Title and Abstract

Title:  Machine Learning for Space Weather Analytics

Abstract:  In this talk we introduce an emerging interdisciplinary field, space weather analytics, which aims to (i) understand the onset of solar eruptions and assess space weather effects on Earth through big solar and space data analysis, (ii) perform near real-time long-term predictions of extreme space weather events including solar flares, coronal mass ejections and solar energetic particles, and (iii) understand solar wind turbulence and geomagnetic storms, by using advanced artificial intelligence (AI) and machine learning (ML) techniques. We present a suite of end-to-end deep learning AI models and tools for performing space weather analytics. These AI/ML tools have been used to (i) predict solar flares and coronal mass ejections, (ii) trace important structures such as magnetic flux elements and Halpha fibrils in solar active regions, and (iii) perform Stokes inversion. Finally, we describe our efforts of incorporating some of these AI/ML tools into a community-coordinated cyberinfrastructure (CI) platform for space weather science and point out some directions of future research for expanding our CI facility to serve the space weather community.