SH036-0019
Italian Solar Orbiter-SWA Working Group on Machine Learning and Artificial Intelligence
Monday, 14 December 2020
Poster
Rossana De Marco1, Tommaso Alberti2, Jorge Amaya3, Roberto Bruno4, Francesco Califano5, Enrico Camporeale6, Giuseppe Consolini7, Raffaello Foldes8, Raffaella D'Amicis1, Romain Dupuis9, Luca Franci10, Luiz Fernando Guedes dos Santos11, Maria Elena Innocenti12, Vamsee Krishna Jagarlamudi13, Giovanni Lapenta14, Monica Laurenza15, Maria Federica Marcucci2, Ayris Narock16, Emanuele Papini17, Silvia Perri18, Denise Perrone19, Alessandro Retino20, Sergio Servidio21, Manuela Sisti22, Luca Sorriso-Valvo23 and Francesco Valentini24, (1)INAF-IAPS, Rome, Italy, (2)INAF - IAPS, Rome, Italy, (3)KU Leuven, Leuven, Belgium, (4)INAF-IFSI, Rome, Italy, (5)University of Pisa, Pisa, Italy, (6)University of Colorado, Cooperative Institute for Research in Environmental Sciences, Boulder, United States, (7)Ist. Nazionale di Astrofisica, Roma, Italy, (8)University of L'Aquila, L'Aquila, Italy, (9)Katholieke Universiteit Leuven, Leuven, Belgium, (10)Queen Mary University of London, School of Physics and Astronomy, London, United Kingdom, (11)Catholic University of America, Physics, Washington, DC, United States, (12)University of Leuven, Leuven, Belgium, (13)LPC2E, CNRS and University of Orléans, Orléans, France, (14)Katholieke Universiteit Leuven, Department of Mathematics, Leuven, Belgium, (15)INAF-Istituto di Astrofisica e Planetologia Spaziali, Roma, Italy, (16)NASA Goddard Space Flight Center, Greenbelt, MD, United States, (17)Università degli Studi di Firenze, Florence, Italy, (18)Dipartimento di Fisica, Fisica, Arcavacata di Rende, Italy, (19)ASI, Rome, Italy, (20)Laboratoire de Physique des Plasmas, Palaiseau, France, (21)Università della Calabria, Department of Physics, Rende, Italy, (22)INSIS, University of Aix-Marseille, Marseille, France, (23)ISTP-CNR, Bari, Bari, Italy, (24)University della Calabria, Rende, Italy
Abstract:
The exponential growth of data volume experienced by astronomy and astrophysics causes new disciplines like machine learning (ML) and data mining (DM) to gain more and more ground in these fields. Applications like clustering, feature selection, automatic classification of events are proving to be a valuable aid in exploiting space data in the era of the synergy between "pure" science and "data-driven" science.
The Italian Solar Orbiter-SWA Working Group on Machine Learning and Artificial Intelligence1 together with the European Commission Horizon 2020 project AIDA2, has the scope of applying ML and DM analysis techniques to the Solar Orbiter data. The implementations are numerous. First of all these new techniques can be used to discover unexpected relations between data, can automate tasks so that they can be carried out without human intervention, and can help to forecast physical properties and events. A non-exhaustive list of these activities includes automatic detection of coronal holes in images, automatic recognition of plasma regions, prediction of solar wind properties at 1 AU, classification of solar wind type based on new indicators, analysis of particle velocity distribution functions.
In addition, this Working Group will integrate the existing software developed in the context of the various heliospheric missions with the parts regarding Solar Orbiter. These packages are able to handle complex data set with ease and provide statistical analysis and visualization tools. Catalogs of scientific data are also produced, which report, among others, magnetic reconnection and particle acceleration events, detected by routines trained to browse data and select physical processes and features of interest.
Here we present the project overview along with the ML and DM tools which will be used to handle and analyse Solar Orbiter data.
1.https://sites.google.com/view/italian-solar-orbiter-swa/
2.http://www.aida-space.eu/
