Extreme events in globally coupled chaotic maps

Abstract

Understanding and predicting uncertain things are the central themes of scientific evolution. Human beings revolve around these fears of uncertainties concerning various aspects like a global pandemic, health, finances, to name but a few. Dealing with this unavoidable part of life is far tougher due to the chaotic nature of these unpredictable activities. In the present article, we consider a global network of identical chaotic maps, which splits into two different clusters, despite the interaction between all nodes are uniform. The stability analysis of the spatially homogeneous chaotic solutions provides a critical coupling strength, before which we anticipate such partial synchronization. The distance between these two chaotic synchronized populations often deviates more than eight times of standard deviation from its long-term average. The probability density function of these highly deviated values fits well with the generalized extreme value distribution. Meanwhile, the distribution of recurrence time intervals between extreme events resembles the Weibull distribution. The existing literature helps us to characterize such events as extreme events using the significant height. These extremely high fluctuations are less frequent in terms of their occurrence. We determine numerically a range of coupling strength for these extremely large but recurrent events. On-off intermittency is the responsible mechanism underlying the formation of such extreme events. Besides understanding the generation of such extreme events and their statistical signature, we furnish forecasting these events using the powerful deep learning algorithms of an artificial recurrent neural network. This long short-term memory (LSTM) can offer handy one-step forecasting of these chaotic intermittent bursts. We also ensure the robustness of this forecasting model with two hundred hidden cells in each LSTM layer.

Prediction of chaotic attractors in quasiperiodically forced logistic map using deep learning

“…For the present study we consider a DL framework known as LSTM [21] which is a special kind of RNNs. In recent years, LSTM framework has proven to be capable of forecasting time series of the chaotic systems even when there are extreme events in the time series [10,13,14]. The main feature that differentiates LSTM from the other RNNs is that the latter has only one activation function for the neurons that is tanh but in the case of the former, a sigmoid function is used for recurrent activations and tanh is used for the activation of neurons.…”

“…In the study of dynamics of nonlinear systems, ML and DL algorithms are extensively used for the prediction and discovery of the behaviour of the chaotic and complex systems. For example, they have been used to identify chimera states [4,5], in the replication of chaotic attractors [6], using symbolic time series for network classification [7], separating chaotic signals [8], learning dynamical systems in noise [9] and in the prediction of extreme events [10][11][12][13][14][15]. Very recently, the authors of Ref.…”

Swarmalators under competitive time-varying phase interactions

“…Such signed networks can display fascinating macroscopic dynamics [57,63,64], including the π state, the traveling wave state, and the mixed state. Recently, the impact of such competitive interactions through the concurrence of positive-negative coupling has been investigated on the time-evolving networks of mobile agents [43,65], leading to diverse peculiar dynamical states, including extreme events [66,67]. However, these studies consider only the unidirectional influence of spatial dynamics toward the oscillator's amplitude and phase dynamics.…”

Data driven soliton solution of the nonlinear Schrödinger equation with certain $\mathcal{PT}$-symmetric potentials via deep learning

“…Nowadays Machine Learning (ML) and Deep Learning (DL) approaches have become important tools in the prediction task in various fields of physics [15][16][17][18] . In the field of nonlinear dynamics, ML methods have been used for the replication of chaotic attractors 19 , prediction of chaotic laser pulses amplitude 20 , detection of unstable periodic orbits 21 , chaotic signals separation 22 , network classification from symbolic time series 23 , identification of chimera states [24][25][26] and also in the study of extreme events [27][28][29][30][31] .…”