September 2024, two papers accepted to NeurIPS '24; March 2024, one paper accepted to ICLR '24; January 2024, gave an oral talk at CPAL '24. Involved in research projects including: 'Concept Lancet: Representation Decomposition and Transplant for Diffusion-Based Image Editing', 'Token Statistics Transformer: Linear-Time Attention via Variational Rate Reduction', and 'PaCE: Parsimonious Concept Engineering for Large Language Models'.
Research Experience
Before joining the University of Pennsylvania, he was a research assistant at ShanghaiTech University, working closely with Professor Manolis C. Tsakiris and collaborating with Laurent Kneip.
Education
PhD student at the University of Pennsylvania, advised by René Vidal; Research Assistant at ShanghaiTech University, advised by Manolis C. Tsakiris and collaborating with Laurent Kneip; Master’s in Applied Mathematics and Statistics from Johns Hopkins University; Bachelor’s in Computer Science with honors from ShanghaiTech.
Background
Research interests center on the theoretical foundations of machine learning and emerging applications. On one hand, he uses mathematics to understand when and why existing empirical paradigms work. On the other hand, these insights allow him to develop practical algorithms that are more robust, trustworthy, and efficient.
Miscellany
Open to chatting about research, advising, collaborations, life, and fun. For undergraduate and MS students interested in doing research with him, a recommended time investment is at least 15 hours per week. Students he has mentored have gone on to PhD programs at UC Berkeley, Hong Kong University, MIT, NYU, and to full-time roles at Google and Meta.