NASA’s Solar Dynamics Observatory gathers significant data about the Sun and in terms of the amount of data collected, it outperforms all other satellites sent by the agency. It in fact collects 1.5 terabytes of data per day and parsing this much data by the researchers is quite tough. To resolve the problem the scientists at Stanford University are making use of artificial intelligence to sort out the huge amount of information.
The process has been termed as “machine learning” in which data can be sort out by artificial intelligence into relevant categories. The sorting process becomes better as more and more data pours in. Researcher Monica Bobra stated “Machine learning is a sophisticated way to analyze a ton of data and classify it into different groups.” Bobra and her fellow researcher Sebastien Couvidat also analyzed if the propensity of the machine to find patterns can be used for predicting the strength of potential solar flares. This is crucial information as massive bursts of solar energy can damage communication equipments on Earth.
Active and non-active areas of the Sun’s surface were catalogued by the researchers with the help of a database of nearly 2000 active regions. The numbers were then fed to their machines. The process of identifying the relevant features of active as well as dormant regions was quickly learnt by the machines. These active regions threaten to send out solar flares.
The prediction of active regions would help the communication workers to prevent damage of the equipments in times when solar flares make way towards our planet. Researchers also observed that all 25 features of the machine which it uses to differentiate between active and non-active regions were not essential for accurate prediction.