Primary Convener:
Enrico Camporeale, University of Colorado, Queen Mary University of London

Convener:
Jacob Bortnik, University of California Los Angeles
Ryan McGranaghan, NASA Jet Propulsion Laboratory
Tomoko Matsuo, University of Colorado Boulder

Chair:
Enrico Camporeale, University of Colorado, Queen Mary University of London
Jacob Bortnik, University of California Los Angeles

Since 2017, this session has brought together researchers applying machine learning (ML) across all domains of Heliophysics and Space Weather. We invite contributions using a range of techniques—including classification, regression, dimensionality reduction, inverse problems, deep learning, Bayesian methods, uncertainty quantification, and physics-informed ML—with a strong emphasis on scientific advancement. We are not seeking proof-of-concept studies; instead, we will prioritize work that clearly demonstrates how ML enables or enhances physical understanding, drives discovery, or improves the scientific value of observations and models. We also encourage discussion of emerging themes such as reproducibility, interpretability, and the use of generative AI—including Large Language Models and other foundation models—to support research workflows, foster hypothesis generation, or augment scientific analysis. This session aims to showcase the cutting edge of science-driven ML in Heliophysics, bringing together method developers and domain experts committed to rigorous, impactful, and interpretable applications of AI.

Index Terms
1942 Machine learning
7924 Forecasting
7959 Models

Suggested Itineraries:
Disasters‚ Calamities and Extreme Events
Space Weather
Machine Learning and AI

Cross-Listed:
SH - SPA-Solar and Heliospheric Physics
IN - Informatics
SA - SPA-Aeronomy
SM - SPA-Magnetospheric Physics

Co-Sponsored Sessions:
AMS： American Meteorological Society
EGU： European Geosciences Union

Co-Organized Sessions:
Informatics

Neighborhoods:
1. Science Nexus