Determining properties of solar active regions using machine learning

Karen A. Meyer; Nikolay Panayotov; Salma Hamdy Elsayed; Kean Lee Kang; Duncan Mackay

Determining properties of solar active regions using machine learning

Karen A. Meyer^*, Nikolay Panayotov, Salma Hamdy Elsayed, Kean Lee Kang, Duncan Mackay

^*Corresponding author for this work

Division of Cybersecurity

Research output: Contribution to conference › Abstract

Abstract

Advances in Machine Learning techniques mean that we now have the ability to analyse and derive statistics from large datasets of complex features. This project aims to train an artificial intelligence (including a deep neural network) to extract a consistent dataset of desired properties of solar Active Regions (ARs) from three decades of observed synoptic magnetograms. The methodology is initially developed using simulation data from a well-established flux transport model. This will then be adapted for application to observed synoptic maps, a subset of which have the desired AR properties catalogued, for verification of the technique.

Original language	English
Pages	50-51
Number of pages	1
Publication status	Published - 16 Sept 2019
Event	Machine Learning in Heliophysics - The Royal Tropical Institute - Koninklijk Instituut voor de Tropen (KIT), Amsterdam, Netherlands Duration: 16 Sept 2019 → 20 Sept 2019 https://ml-helio.github.io/

Conference

Conference	Machine Learning in Heliophysics
Abbreviated title	ML-Helio 2019
Country/Territory	Netherlands
City	Amsterdam
Period	16/09/19 → 20/09/19
Internet address	https://ml-helio.github.io/

Access to Document

http://web.archive.org/web/20201021165257/https://ml-helio.github.io/abstracts_book.pdf

Cite this

@conference{ffcb8c55ccb2496c94a54934cdf77e32,

title = "Determining properties of solar active regions using machine learning",

abstract = "Advances in Machine Learning techniques mean that we now have the ability to analyse and derive statistics from large datasets of complex features. This project aims to train an artificial intelligence (including a deep neural network) to extract a consistent dataset of desired properties of solar Active Regions (ARs) from three decades of observed synoptic magnetograms. The methodology is initially developed using simulation data from a well-established flux transport model. This will then be adapted for application to observed synoptic maps, a subset of which have the desired AR properties catalogued, for verification of the technique.",

author = "Meyer, {Karen A.} and Nikolay Panayotov and Elsayed, {Salma Hamdy} and Kang, {Kean Lee} and Duncan Mackay",

year = "2019",

month = sep,

day = "16",

language = "English",

pages = "50--51",

note = "Machine Learning in Heliophysics, ML-Helio 2019 ; Conference date: 16-09-2019 Through 20-09-2019",

url = "https://ml-helio.github.io/",

}

TY - CONF

T1 - Determining properties of solar active regions using machine learning

AU - Meyer, Karen A.

AU - Panayotov, Nikolay

AU - Elsayed, Salma Hamdy

AU - Kang, Kean Lee

AU - Mackay, Duncan

PY - 2019/9/16

Y1 - 2019/9/16

N2 - Advances in Machine Learning techniques mean that we now have the ability to analyse and derive statistics from large datasets of complex features. This project aims to train an artificial intelligence (including a deep neural network) to extract a consistent dataset of desired properties of solar Active Regions (ARs) from three decades of observed synoptic magnetograms. The methodology is initially developed using simulation data from a well-established flux transport model. This will then be adapted for application to observed synoptic maps, a subset of which have the desired AR properties catalogued, for verification of the technique.

AB - Advances in Machine Learning techniques mean that we now have the ability to analyse and derive statistics from large datasets of complex features. This project aims to train an artificial intelligence (including a deep neural network) to extract a consistent dataset of desired properties of solar Active Regions (ARs) from three decades of observed synoptic magnetograms. The methodology is initially developed using simulation data from a well-established flux transport model. This will then be adapted for application to observed synoptic maps, a subset of which have the desired AR properties catalogued, for verification of the technique.

M3 - Abstract

SP - 50

EP - 51

T2 - Machine Learning in Heliophysics

Y2 - 16 September 2019 through 20 September 2019

ER -

Determining properties of solar active regions using machine learning

Abstract

Conference

Access to Document

Fingerprint

Cite this