Full Professor, University of California, Davis
2 papers at NeurIPS 2025
We introduce an invertible Wasserstein slicing map that embeds multivariate distributions into a Hilbert space, enabling functional data analysis with theoretical guarantees and application to real-world data.
We propose FGBoost, a gradient boosting framework designed to intrinsically model complex regression relationships with non-Euclidean outputs in geodesic metric spaces.
