Transfer Learning in Spatial-Temporal Forecasting of the Solar Magnetic Field

Abstract

Machine learning techniques have been widely used in attempts to forecast several solar datasets. Most of these approaches employ supervised machine learning algorithms which are, in general, very data hungry. This hampers the attempts to forecast some of these data series, particularly the ones that depend on (relatively) recent space observations. Here we focus on an attempt to forecast the solar surface longitudinally averaged radial magnetic field distribution using a form of spatial-temporal neural networks. Given that the recording of these spatial-temporal datasets only started in 1975 and are therefore quite short, the forecasts are predictably quite modest. However, given that there is a potential physical relationship between sunspots and the magnetic field, we employ another machine learning technique called transfer learning which has recently received considerable attention in the literature. Here, this approach consists in first training the source spatial-temporal neural network on the much longer time/latitude sunspot area dataset, which starts in 1874, then transferring the trained set of layers to a target network, and continue training the latter on the magnetic field dataset. The employment of transfer learning in the field of computer vision is known to obtain a generalized set of feature filters that can be reused for other datasets and tasks. Here we obtain a similar result, whereby we first train the network on the spatial-temporal sunspot area data, then the first few layers of the neural network are able to identify the two main features of the solar cycle, i.e. the amplitude variation and the migration to the equator, and therefore can be used to train on the magnetic field dataset and forecast better than a prediction based only on the historical magnetic field data.

References in corpus (17)

• Solar Flare Prediction Using SDO/HMI Vector Magnetic Field Data with a Machine-Learning Algorithm
• Solar activity forecast with a dynamo model
• Predicting Coronal Mass Ejections Using Machine Learning Methods
• Transfer learning for music classification and regression tasks
• An Updated Solar Cycle 25 Prediction with AFT: The Modern Minimum
• The solar magnetic field since 1700: I. Characteristics of sunspot group emergence and reconstruction of the butterfly diagram
• Predictability of the solar cycle over one cycle
• Deciphering Solar Magnetic Activity I: On The Relationship Between The Sunspot Cycle And The Evolution Of Small Magnetic Features
• Solar cycle 25: another moderate cycle?
• Effects of Meridional Flow Variations on Solar Cycles 23 and 24
• A Comparison Between Global Proxies of the Sun's Magnetic Activity Cycle: Inferences from Helioseismology
• Using the Dipolar and Quadrupolar Moments to Improve Solar-Cycle Predictions Based on the Polar Magnetic Fields
• How many active regions are necessary to predict the solar dipole moment?
• Solar activity in the past and the chaotic behaviour of the dynamo
• Solar activity simulation and forecast with a flux-transport dynamo
• Spatial-temporal forecasting the sunspot diagram
• Optimal Neural Network Feature Selection for Spatial-Temporal Forecasting

Cited by in corpus (2)

• Predicting Solar Flares Using CNN and LSTM on Two Solar Cycles of Active Region Data
• Predicting the Evolution of Photospheric Magnetic Field in Solar Active Regions Using Deep Learning