IBM and NASA Release Groundbreaking Open-Source AI Model on Hugging Face to Predict Solar Weather and Help Protect Critical Technology
IBM and NASA have unveiled Surya, an advanced open-source foundation model designed to interpret high-resolution solar data and predict space weather. This tool addresses the growing vulnerability of modern infrastructure to solar storms, which can disrupt GPS, power grids, and telecommunications. By treating space weather similarly to terrestrial climate forecasting, Surya aims to mitigate economic and technological risks associated with solar activity, ensuring better protection for satellites, aviation, and critical systems essential to daily life. The model represents a significant leap in AI capabilities, trained on nine years of high-resolution heliophysics data from NASA. It overcomes traditional limitations in solar forecasting by utilizing a massive, curated dataset to achieve higher accuracy in predicting solar flares and wind speeds. Unlike previous methods that rely on partial views, Surya provides visual predictions of flare locations up to two hours in advance. This capability stems from custom architectural solutions that handle the immense scale of the imagery, offering unprecedented spatial resolution for scientific analysis. Surya’s release on Hugging Face highlights a crucial trend in open data: democratizing access to complex scientific tools. By making these advanced foundation models publicly available, IBM and NASA enable researchers worldwide to build specialized applications and contribute to broader scientific discovery. This openness fosters collaboration across industries and regions, allowing the global community to leverage NASA’s deep expertise combined with cutting-edge AI. Ultimately, this initiative transforms proprietary research into a shared resource, accelerating innovation in how humanity prepares for and responds to space weather events.
Source: newsroom.ibm.comPublished on 2025-09-26
Related news
- Tevogen Marks Major Milestone in Its AI Initiative to Enhance Efficacy of T Cell–Based Therapies with 100x PredicTcell™ Beta Data Expansion
- NanoEdge AI Studio v5, 1st AutoML tool with Synthetic Data Generation
- Nigeria Unveils N-ATLAS Language AI Model at UNGA80, Signals Bold Step in Global AI Inclusion
- La pobreza en la Argentina bajó 21,3% en el primer año de Milei, según el INDEC
- Canada, Australia, Belgium, pledge to support NEITI on open data policy