When
Where
Speaker: Teddy Meissner, Program in Applied Mathematics
Title: System Identification for Differential Equations
Abstract: Understanding the underlying equations that govern a system’s behavior is essential for modeling, prediction, and control. System identification seeks to infer these governing differential equations directly from data, making it a powerful tool for studying complex dynamical systems. This process involves optimization-based methods, including sparse regression, and regularization techniques to handle noisy or incomplete data.
Naturally, this field builds upon many classical parameter estimation techniques. I will discuss how these frameworks can be used to extract interpretable models, the challenges of fitting differential equations to data, and connections to methods like multiple shooting.