Machine Learning for System Identification and Parameter Estimation

Ho, Caitlin

Student Work

Machine Learning for System Identification and Parameter Estimation

Public Deposited

Incorporating a priori physics knowledge into machine learning leads to more robust and interpretable algorithms. In this work, we combine deep learning techniques with numerical methods to solve two problems in dynamical systems theory: dynamics discovery and parameter estimation. We present numerical results of applying these proposed approaches to highly oscillatory and chaotic problems. Finally, we compare the performance of various numerical schemes, such as Runge-Kutta and linear multistep families of methods, in predicting system dynamics and estimating physical parameters for these problems.

This report represents the work of one or more WPI undergraduate students submitted to the faculty as evidence of completion of a degree requirement. WPI routinely publishes these reports on its website without editorial or peer review.

Creator