IBM and NASA Develop a Digital Twin of the Sun to Predict Future Solar Storms

The most of the sun Thanks to the complex mysteries can be solved soon Artificial intellectThe August 20, IBM And Nasal The sun has announced to launch, a Foundation model For the sun. After being trained in large datasets of solar activity, this AI equipment deepens humanity’s understanding about solar weather and solar flames correctly predicts our stars that threatens and threatened both astronauts and both astronauts emitted by the stars Earth’s communication infrastructureThe

The sun was trained with nine-year data collected by the NASA solar dynamics observatory (SDO), which has been a material that has circulated the sun since the 21st, adopting high-resolution images every 12 seconds. SDOs capture the sun’s observations at various electronic magnetic wavelengths to estimate the temperature of the stars. It also takes a specific measure of the magnetic field of the sun – to understand how energy goes through the stars and to predict the solar storm.

ICally, this huge amount of diversified and complex data has become a challenge for Heliophysicists. To deal with this challenge, Says IBM Sun developers used SDO data to create a digital twin in the sun – a dynamic virtual replica of the stars that are updated when capturing new data and which is manipulated and more easily studied.

The process started by combining the various data forms feeding in the model, allowing them to process consistently. Subsequently, a long range of Vision Transformer was appointed-a II architecture that enables a detailed analysis of very high-resolution images and the identification of the relationship between their components, regardless of their distance.

The model’s performance was favorable using a process called spectral gating, which reduces the use of memory by filtering the data on the data, thereby increasing the quality of processed data.

More accurate predictions in less time

Its developers say that this design gives the sun a significant benefit: opposite other algorithms that require broad labeling of their feeding data, the sun can learn directly from raw data. It allows it to quickly adapt to different tasks and provide reliable results in less time.

During the test, the sun demonstrated its versatility in consolidating data from other instruments, such as Parker solar probe And solar and heliosphoric observatory (soho), two more spacecraft that observe the sun. The sun was also proven to be effective in various predictions, which include Flair activities and solar wind speed.

According to the IBM, the traditional pranted prediction models can predict a flame just one hour before the sun detected in the sun. In contrast, “the sun provided two hours of lead using visual data. The model was considered the first to provide such a warning. In the preliminary examination of the model, the team said that they had achieved a 16 percent improvement in the accuracy of the solar flame classification compared to the existing procedure,” the company said at one StatementThe

NASA emphasized that although the model was designed to study Heliophagics, its architecture is adapted to various fields from planetary science to earth observation. “By developing a foundation model trained in NASA Heliophysics data, we make it easier to analyze the complexities of the sun’s behavior with unprecedented speed and accuracy,” Kevin Murphy, director of NASA Data Science, says NASA’s Data Science director Kevin Murphy said. StatementThe “This model affects the critical systems and technologies that we all depend on the world here that we all give a wide understanding understanding.”

The risk raised by abnormal solar activity is not slight. A large solar storm can directly disturb global telecommunications, break electric grids and disturb the GPS, satellite operation, internet connection and radio infection.

The solar physicist of the South West Research Institute of San Antonio, Texas, and the top scientist of the project, Andreas Muse-Zamillo, emphasized that the goal of the sun is to maximize leadership time in this possible situation. “We want to give the world the longest leadership time. Our hope is that the model has learned all the critical process behind the evolution of our stars over time so that we can figure out functional insights.”

This story was originally attended Wired In Spanish And have been translated from Spanish.