제목:  What kinds of problems in space weather and weather are well solved by deep learning?​​​

​

​​​​​​​​​​​​​​​​​​연사: 문용재 교수 (경희대학교)​​​​​​​​​​​​

​

일시: 2026년 3월 17일 화요일 16:00

장소: 과학관 B102호​​

Abstract:

​In this talk, I introduce our recent deep learning applications to space weather and weather data. We have successfully applied novel deep learning methods to the following applications: (1) image translations between solar images such as solar farside magnetograms, (2) improvement of empirical models, (3) inversion problem, (4) super resolution, (5) pixel-to-pixel translation, (6) reduction of simulation times, (7) 2D to 3D conversion, and (8) space weather forecasting. We also discuss the applications of symbolic regression to find data-driven empirical relationships and GPT for space weather.