Good and bad predictions: Assessing and improving the replication of chaotic attractors by means of reservoir computing

DOI10.1063/1.5118725zbMath1425.37025arXiv1907.05639OpenAlexW2959676983WikidataQ91057939 ScholiaQ91057939MaRDI QIDQ4973018

Christoph Räth, Alexander Haluszczynski

Publication date: 29 November 2019

Published in: Chaos: An Interdisciplinary Journal of Nonlinear Science (Search for Journal in Brave)

Full work available at URL: https://arxiv.org/abs/1907.05639

Mathematics Subject Classification ID

Strange attractors, chaotic dynamics of systems with hyperbolic behavior (37D45) Simulation of dynamical systems (37M05)

Related Items (11)

Robustness of LSTM neural networks for multi-step forecasting of chaotic time series ⋮ Modeling chaotic systems: dynamical equations vs machine learning approach ⋮ High-efficiency chaotic time series prediction based on time convolution neural network ⋮ Proper choice of hyperparameters in reservoir computing of chaotic maps ⋮ Robust optimization and validation of echo state networks for learning chaotic dynamics ⋮ Reducing network size and improving prediction stability of reservoir computing ⋮ Global optimization of hyper-parameters in reservoir computing ⋮ Path length statistics in reservoir computers ⋮ Breaking symmetries of the reservoir equations in echo state networks ⋮ Transfer learning of chaotic systems ⋮ Symmetry kills the square in a multifunctional reservoir computer

Uses Software

• NetworkX

Cites Work

• Unnamed Item
• Unnamed Item
• Unnamed Item
• Reservoir computing approaches to recurrent neural network training
• Testing for nonlinearity in time series: the method of surrogate data
• Determining Lyapunov exponents from a time series
• Measuring the strangeness of strange attractors
• A practical method for calculating largest Lyapunov exponents from small data sets
• An equation for continuous chaos
• The architecture of dynamic reservoir in the echo state network
• Statistical mechanics of complex networks
• Real-Time Computing Without Stable States: A New Framework for Neural Computation Based on Perturbations
• Using machine learning to replicate chaotic attractors and calculate Lyapunov exponents from data
• Deterministic Nonperiodic Flow
• Path length statistics in reservoir computers
• Collective dynamics of ‘small-world’ networks
• Ridge Regression: Biased Estimation for Nonorthogonal Problems

This page was built for publication: Good and bad predictions: Assessing and improving the replication of chaotic attractors by means of reservoir computing