Japan Geoscience Union Meeting 2021

Presentation information

Print

[J] Poster

M (Multidisciplinary and Interdisciplinary) » M-GI General Geosciences, Information Geosciences & Simulations

[M-GI34] Earth and planetary informatics with huge data management

Thu. Jun 3, 2021 5:15 PM - 6:30 PM Ch.18

convener:Ken T. Murata(National Institute of Information and Communications Technology), Susumu Nonogaki(Geological Survey of Japan, National Institute of Advanced Industrial Science and Technology), Rie Honda(Department of Science and Technology, System of Natual Science, Kochi University), Keiichiro Fukazawa(Academic Center for Computing and Media Studies, Kyoto University)

5:15 PM - 6:30 PM

[MGI34-P07] Activity of preparing training data for study of space plasma with collaboration of observation, MHD simulation and machine leaning

*Keiichiro Fukazawa1, Tomoki Kimura2, Terumasa Tokunaga3, Shin ya Nakano4 (1.Academic Center for Computing and Media Studies, Kyoto University, 2.Frontier Research Institute for Interdisciplinary Sciences, Tohoku University, 3.Department of Systems Design and Informatics, Faculty of Computer Science and Systems Engineering, Kyushu Institute of Technology, 4.Department of Statistical Modeling, Institute of Statistical Mathematics)

Keywords:machine learning, space plasma, database

The machine learning (ML) has become a powerful tool to find the relation between variables thanks to the deep learning technique. This performs greatly in the classification, regression and recently generative modeling in the engineering and commercial areas. However, due to the satisfaction of physical laws in the scientific research area, the application of ML has some difficulties. In particular, the generative modeling is very sensitive to scientific data since the generated data is not guaranteed by the physical laws.

To overcome these problems, we have tried to apply ML to space plasma physics for several years and then reconfirmed the importance of preparing the training data. For example, when we classify the aurora in the observation results with ML, we have to prepare the huge amount of aurora observational image to learn the feature with high accuracy. It takes a lot of time to prepare the aurora image if considering the augmentation. Thus, we are developing the web site to classify the image including the aurora or not with the expectation of citizen science cooperation. In this study we show the web site for preparing the classification image and the initial results of it. In addition, our database of global simulation data of magnetosphere using real solar wind data and formatted aurora image by All-Sky-Imager will be shown.