
Peking University
Vision-Language-Action Models, Generative Models, Robotic Manipulation
2025 – Present
B.S. Candidate — Computer Science & Materials Science
I work on generative models and vision-language-action systems — toward machines that can perceive, reason, and act in the physical world.
Stanford Bunny — an interactive point-cloud study
Generative AI for embodied intelligence. I am interested in connecting generative AI with robotics, particularly through vision-language-action models that integrate perception, reasoning, and physical action.
Robot learning and manipulation. My work explores how multimodal representations and generative models can help robots understand instructions, interact with their environment, and perform robust manipulation tasks.

Vision-Language-Action Models, Generative Models, Robotic Manipulation
2025 – Present
Mathematical Reasoning, LLM Inference
2025 – 2026

Generative Models
2024 – 2025