Operational solar flare forecasting via video-based deep learning

Solar flare prediction is an important task in the context of space weather research, as it has to address open problems in both solar physics and operational forecasting (Schwenn, 2006; McAteer et al., 2010). Although it is well-established that solar flares are a consequence of reconnection and reconfiguration of magnetic field lines high in the solar corona (Shibata, 1996; Sui et al., 2004; Su et al., 2013), yet there is still no agreement about the physical model that better explains the sudden magnetic energy release and the resulting acceleration mechanisms (Aschwanden, 2008; Shibata, 1996; Sui et al., 2004; Su et al., 2013). Further, solar flares are the main trigger of other space weather phenomena, and it is a challenging forecasting issue to predict the chain of the events that from solar flares lead to possible significant impacts on both in-orbit and on-Earth assets (Crown, 2012; Murray et al., 2017). Flare forecasting rely on both statistical (Song et al., 2009; Mason and Hoeksema, 2010; Bloomfield et al., 2012; Barnes et al., 2016) and deterministic (Strugarek and Charbonneau, 2014; Petrakou, 2018) methods.