Flare Transformer: Solar Flare Prediction Using Magnetograms and Sunspot Physical Features

Kaneda, Kanta; Wada, Yuiga; Iida, Tsumugi; Nishizuka, Naoto; Kubo, Yûki; Sugiura, Komei

doi:10.1007/978-3-031-26284-5_27

Cited by 3 publications

(7 citation statements)

References 23 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Regarding the computational demand, self-attention layers are faster than recurrent and convolutional layers, allowing parallelization (Vaswani et al, 2017). Kaneda et al (2023) recently introduced the Flare Transformer to forecast ≥M-class flares in the next 24 h. It is a hybrid system that consists of two combined transformer-based modules: the Sunspot Feature Module, which receives numerical time series as input, and the Magnetogram Module, which receives images and time-series data from magnetograms.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Solar Flare Forecasting Based on Magnetogram Sequences Learning with Multiscale Vision Transformers and Data Augmentation Techniques

Grim,

Gradvohl

2024

Sol Phys

View full text Add to dashboard Cite

Solar flares are releases of electromagnetic energy generally occurring in active solar regions with magnetic fields, known as sunspots. The burst of radiation released by a solar flare can reach Earth's atmosphere in a few minutes. Highintensity solar flares, i.e., M-or X-class flares, can significantly impact some of Earth's activities and technologies, such as satellites, telecommunications, and electrical power systems. Therefore, driving efforts in high-intensity solar flare forecasting systems is crucial. A forecasting model that observes the evolution of active regions may analyze a set of attributes that indicate which active regions can be precursors to solar flares. Recent work has focused on deep learning models that consider the evolution of active regions in the Sun. However, M-and X-class flares are spurious in the solar cycle period. That situation leads to an imbalanced dataset, increasing the effort to develop machine learning models for forecasting. Therefore, we proposed transformers-based models to forecast ≥Mclass flares, taking sequences of line-of-sight magnetogram images as input. In addition, we apply data augmentation techniques and other methods to deal with training on imbalanced datasets. Our fine-tuned models outperformed (TSS ≈ 0.80) state-of-the-art work using image processing to forecast ≥M-class flares in the next 48 h. Moreover, the data augmentation techniques applied to the training set kept the TSS stable and improved most of the secondary performance metrics analyzed.

show abstract

Section: Introductionmentioning

confidence: 99%

“…Our approach differs from most related works, which focus on training from scratch on models built explicitly for solar flare forecasting. Those models are derived from CNN-based models or, as in Kaneda et al (2023), based on a simplified transformer.…”

Section: Introductionmentioning

confidence: 99%

Solar Flare Forecasting Based on Magnetogram Sequences Learning with Multiscale Vision Transformers and Data Augmentation Techniques

Grim,

Gradvohl

2024

Sol Phys

View full text Add to dashboard Cite

show abstract

“…Solar flares also cause ionospheric heating which can lead to inaccurate information from satellites due to disrupted wave propagation [2]. Based on approximations, a major flare can result in a loss of US $163, 000, 000, 000.00 (≈ R3, 056, 908, 534, 521.05 in South African Rand (ZAR) [10]) in North America alone [11]. Global financial estimates make solar flares a significant area of attention.…”

Section: List Of Figuresmentioning

confidence: 99%

“…The occurrence of solar flares is mainly stochastic [6]. Most present data is based on ARs [5,7,11,14,17]. Existing solutions based on Recurrent Neural Networks (RNNs) that attempted to predict the recurrence of ≥C solar flares have shown poor calibration [7].…”

Section: Problem Statementmentioning

confidence: 99%

See 1 more Smart Citation

Solar flare recurrence prediction & visual recognition.

Mngomezulu

View full text Add to dashboard Cite

Solar flare forecasting based on magnetogram sequences learning with multiscale vision transformers and data augmentation techniques

Grim,

Gradvohl

2023

Preprint

View full text Add to dashboard Cite

Solar flares are releases of electromagnetic energy generally occurring in active solar regions with magnetic fields, known as sunspots. The burst of radiation released by a solar flare can reach Earth's atmosphere in a few minutes. High-intensity solar flares, i.e., M- or X-class flares, can significantly impact some of Earth’s activities and technologies, such as satellites, telecommunications, and electrical power systems. Therefore, driving efforts in high-intensity solar flare forecasting systems is crucial. A forecasting model that observes the evolution of active regions may analyze a set of attributes that indicate which active regions can be precursors to solar flares. Recent work has focused on deep learning models that consider the evolution of active regions in the Sun. However, M- and X-class flares are spurious in the solar cycle period. That situation leads to an imbalanced dataset, increasing the effort to develop machine learning models for forecasting. Therefore, we proposed transformers-based models to forecast >= M-class flares, taking sequences of line-of-sight magnetogram images as input. In addition, we apply data augmentation techniques and other methods to deal with training on imbalanced datasets. Our fine-tuned models outperformed (TSS ≈ 0.80) state-of-the-art work using image processing to forecast >= M-class flares in the next 48 h. Moreover, the data augmentation techniques applied to the training set kept the TSS stable and improved most of the secondary performance metrics analyzed.

show abstract

Flare Transformer: Solar Flare Prediction Using Magnetograms and Sunspot Physical Features

Cited by 3 publications

References 23 publications

Solar Flare Forecasting Based on Magnetogram Sequences Learning with Multiscale Vision Transformers and Data Augmentation Techniques

Solar Flare Forecasting Based on Magnetogram Sequences Learning with Multiscale Vision Transformers and Data Augmentation Techniques

Solar flare recurrence prediction & visual recognition.

Solar flare forecasting based on magnetogram sequences learning with multiscale vision transformers and data augmentation techniques

Contact Info

Product

Resources

About