IBM and NASA create first-of-its-kind AI that can accurately predict violent solar flares

IBM and NASA have developed Surya, a groundbreaking open-source AI model capable of accurately predicting solar flares and coronal mass ejections. By analyzing vast amounts of historical satellite imagery from the Solar Dynamics Observatory, the model identifies complex patterns in the sun’s magnetic fields and atmospheric conditions. This capability allows for forecasts of dangerous solar activity up to two hours in advance, offering a significant improvement over previous prediction methods and providing crucial lead time for protecting astronauts and critical infrastructure. The relevance of this development to open data lies in the democratization of specialized scientific resources. The researchers have released the model’s architecture and a curated benchmark dataset, SuryaBench, to the public domain via platforms like GitHub and Hugging Face. This transparency enables the global scientific community to access, test, and build upon these sophisticated tools without proprietary barriers. Consequently, it fosters collaborative research and accelerates innovation in space weather monitoring by allowing diverse experts to refine solar prediction algorithms using standardized, publicly available data. This initiative demonstrates the transformative power of open-source artificial intelligence in handling large-scale, complex scientific datasets. By making high-quality data and processing models freely accessible, the project lowers the threshold for advanced astrophysical research. It encourages wider participation in solving critical technological challenges, ensuring that the benefits of improved space weather forecasting extend beyond a single organization to benefit society at large, particularly in sectors vulnerable to solar disruptions.

Source: livescience.com
Published on 2025-08-22

IBM and NASA create first-of-its-kind AI that can accurately predict violent solar flares

Related news