Thanks to the National Science Foundation (NSF), particularly the AGS (Atmospheric and Geospace Sciences) and OAC (Office of Advanced Cyberinfrastructure) directorates, for funding my research projects.
This collaborative NSF project will generate realistic synthetic data of Solar Energetic Particle (SEP) events by integrating multimodal satellite observations and developing advanced generative AI frameworks, addressing the critical scarcity of training data in space weather forecasting. The resulting datasets and predictive tools will improve SEP prediction, enhance space infrastructure resilience, and foster broad educational and community impacts through student training, curriculum integration, and public outreach.
This project integrates heliophysics and computer science by using NASA and NOAA solar observations with graph neural networks to improve solar flare prediction from multivariate magnetic field time series data. It will advance flare forecasting methods, train graduate and undergraduate students, and broaden participation in data science and space weather research through mentorship and distance learning programs.
This project developed a public, web-based cyberinfrastructure with a GUI that enabled machine learning on both multivariate time series (MVTS) and functional network representations, supporting predictive, exploratory, and generative tasks across domains such as solar physics and neuroscience. By integrating advanced ML models and synthetic data generation, the platform improved applications like solar flare forecasting and neurological disease prediction, while fostering interdisciplinary research, curriculum development, and outreach to underrepresented students.