Published multiple papers, such as 'A Theoretical and Empirical Analysis of Divide-and-Conquer Prompting in LLMs' (COLM 2025), 'InstaSHAP: Interpretable Additive Models Explain Shapley Values Instantly' (ICLR 2025), etc. Served as general chair or co-chair for several international conferences, including ICLR 2023 and ACM KDD 2020.
Research Experience
Was an assistant professor from 2010 to 2016 and an associate professor from 2016 to 2020. Served as a research staff member at the IBM T.J. Watson Research Center from 2006 to 2010. Currently, a full professor in the Thomas Lord Department of Computer Science at USC, and also the Director of the USC Machine Learning Center and co-Director of the USC Institute of Ethics and Trust in Computing.
Education
Received M.S. and Ph.D. from Carnegie Mellon University.
Background
Research interests include machine learning for time series, physics-informed machine learning, and interpretable machine learning, with applications to health, sustainability, and social media.
Miscellany
Some interesting stories about AI research work: 'The Perfect Stitch: Training for Surgery Using AI', 'USC Researchers Analyze Coronavirus Misinformation on Twitter', 'Fighting Addiction Before It Begins'.