Automatic Plasma Processes Detection with Supervised Machine Learning: Application to Plasma Jet Dipolarization Fronts in Near-Earth Space using MMS Data

The detection of plasma processes in near-Earth space is crucial to perform unambiguous statistical studies of fundamental plasma processes such as shocks, magnetic reconnection, waves and turbulence, jets and their combinations. The majority of available studies have been performed by using human-driven methods, such as visual data selection or the application of predefined thresholds to different observable plasma quantities. While human-driven methods have allowed performing many statistical studies, these methods are often time-consuming and can introduce important biases. On the other hand, the recent availability of large, high-quality spacecraft databases, together with major advances in machine-learning algorithms, now allow meaningful applications of machine learning to in situ plasma data. In this study, we apply a fully convolutional neural network (FCN) deep machine-leaning algorithm to the recent Magnetospheric Multi Scale (MMS) mission data in order to classify dipolarization fronts in the Earth's magnetotail. For this purpose, we use available intervals of time series which were labeled as such by using human-driven selective downlink applied to MMS data. We discuss the data selection and the chosen model and its parameters. Our results show that the FCN method employed is reliable to identify dipolarization fronts in the time series data since it takes into account the dynamical features of the plasma. We also show that the model is able to detect dipolarization fronts in time series not labeled as such, indicating that such method can be potentially applied to other in situ spacecraft plasma process.