IBM and NASA Launch Surya AI Model to Advance Solar Weather Forecasting

Details

• IBM and NASA have released Surya, an open-source AI model for predicting solar weather, now available on Hugging Face.
• Trained on nine years of high-resolution solar observation data from NASA's Solar Dynamics Observatory, Surya leverages self-supervised learning to improve accuracy.
• The model outperforms previous benchmarks, offering a 16% boost in solar flare classification accuracy.
• Surya can predict solar flares up to two hours in advance, providing visual forecasts for enhanced preparedness.
• This release supports NASA's broader 5+1 strategy for integrating AI in scientific research and addresses the rising risks associated with increased reliance on technology vulnerable to space weather.
• Economic assessments from Lloyd’s warn that severe solar storms could cause up to $2.4 trillion in losses over five years.

Impact

Surya's debut marks a major step in leveraging AI for space weather prediction at a time when technological infrastructure is increasingly susceptible to solar disturbances. Its open-source availability encourages broad research and innovation, setting a new standard for the field. With potential to influence global science and infrastructure resilience, this collaboration positions IBM and NASA at the forefront of specialized AI-driven weather forecasting.