Matthew Golden, Los Alamos National Laboratory
When
Refreshments served at 3:00 pm in the 3rd floor atrium
TAP Colloquium
Matthew Golden, Los Alamos National Laboratory
Visit Dates: October 26 - 28, 2026
Title: Fighting the Curse of Dimensionality in Kinetic Simulations
Abstract: Properly modeling kinetic phenomena in the magnetosphere requires solving the six-dimensional Vlasov-Maxwell equations. Discretizing this phase space requires enormous computational resources, a challenge known as the curse of dimensionality. I will discuss Time Dependent Bases (TDB) for the kinetic equations, where a small set of basis functions coevolve with the solution to minimize the error in the governing equations.
This approach can reduce memory requirements by orders of magnitude while retaining much of the solution’s accuracy. This tradeoff between memory and accuracy enables multi-fidelity modeling of kinetic processes, where the resources devoted to a problem are proportional to the desired accuracy. I will demonstrate the TDB formalism with several examples of increasing dimensionality.
Bio: Matthew earned his B.S. in Physics and Mathematics from Ohio Northern University in 2017, followed by a Ph.D. in Physics from Georgia Tech in 2023. His doctoral research focused on equation learning via sparse regression, with applications to simulated and experimental biological flows. From 2023 to 2026, he was a postdoctoral fellow in the Black Hole Astrophysics Group at Georgia Tech, working with Dimitrios Psaltis. He joined the Applied Mathematics and Plasma Physics group at Los Alamos National Laboratory in early 2026 as a postdoc, where his current work focuses on low-rank methods for the Vlasov-Maxwell equations. Matthew's research interests span nonlinear dynamics, relativistic astrophysics, machine learning, and turbulence.
Matthew Golden's "T-5" Website