IMFUFA Seminar: Inferring Dynamical Properties from Time Series Data

Rok Cestnik:

Inferring Dynamical Properties from Time Series Data

In this talk, I will present approaches for inferring dynamical properties directly from time series data. The main emphasis will be on oscillatory dynamics, with a theme of estimating phase dynamics from observations. I will first show methods that assume a phase description and fit it directly. Then I will move to machine learning approaches, where a model is trained to reproduce the system and can then be used to infer dynamical properties. I will highlight recent work with reservoir computing as a simple and versatile tool for modeling and inference.
 

IMFUFA is the section/research group for mathematics and physics under Department of Science and Environment.