Abstract In this study, we present the application of Convolutional Neural Network (CNN) to forecast solar flare occurrence. For this, consider three CNN models (two pretrained models, AlexNet and GoogLeNet, one newly proposed model). Our inputs are SOHO /Michelson Doppler Imager (from 1996 May 2010 December) SDO /Helioseismic Magnetic 2011 January 2017 June) full-disk magnetograms at 00:00 UT. Model outputs “Yes or No” daily occurrence (C, M, X classes) they compared with GOES observations....

In this study, we address the question of whether can properly determine differential emission measures (DEMs) using Solar Orbiter/Extreme Ultraviolet Imager (EUI)/Full Sun (FSI) and AI-generated extreme UV (EUV) data. The FSI observes only two full-disk EUV channels (174 304 Å), which is insufficient for accurately determining DEMs lead to significant uncertainties. To solve problem, trained tested deep learning models based on Pix2PixCC Dynamics Observatory (SDO)/Atmospheric Imaging...

In this study, we determine the differential emission measures (DEMs) using Solar Orbiter/Extreme Ultraviolet Imager (EUI)/Full Sun (FSI) and AI-generated EUV data. The FSI observes only two full-disk extreme UV (EUV) channels (174 304 Å), which poses a limitation for accurately determining DEMs. We address problem deep learning models based on Pix2PixCC, trained Dynamics Observatory (SDO)/Atmospheric Imaging Assembly (AIA) dataset. These successfully generate five-channel (94, 131,...

Abstract In this Letter, we apply deep-learning methods to the image-to-image translation from solar magnetograms ultraviolet (UV) and extreme UV (EUV) images. For this, We consider two convolutional neural network models with different loss functions, one (Model A) is L1 ( L 1 ), other B) cGAN ). train using pairs of Solar Dynamics Observatory SDO )/Atmospheric Imaging Assembly (AIA) nine-passband (94, 131, 171, 193, 211, 304, 335, 1600, 1700 Å) UV/EUV images their corresponding...

Abstract In this study, we present a visual explanation of deep learning solar flare forecast model and its relationship to physical parameters active regions (ARs). For this, use full-disk magnetograms at 00:00 UT from the Solar Heliospheric Observatory/Michelson Doppler Imager Dynamics Observatory/Helioseismic Magnetic Imager, Space-weather HMI Active Region Patch (SHARP), Geostationary Operational Environmental Satellite X-ray data. Our based on Convolutional Neural Network (CNN) predicts...

Background: Nontargeted analysis (NTA) methods identify novel exposures; however, few chemicals have been quantified and interrogated with pregnancy complications. Objectives: We characterized levels of nine exogenous endogenous in maternal cord blood identified, selected, confirmed prior NTA steps, including linear branched isomers perfluorooctane sulfonate (PFOS), perfluorohexane (PFHxS), monoethylhexyl phthalate, 4-nitrophenol, tetraethylene glycol, tridecanedioic acid, octadecanedioic...

Abstract Solar magnetic fields play a key role in understanding the nature of coronal phenomena. Global are usually extrapolated from photospheric fields, for which farside data is taken when it was at frontside, about two weeks earlier. For first time we have constructed extrapolations global using frontside and artificial intelligence (AI)-generated near-real basis. We generate magnetograms three channel observations Terrestrial Relations Observatory (STEREO) Ahead (A) Behind (B) by our...

ABSTRACT We investigate 306 LASCO front-side halo (partial and full) CMEs from 2009 to 2013, which are well-observed by both the Solar Heliospheric Observatory ( SOHO ) TErrestrial RElations STEREO ). These have two-dimensional (2D) parameters, such as speed, angular width, propagation direction, a single spacecraft ), well three-dimensional (3D) parameters multi-spacecraft 2D CME based on plane-of-sky observations, taken catalog NGDC flare catalog. determined their 3D using Stereoscopic...

Abstract In this Letter, we generate realistic high-resolution (1024 × 1024 pixels) pseudo-magnetograms from Ca ii K images using a deep learning model based on conditional generative adversarial networks. For this, consider “pix2pixHD” that is specifically devised for image translation tasks. We use 393.3 nm the Precision Solar Photometric Telescope at Rome Observatory and line-of-sight magnetograms Helioseismic Magnetic Imager (HMI) Dynamics 2011 January to 2015 June. 2465 pairs of HMI are...

