Extreme events prediction from nonlocal partial information in a spatiotemporally chaotic microcavity laser

Abstract

The forecasting of high-dimensional, spatiotemporal nonlinear systems has made tremendous progress with the advent of model-free machine learning techniques. However, in real systems it is not always possible to have all the information needed; only partial information is available for learning and forecasting. This can be due to insufficient temporal or spatial samplings, to inaccessible variables or to noisy training data. Here, we show that it is nevertheless possible to forecast extreme events occurrence in incomplete experimental recordings from a spatiotemporally chaotic microcavity laser using reservoir computing. Selecting regions of maximum transfer entropy, we show that it is possible to get higher forecasting accuracy using nonlocal data vs local data thus allowing greater warning times, at least twice the time horizon predicted from the nonlinear local Lyapunov exponent.

References in corpus (10)

• Measuring Information Transfer
• Data-Driven Forecasting of High-Dimensional Chaotic Systems with Long Short-Term Memory Networks
• Spatiotemporal chaos induces extreme events in an extended microcavity laser
• Machine-learning inference of fluid variables from data using reservoir computing
• Machine learning algorithms for predicting the amplitude of chaotic laser pulses
• Early Detection of Rogue Waves by the Wavelet Transforms
• Predictability of Extreme Intensity Pulses in Optically Injected Semiconductor Lasers
• Features of the extreme events observed in the all-solid state laser with a saturable absorber
• Precursor-driven machine learning prediction of chaotic extreme pulses in Kerr resonators
• Extreme Events related with spatial patterns in an all-solid-state laser with saturable absorber