Berlin 2024 – scientific programme
Parts | Days | Selection | Search | Updates | Downloads | Help
DY: Fachverband Dynamik und Statistische Physik
DY 42: Focus Session: Computing with Dynamical Systems: New Perspectives on Reservoirs and Applications I – Fundamentals
DY 42.5: Talk
Thursday, March 21, 2024, 11:00–11:15, BH-N 243
Prediction of spatio-temporal chaos using parallel reservoir computing in combination with dimensionality reduction — Kai-Uwe Hollborn1, Luk Fleddermann1,2, •Gerrit Wellecke1, and Ulrich Parlitz1,2 — 1Max Planck Institute for Dynamics and Self-Organization, Göttingen, Germany — 2Institute for the Dynamics of Complex Systems, University of Göttingen, Germany
Reservoir computers can be used to predict time series of spatio-temporal chaotic systems. Using multiple reservoirs in parallel has shown improved performances for these predictions. Similarly, enhancements can be achieved by reducing the dimensionality of the input data. To utilise both options for better predictions, we combine parallel reservoirs with a dimensionality reduction of the input data by Fourier transformation or principal component analysis. The performance of this hybrid approach will be illustrated and evaluated for data generated by a one-dimensional Kuramoto-Sivashinsky equation and a two-dimensional Aliev-Panfilov model describing chaotic spiral wave dynamics.
Keywords: reservoir computing; dimension reduction; time series prediction; spatio-temporal chaos; echo state networks