My current research focuses on the interdisciplinary area of generative AI and physics-based simulation. To be more specific, I aim to enhance current image/video generative models with real-world physics knowledge, enabling
the generation of more physically plausible contents.
PhysMotion: Physics-Grounded Dynamics From a Single Image Xiyang Tan*,
Ying Jiang*,
Xuan Li* (Equal Contributions),
Zeshun Zong,
Tianyi Xie,
Yin Yang,
Chenfanfu Jiang
arXiv, 2024
project page /
arXiv
PhysMotion is a novel framework that leverages principled physics-based simulations to guide intermediate 3D representations generated from a single image and input conditions (e.g., applied force and torque), producing high-quality, physically plausible video generation.
Real-Time Ocean Surface Simulation Based on GPGPU and IFFT Xiyang Tan,
advisor:
Renjie Chen Bachelor Thesis (written in Chinese), 2023
We apply Inverse Fast Fourier Transform (IFFT) algorithm on the Phillips spectrum on GPU to simulate ocean surface in real-time.
Miscellaneous
I'm a huge fan of sports, especially badminton. I was trained professionally as a badminton player since I was 8, and I was entitled national second-class badminton player
back in high school. I won the third place in Men's Doubles at the Youth Badminton Tournament of Sichuan Province, and I won gold medal in BUCS badminton winter tournament as a member of
LU Badminton men's first team.
I also enjoy other sports like volleyball, basketball and Formula 1. I was a starting setter at USTC volleyball team, and we won second place in University Volleyball Tournament of Anhui Province.