Abstract Background Laparoscopic total gastrectomy for gastric cancer is feasible but less commonly performed compared to laparoscopic distal due technical difficulties such as reconstruction. There no standard esophagojejunal anastomosis technique in a lack of evidence. Methods We retrospectively analyzed data from 213 patients with who underwent October 2012 December 2016. Of these, 109 and 104 esophagojejunostomy linear circular stapling, respectively. short-term postoperative outcomes,...

Here we greatly improve Artificial Intelligence (AI)-generated solar farside magnetograms using data sets of Solar Terrestrial Relations Observatory (STEREO) and Dynamics (SDO). We modify our previous deep learning model configuration input to generate more realistic than before. First, model, which is called Pix2PixCC, uses updated objective functions include correlation coefficients (CCs) between the real generated data. Second, construct model: STEREO extreme ultraviolet (EUV)...

Measurement of the solar wind speed near Sun is important for understanding acceleration mechanism wind. In this Letter, we determine 2D speeds from 6 to 26 radii by applying Fourier motion filters SOHO/LASCO C3 movies observed 1999 2010. Our method successfully reproduces original flow in artificially generated data as well streamer blobs. We measure one-day one-year timescales and their variation cycle 24. find that at longer than a month maximum period are relatively uniform azimuthal...

Abstract In astronomy, long-exposure observations are one of the important ways to improve signal-to-noise ratios (S/Ns). this Letter, we apply a deep-learning model de-noise solar magnetograms. This is based on deep convolutional generative adversarial network with conditional loss for image-to-image translation from single magnetogram (input) stacked (target). For input magnetogram, use Solar Dynamics Observatory ( SDO )/Helioseismic and Magnetic Imager (HMI) line-of-sight magnetograms at...

Sleep staging is an important task that enables sleep quality assessment and disorder diagnosis. Due to dependency on manually labeled data, many researches have turned from supervised approaches self-supervised learning (SSL) for staging. While existing SSL methods made significant progress in terms of its comparable performance methods, there are still some limitations. Contrastive could potentially lead false negative pair assignments signal data. Moreover, data augmentation techniques...

Abstract To prepare for when only single‐view observations are available, we have made a test whether the 3‐D parameters (radial velocity, angular width, and source location) of halo coronal mass ejections (HCMEs) from consistent with those multiview observations. For this test, select 44 HCMEs December 2010 to June 2011 following conditions: partial full by SOHO limb CMEs twin STEREO spacecraft they were approximately in quadrature. In study, compare three different methods: (1) geometrical...

Abstract In this Letter, we examine the relationship between 3D magnetic free energy (MFE) and flaring activity using 83 major solar flares (M-class X-class) in nine active regions (ARs). For this, use 998 nonlinear force-free field extrapolations compiled by “Institute for Space-Earth Environmental Research Database” at Nagoya University. These ARs produced least three with distinct rising peak phases of MFE. each phase, flare occurrence rate (FOR) is calculated as a ratio number to...

We present the first observational detection of radial and azimuthal oscillations in full halo coronal mass ejections (HCMEs). analyze nine HCMEs well-observed by Large Angle Spectrometric Coronagraph (LASCO) from 2011 February to June. Using LASCO C3 running difference images, we estimated instantaneous apparent speeds different directions solar disk center. find that development all these is accompanied quasi-periodic variations velocity with periods ranging 24 48 minutes. The amplitudes...

Abstract It is essential to determine three-dimensional parameters (e.g., radial speed, angular width, and source location) of coronal mass ejections (CMEs) for the space weather forecast. In this study, we investigate which cone type represents a halo CME morphology using 29 CMEs (12 Solar Heliospheric Observatory (SOHO) /Large Angle Spectrometric Coronagraph (LASCO) 17 Terrestrial Relations ( STEREO )/Sun–Earth Connection Coronal Investigation COR2 CMEs) from 2010 December 2011 June. These...