Two papers were accepted to NeurIPS'25; one paper (LightGaussian) was selected as a spotlight presentation at NeurIPS'24; published a research paper on DynamicVerse, which introduces a physically-aware multimodal framework for 4D world modeling; also published a research paper on InstantSplat, which presents a method for sparse-view SfM-free Gaussian splatting in seconds.
Research Experience
Has research experience with leading research groups across academia and industry, including VITA-Group@UT Austin and InternRobotics@Shanghai AI Lab.
Education
Incoming Ph.D. student at MMLab, The Chinese University of Hong Kong, advised by Prof. Hongsheng Li; currently pursuing a master’s degree in the School of Informatics at Xiamen University, advised by Prof. Xinghao Ding.
Background
Research interests include 4D scene understanding, efficient 3D modeling, 3D perception, and embodied AI agents. The long-term goal is to develop an efficient and scalable embodied agent that can continuously perceive, understand, and interact with the physical world through multi-modal spatial intelligence.