AI Forecasts Solar Activity Nine Hours Ahead
A team from the New Jersey Institute of Technology developed an artificial intelligence model called EarlyDetect, which can predict the appearance of active regions on the Sun approximately 9.24 hours in advance on average. The model relies on analyzing sound waves that pass through the Sun, capturing subtle changes in magnetic activity, which leave faint signatures in these waves. After training on data from NASA's observatory and addressing the challenge of removing weak signals, the model was able to improve its accuracy and outperform previous models. This advancement provides a greater opportunity to warn of solar flares before they occur, which is crucial for enhancing the readiness of communication and energy systems for space weather. Additionally, the team launched an interactive platform and open data to support future research.
Notice: This Is an AI-Generated Summary
Comments (0)