What question did this study set out to answer?

Determine whether parallel reservoir computing can achieve model-free forecasting of high-dimensional, large-scale spatiotemporally chaotic systems using only historical state measurements.

January 12, 2018Physical Review Letters

Model-Free Prediction of Large Spatiotemporally Chaotic Systems from Data: A Reservoir Computing Approach

View Full Paper

Ask AI

Bookmark

Authors

JPJaideep PathakNvidia (United Kingdom)BHBrian R. HuntMIT University MGMichelle GirvanSanta Fe Institute

Discussion

Member takes

Implication

Computational study demonstrates model-free prediction in large spatiotemporally chaotic systems using parallel reservoir computing, indicating high scalability for complex physical dynamics.

Key Points

Determine whether parallel reservoir computing can achieve model-free forecasting of high-dimensional, large-scale spatiotemporally chaotic systems using only historical state measurements.
Designed a parallelized machine learning architecture based on the reservoir computing framework to handle distributed spatial dynamics.
Tested the system's predictive performance and scalability using time-series data generated from the Kuramoto-Sivashinsky equation across large spatial domains and high attractor dimensions.
Demonstrated accurate, model-free time-series forecasting of spatiotemporal chaotic dynamics without relying on knowledge of the governing physical equations.
Confirmed that the parallel reservoir computing scheme scales efficiently to systems with arbitrarily large spatial extents and attractor dimensions.

Cite This Study

Pathak et al. (2018) studied this question.

synapsesocial.com/papers/69d8d89e183921ebcaae3dea https://doi.org/10.1103/physrevlett.120.024102

View Full Paper

Ask AI

Bookmark

Implication

Computational study demonstrates model-free prediction in large spatiotemporally chaotic systems using parallel reservoir computing, indicating high scalability for complex physical dynamics.

Model-Free Prediction of Large Spatiotemporally Chaotic Systems from Data: A Reservoir Computing Approach

Discussion

Member takes

Implication

Key Points

Cite This Study

Model-Free Prediction of Large Spatiotemporally Chaotic Systems from Data: A Reservoir Computing Approach

Discussion

Member takes

Implication

Key Points

Cite This Study

Also Consider

Also Consider