IBM and NASA have collaborated to develop Surya, an AI model trained on nine years of data from NASA’s Solar Dynamics Observatory, aiming to enhance our understanding of solar weather and predict solar flares accurately. Surya serves as a digital twin of the sun, constantly updated with new data to provide insights into solar activity and improve predictions of electromagnetic radiation bursts that can impact astronauts and Earth’s communications infrastructure. By leveraging AI technology, this innovation addresses the challenge of interpreting complex and diverse solar data, offering a dynamic virtual replica of the sun for detailed study.
The development of Surya involved unifying diverse data formats from the SDO, utilizing a long-range vision transformer for high-resolution image analysis, and implementing spectral gating to enhance data processing quality by filtering out noise. Unlike traditional algorithms requiring extensive data labeling, Surya can learn directly from raw data, enabling quick adaptation to different tasks and delivering reliable results efficiently. In testing, Surya showcased its versatility by integrating data from various spacecraft observing the sun, demonstrating effectiveness in predicting flare activity and solar wind speed with a two-hour lead time, surpassing existing prediction models.
NASA highlights the adaptability of Surya’s architecture beyond heliophysics, emphasizing its potential applications in planetary science and Earth observation. The model’s capacity to analyze solar behavior with speed and precision opens avenues for understanding how solar activity impacts critical systems on Earth, such as telecommunications, electrical grids, GPS navigation, and satellite operations. With the capability to offer early warnings of abnormal solar activity, Surya aims to provide crucial lead time for potential scenarios, maximizing preparedness and enabling actionable insights to mitigate the impact of solar storms on Earth’s technological infrastructure.
Andrés Muñoz-Jaramillo, lead scientist on the Surya project, underscores the significance of extending lead time for Earth in the face of solar threats, aiming to leverage the model’s comprehensive understanding of solar processes to extract actionable insights for safeguarding critical systems. The collaboration between IBM and NASA in developing Surya represents a milestone in leveraging AI and digital twin technology to advance solar research, enhance predictive capabilities, and mitigate risks associated with solar storms. By empowering a deeper understanding of solar dynamics and their implications for Earth, Surya signifies a significant step towards harnessing technology for proactive space weather management.
Key Takeaways:
– Surya, IBM and NASA’s AI model, serves as a digital twin of the sun, enabling accurate prediction of solar flares with a two-hour lead time.
– The model’s adaptability extends beyond heliophysics, offering applications in planetary science and Earth observation, enhancing understanding of solar impact on critical systems.
– With the potential to provide early warnings of abnormal solar activity, Surya aims to maximize lead time for Earth, extracting actionable insights to mitigate risks associated with solar storms.
– The collaboration between IBM and NASA in developing Surya represents a significant advancement in leveraging AI and digital twin technology for proactive space weather management.
Read more on wired.com