CSE Community Seminar | November 1, 2024

Abstract
Numerical algorithms for parabolic PDE underlie many techniques dealing with geometric domains. The most popular PDE solvers in geometric data processing, however, favor linear PDE and deterministic solutions. In this talk, I will describe efforts to design tools for solving parabolic PDE beyond these favored regimes — in what I call the differential equations multiverse. Specifically, I will cover a convex optimization strategy tackling nonlinearity on curved triangle meshes, as well as recent work involving the stochastic counterpart to one of these equations. This talk will cover collaborative research with Oded Stein, Silvia Sellán and Justin Solomon.