Dynamical Measure Transport
I am interested in continuous-time formulations of measure transport, including velocity-field descriptions, flow maps, variational structure, and computational methods for moving mass between distributions.
Research
Dynamical views of transport, sampling, and control.
My interests center on how probability measures move under dynamics, and how those dynamics can be designed, analyzed, and computed for sampling and related inference problems.
Sampling
Sampling problems provide a natural setting where transport, stochastic dynamics, and optimization meet. I am especially interested in methods that make the geometry of the target distribution explicit.
Optimization and Control
Many transport and sampling questions have useful control-theoretic or optimization-theoretic formulations. These viewpoints are part of my broader collaboration with Pablo Parrilo, Asu Ozdaglar, and Russ Tedrake.