Modeling and Simulation Group Meeting

Quantifying extrinsic and intrinsic noise using Wasserstein distance

Time and Location:

Sept. 19, 2024 at 12:30PM; Warren Weaver Hall, Room 202

Speaker:

Mingtao Xia, NYU

Abstract:

In this talk, I will introduce my recent work on developing Wasserstein-distance-based approaches for efficient reconstruction of extrinsic noise (parameter heterogeneity) and intrinsic fluctuations in uncertainty quantification (UQ) with potential applications in reconstructing noisy cell dynamics. I shall discuss the application of Wasserstein distance in reconstructing uncertainty models of different types such as models subject to noise, models with latent parameters, as well as stochastic differential equations with unknown terms. (Note: This is one of my tentative presentation topics for my future job talk and I will greatly appreciate your coming and providing feedback for me.)