The exact mechanisms leading to chromospheric heating are still ill-defined. While the presence of magnetic elements is undoubtedly necessary, the details of the heating, and its spatio-temporal distribution remain poorly understood. We contribute to this topic by analyzing the behavior of hot chromospheric fibrils surrounding network and plage elements, identified via the broader Hα profiles observed along their length; the H-α spectral line width has been shown to correlate with the local chromospheric temperatures through comparison with the ALMA millimeter-continuum brightness temperature. We make use of loop tracing and analysis software to investigate characteristics of the chromospheric hot fibrils including their length, number density, and transverse spatial extension in an enhanced network region.

In this review, we introduce our recent applications of deep learning to solar and space weather data. We have successfully applied novel deep learning methods to the following applications: (1) generation of solar farside/backside magnetograms and global field extrapolation based on them, (2) generation of solar UV/EUV images from other UV/EUV images and magnetograms, (3) denoising solar magnetograms using supervised learning, (4) generation of UV/EUV images and magnetograms from Galileo sunspot drawings, (5) improvement of global IRI TEC maps using IGS TEC ones, (6) one-day forecasting of global TEC maps through image translation, (7) generation of high-resolution magnetograms from Ca II K images, (8) super-resolution of solar magnetograms, (9) flare classification by CNN and visual explanation by attribution methods, and (10) forecasting GOES solar X-ray profiles. We present major results and discuss them. We also present future plans for integrated space weather models based on deep learning.

The signatures of small-scale features in the solar atmosphere are severely degraded by limited angular resolution of the observations. The Deep Solar ALMA Neural Network Estimator (Deep-SANNE) is trained towards synthetic observables from 3D magnetohydrodynamic simulations to recognize the small-scale dynamic features in data at limited observational resolution, and provide maps of correction factors across the field of view. The correction factors can be used to acquire deconvolved refined images with significantly improved brightness temperature contrasts, where the strength of brightening events are reproduced to an accuracy of 94.0% instead of the 43.7% at observational resolution. Deep-SANNE can also provide masks of the most probable locations with large accuracies, and estimations on the radiation formation heights in connection to the small-scale features. The Deep-SANNE refined images and estimations of radiation formation heights allow for larger accuracy and meaningful analysis of solar ALMA data.