Surya: New AI Model for Solar Physics Released

IBM and NASA have open-sourced 'Surya', a new AI model for solar physics. This model can predict powerful solar outbursts that pose threats to astronauts and affect satellites, power grids, and communications on Earth.

Charged particles from the Sun can impact power grids, GPS, and the internet. NASA's Solar Dynamics Observatory (SDO) has been observing the Sun for 15 years, and IBM and NASA have used this data to develop the Surya model.

Surya utilizes SDO data to predict solar eruptions. The model is available on Hugging Face, GitHub, and IBM's TerraTorch library. Additionally, SuryaBench, a dataset and benchmark, has been released.

Predicting solar storms is more challenging than predicting Earth's storms due to limited understanding of solar physics. Surya learns about the Sun's magnetic fields and rotation to forecast eruptions.

Surya is also used to predict the Sun's activity cycle. During the solar maximum, active regions appear on the Sun's surface, leading to solar flares and coronal mass ejections, which can disrupt Earth's GPS and communications.

Surya and SuryaBench offer a new way to study the Sun. The model is a tool for rapidly predicting and responding to solar flares.