Solar Flare Detection

Developed a novel algorithm to predict the magnitude of solar flares based on NASA satellite imagery of the Sun

Developed an innovative algorithm that analyzes over 40,000 satellite images of the Sun from NASA's Solar Dynamics Observatory, achieving an impressive > 90% accuracy in classifying solar flares (B, C, or M class) based on the previous days satellite imagery. This project was demoed at Georgia Tech's inaugural Scientific Machine Learning Symposium. Scientific Machine Learning is a subset of ML that focuses on combining known physics principles with ML to achieve higher accuracy in Scientific models. For this project, this involved adapting a traditional CNN architecture to incorporate essential physics principles (like Wien's Law) that influence the occurrence of a solar flare.