Joint Physics Dept.–Inst. for Space Weather Sci. Colloquium

April 16, Thursday (** SPECIAL DAY**)

Space Weather Forecasting with Machine Learning – Lessons Learned from Research to Operation

Prof. Yang Chen

University of Michigan-Ann Arbor

(Solar Physics, Host: Bo Shen of Industrial Engineering)

**Special Room: CKB 116

**Special Time: 1pm - 2pm with 12:45 pm teatime

**ZOOM Meeting ID for those who cannot attend in-person:

955 9399 6954

(APPROVAL by Prof Ahn REQUIRED for APPH/MTSE PhD Students to attend online)

*Password: check email or request from kenahn@njit.edu

Machine learning is rapidly changing the landscape of space weather forecasting, offering new ways to detect precursors of solar flares, coronal mass ejections, solar energetic particle events, and geomagnetic disturbances. In this talk, I will review lessons learned from developing machine-learning-based space weather models and translating them from research prototypes into operational settings, with examples and context from our work at spaceweatherlab.org. I will focus on key challenges that arise in practice, including data quality, class imbalance, uncertainty quantification, model interpretability, and the need to evaluate performance under operational constraints rather than only on retrospective test sets. I will also discuss how combining data-driven methods with physics-based insight can improve robustness and make forecasts more useful for real-world decision-making. Finally, I will highlight opportunities for future progress as the community moves toward more reliable, scalable, and operationally relevant space weather prediction systems.