Machine learning approaches for Kuramoto coupled oscillator systems

Abstract

Recently, there has been significant advancement in the machine learning (ML) approach and its application to diverse systems ranging from complex to quantum systems. As one of such systems, a coupled-oscillators system exhibits intriguing collective behaviors, synchronization phase transitions, chaotic behaviors and so on. Even though traditional approaches such as analytical and numerical methods enable to understand diverse properties of such systems, some properties still remain unclear. Here, we applied the ML approach to such systems particularly described by the Kuramoto model, with the aim of resolving the following intriguing problems, namely determination of the transition point and criticality of a hybrid synchronization transition; understanding network structures from chaotic patterns; and comparison of ML algorithms for the prediction of future chaotic behaviors. The proposed method is expected to be useful for further problems such as understanding a neural network structure from electroencephalogram signals.