robot control, deep-learning course projects, and quadruped control.
Undergraduate student, Department of Automation. Minor in Statistics.
Explores Sinkhorn optimal transport to mitigate codebook collapse in VQ-VAE style discrete vision tokenizers.
Builds an embodied-intelligence benchmark focused on critical intervals in agent behavior.
Investigates active learning for compositional robotic tasks, with an emphasis on efficient data collection and feedback-driven policy improvement.
Studies tactile representation and control for robotic whole-body locomotion and manipulation, using tactile-skin hardware to enrich the control modality.
A deep-learning project for image recognition tasks.
GitHubA practical tool for processing Excel files and automating spreadsheet workflows.
GitHubControl algorithms for a quadruped robot competition.
GitHub