Weihan Xu

I am a researcher at MIT EECS and MIT Media Lab. I am broadly interested in multimodal learning in creative application. My research vision is to build multimodal systems with user-friendly interfaces that can reason, retrieve, generate, and edit content, ultimately making advanced AI capabilities accessible and useful in everyday life.

I have been working closely with researchers at Sony Research, MIT EECS, and Berkeley AI Research on multimodal content stylization.

I have been working with Dr. Julian McAuley, Dr. Taylor Berg-Kirkpatrick, Dr. Hao-Wen Dong, Dr. Paul Liang, and Dr. Shlomo Dubnov on multimodal retrieval and content generation.

I was fortunate to work with Dr. Cynthia Rudin on interpretable music analysis, and Dr. Pardis Emami-Naeini on human-AI interaction.

Previously, I did my undergraduate at the University of Michigan with a double major in computer science (honors) and data science, where I worked with Dr. Sardar Ansari and Dr. Kayvan Najarian on time-series analysis in clinical settings. I also worked with Dr. Gongjun Xu on statistical analysis of educational assessment data.

I began studying music at six and have training in piano, violin, and the French horn.