Ten Years of Machine Learning in Heliophysics

Session ID#: 281102

Session Description:
Since its inception, this session has been a focal point for machine learning (ML) in Heliophysics and Space Weather. Now marking its tenth edition, we reflect on progress from early exploration to scientifically grounded applications that enhance physical understanding and predictive capability.

We invite contributions spanning deep learning, generative AI, inverse problems, Bayesian methods, uncertainty quantification, and physics-informed ML. Priority will be given to studies that move beyond proof-of-concept and clearly demonstrate how ML advances domain science.

We particularly encourage submissions at the intersection of AI and scientific computing, including foundation and generative models in research workflows, agentic AI for autonomous or semi-autonomous discovery, and new integrations of data-driven and physics-based approaches. Contributions on reproducibility, interpretability, and the evolving role of AI are also welcome.

This session celebrates a decade of progress while defining the path forward for impactful, scientifically meaningful AI in Heliophysics and Space Weather.

Co-Sponsor(s):
  • NH - Natural Hazards
  • SA - SPA-Aeronomy
  • SH - SPA-Solar and Heliospheric Physics
  • SM - SPA-Magnetospheric Physics
Index Terms:

1942 Machine learning [INFORMATICS]
7899 General or miscellaneous [SPACE PLASMA PHYSICS]
7924 Forecasting [SPACE WEATHER]
7999 General or miscellaneous [SPACE WEATHER]
Primary Convener:  Enrico Camporeale, University of Colorado, Queen Mary University of London, Boulder, United States
Conveners:  Ryan McGranaghan, JPL/NASA/Caltech, Pasadena, United States, Jacob Bortnik, University of California Los Angeles, Atmospheric and Oceanic Sciences, Los Angeles, United States and Tomoko Matsuo, University of Colorado Boulder, Boulder, United States
See more of: Nonlinear Geophysics
<< Previous Session Proposal | Next Session Proposal >>