For the first time, an artificial intelligence has identified and classified a supernova

Scientists say that an artificial intelligence tool has succeeded for the first time Fully automatic Observe, identify and classify a supernova. This mechanism can not only increase the speed of analysis and classification of supernovae, but also can prevent human error.

A group of international scientists led by Northwestern University in the United States has developed a new artificial intelligence system that can automatically perform the entire process of searching for supernovae in the night sky. The first success of this system has now been shared with the global astronomical community.

Identification of supernovae with artificial intelligence

Adam Miller, from Northwestern University, who was the leader of this research, says: “For the first time in history, a group of robots and artificial intelligence algorithms were able to observe and then identify the objects and talk to another telescope to discover a supernova. confirm “This achievement represents an important step forward, as further optimizations in these models will allow robots to detect specific subtypes of starbursts.”

In the last six years, researchers have spent a total of about 2,200 hours studying and classifying supernovae. But the new artificial intelligence that is now called BTSbot made available, can help scientists spend their valuable time on more important work.

To train this artificial intelligence, scientists trained a machine learning algorithm with more than 1.4 million images from 16 thousand different sources. Then, to train BTSbot, go to a supernova candidate known as SN2023tyk they went. After the discovery of SN2023tyk, the system requested the spectrum of this possible supernova from the Palomar Observatory. There, a bot responded to BTSbot and sent the relevant information.

Then it was determined that SN2023tyk is a type Ia supernova; A type of starburst in which a white dwarf in a binary star system completely explodes. The results of these findings were automatically shared with the astronomical community on October 7.