About Me 🤔

Before turning to computer science in my junior year, I spent two years studying developmental psychology and the history of philosophy. I earned a B.S. in Computer Science and a B.A. in Psychology at the University of Rochester, where I worked with Prof. Chenliang Xu and Prof. Dongmei Li. I’m especially grateful to Jing Bi and the other PhD students at the University of Rochester.

My research asks how models can turn visual experience into understanding and action. I view perception, memory, reasoning, and action as a continuous loop: gathering evidence, retaining useful experience, reasoning about possibilities, and choosing how to act. I am especially interested in how visual representations can support this process under limited memory and computation.

I see many of the core questions in robotics as questions of video understanding: how the world changes, what actions are possible, and which actions bring us closer to a goal. I think of video intelligence as the soul of a robot. My goal is to build models that can reason from video to action, using what they observe and remember to infer what to do next, anticipate its consequences, and adapt as new evidence arrives.

Conference reviewer: CVPR’26, ECCV’26, AAAI’27, NeurIPS’26, CoRL’26, COLM’26.

Previously, I served as Vice President of the Chinese Students and Scholars Association (CSSA) at the University of Rochester.

Outside research, I enjoy cooking dishes from around the world, experimenting with new recipes, and mixing cocktails. Cooking together, sharing a drink, and playing badminton are some of my favorite ways to spend time with friends. I also love spending time with my dog, Maple.