Specialists from the Moscow Institute of Physics and Technology (MIPT) have presented an innovative "Sentinel" system, designed to help astronomers analyze and classify stellar data using artificial intelligence. This was reported by the university's press service.
The "Sentinel" system, created at MIPT, automatically analyzes telescope data and classifies variable stars with an accuracy of up to 87%. In the future, it will help to find comets and potentially dangerous asteroids, which will make it an important tool for monitoring space threats. The project was implemented on domestic Yandex Cloud facilities.
Every night, telescopes around the world generate huge amounts of data, comparable to 30 thousand hours of HD video. Traditional methods of analysis cannot cope with such a flow of information. This leads to the loss of valuable scientific data.
In the development of "Sentinel", advanced ConvLSTM and WaveNet neural networks were used to analyze FITS files, photo images, and time series. The system can already determine the types of stars—from eclipsing to pulsating—and check with the world's largest catalogs, Gaia DR3 and VSX. This allows identifying new objects that may have been missed due to the limitations of individual telescopes.
With the help of "Sentinel", scientists can improve catalogs, quickly process data from telescopes, and find new objects among hundreds of gigabytes of information. The system is fully ready for operation and can be used to search for new stars, supernova flashes, and identify captured objects.
Earlier, www1.ru reported that the world's most powerful laser will appear in Russia.
Read materials on the topic:
Will Open a New Era of Armaments: Characteristics of the Peresvet Combat Laser Revealed
Russian Lasers Will Become Even Better and More Accurate – A Way to Modernize Them Has Been Found
Young Russian Scientist Presents New Technology for Creating Powerful Lasers