NASA, in cooperation with IBM and Research Partners. SURYA – A new model for the artificial intelligence institution designed to analyze and predict sunshine. SURYA has been trained over nine years of continuous data from the NASA solar -energy observatory, how advanced AI can convert helminths and predict space weather.
SURYA applies the structure of the foundation form that learn directly from raw solar data instead of asking to put extensive marks. Its design combines the layers of spectral mass with the tall spinal, allowing it to discover both long -term solar cycles and short -lived phenomena such as solar spots and glow.
Initial criteria indicate that Surya improves the accuracy of the sun’s sun prediction by up to 16 percent compared to previous models. It can generate visual predictions two hours ago and predict the speed of solar wind up to four days. These capabilities provide valuable tools to expect satellite weather events that can affect satellites, energy networks, navigation systems and human space.
Syria has been validated in four basic research tasks:
- Prediction in the active area in the area-24-hour predators for new solar areas.
- Solar annoyance – estimate the possibility of great torches during the next day.
- Solar wind speed – prediction of solar wind conditions is important for satellites and network stability.
- Euv spectrums – sunflower maps to better understand their effect on Earth’s atmosphere.
The ingenuity of the model highlights its potential not only for avant -garde physics but also for other scientific fields. Architectural and systematic engineering can be adapted to the science of planets and the surveillance of the Earth and beyond.
SURYA was developed under the NASA advanced agency and concepts implementation team at the Marshall Space Flight Center as part of the Agency’s AI initiative for science. The project includes NASA centers, IBM Research, the Southern West Research Institute, the SETI Institute, and universities throughout the United States.
Development requires overcoming great challenges, including memory restrictions and the need to combine the frequency known with the modeling of time chains. By integrating experience from both artificial intelligence engineers and Hellenicologists, the team created a model capable of learning accurate spatial details while capturing long -term time dynamics.
Both SURYA model and training data groups are available on Embroidery and JaytabEnsuring open access to researchers, teachers and students all over the world. The training process has been supported by the NAIRR, an initiative led by the National Science Foundation, with the computer infrastructure provided by NVIDIA and other partners.
In addition to its role in the progress of hallucinogenic aircraft, Syria also supports sustainability by enhancing the flexibility of technologies that support modern society. The accurate weather expectations in space helps protect the renewable energy infrastructure, energy networks, and global communications systems from solar energy disorders, which reduces the risk of costly interruption and environmental impact from damaged equipment. In this way, the solar predictions of AI indirectly contribute to the stability of energy networks and the broader targets of green technology.
SURYA is part of NASA’s “5+1” strategy, which aims to develop the basis forms of both the main science fields in NASA along with a large language model to connect it. By including scientific experience directly in advanced AI systems, NASA creates discovery acceleration tools, improving prediction, and enhancing flexibility against risks posed by space weather.
