Research Area
My primary research area is the design and development of numerical schemes for approximating stochastic differential equations (SDEs), and perform convergence analysis. These SDEs may be subject to boundary conditions, evolve on manifolds, be interlaced with jumps, or exhibit distribution dependence (as in mean-field SDEs). The resulting numerical algorithms have applications in sampling problems arising in Bayesian statistics or molecular dynamics, as well as in non-convex global optimization. I use tools from numerical analysis, stochastic analysis, and geometry, with motivation coming from statistical and optimization problems.
Stochastic Numerics
Numerical integration of SDEs with applications to solving boundary value problems.
Sampling
Developing methods for sampling on manifolds or in bounded domains by utilizing Langevin diffusion and non-reversible Markov jump processes.
Optimization
High-dimensional non-convex optimization via interacting particle system based methods or via controlled processes.