Unveiling the Sun's Mysteries: The Promise of AI - Powered Surya

Artificial intelligence may soon unlock the Sun's most intricate secrets. On August 20, IBM and NASA jointly announced the debut of Surya, a foundational model tailored for solar studies. Trained on extensive datasets of solar activity, this AI - enabled tool is designed to enhance humanity's comprehension of solar weather and precisely forecast solar flares. These flares, bursts of electromagnetic radiation from the Sun, pose threats to orbiting astronauts and Earth - based communication infrastructure.

Data Source: NASA's Solar Dynamics Observatory

Surya was trained using nine years of data amassed by NASA's Solar Dynamics Observatory (SDO). Since 2010, SDO has been orbiting the Sun, capturing high - resolution images every 12 seconds. It observes the Sun at various electromagnetic wavelengths to estimate the temperature of its layers. Additionally, SDO takes precise measurements of the Sun's magnetic field, crucial data for understanding energy transfer within the star and predicting solar storms.

Challenges in Data Interpretation and the Digital Twin Solution

Historically, heliophysicists have grappled with interpreting the vast, diverse, and complex data from SDO. To overcome this hurdle, IBM reports that Surya's developers utilized SDO data to create a digital twin of the Sun. This is a dynamic virtual representation of the star that updates with new data and can be manipulated for more accessible study.

The process commenced with unifying the various data formats input into the model for consistent processing. Subsequently, a long - range vision transformer was employed. This AI architecture enables detailed analysis of high - resolution images and identification of relationships between image components, regardless of their spatial separation.

Optimizing Model Performance: Spectral Gating

The model's performance was optimized through a technique called spectral gating. By filtering out data noise, spectral gating reduces memory usage by up to 5 percent, thereby improving the quality of processed information.

Surya's Advantages: Swift and Accurate Predictions

Developers claim that Surya's design confers a significant edge. Unlike other algorithms that demand extensive data labeling, Surya can learn directly from raw data. This allows it to rapidly adapt to different tasks and produce reliable results in less time.

During testing, Surya demonstrated its ability to integrate data from other instruments, such as the Parker Solar Probe and the Solar and Heliospheric Observatory (SOHO), both of which also observe the Sun. It also proved effective in various predictive functions, including forecasting flare activity and solar wind speed.

IBM notes that traditional prediction models can only anticipate a flare one hour in advance based on signals from specific solar regions. In contrast, "Surya provided a two - hour lead by leveraging visual information. The model is believed to be the first to offer such a warning. In early model testing, the team achieved a 16 percent improvement in solar flare classification accuracy, a notable advancement over existing methods," as stated in an IBM statement.

Adaptability of Surya's Architecture

NASA emphasizes that while Surya was developed for heliophysics, its architecture is adaptable across different fields, from planetary science to Earth observation. "By creating a foundation model trained on NASA's heliophysics data, we're enabling the analysis of the Sun's complex behavior with unprecedented speed and precision," said Kevin Murphy, NASA's data science director. "This model fosters a broader understanding of how solar activity impacts critical systems and technologies on Earth."

The Significance of Predicting Solar Activity

The risks associated with abnormal solar activity are substantial. A major solar storm could disrupt global telecommunications, cause electrical grid failures, and interfere with GPS navigation, satellite operations, internet connections, and radio transmissions.

Andrés Muñoz - Jaramillo, a solar physicist at the Southwest Research Institute in San Antonio, Texas, and the project's lead scientist, stressed that Surya aims to maximize the lead time for such scenarios. "We strive to give Earth the longest possible lead time. Our hope is that the model has learned all the critical processes underlying the Sun's evolution over time, enabling us to extract actionable insights."

This article was originally published on WIRED en Español and has been translated from Spanish.