Scientists trained computers to go through massive amounts of data gathered by telescopes and to detect supernovae events.
Machine learning is an application of artificial intelligence. In this process, a computer learns without direct supervision using mathematical models. Computer systems can learn on their own, based on their own experiences. And this is exactly what scientists did to learn more about the night sky. By training their computers, astronomers at Caltech could classify as many as one thousand supernovae. This was done completely autonomously. They used data gathered by the Zwick Transient Facility (aka ZTF), a sky survey instrument on the Palomar Observatory from Caltech. The Zwick Transient facility peers at the night sky every night and patiently looks for changes in deep space. It looks for changes called transient events. This, say, astronomers, includes everything from asteroids streaking across the night sky to Black Holes devouring defenseless stars. But it can also spot when stars explode in so-called supernovae events.
A handy tool
ZTF comes in very handy for astronomers. This is because every time it detects an event, it alerts astronomers worldwide. Then, astronomers use telescopes to follow up on the detected event and investigate its nature. Now, thanks to ZTF data and machine learning, a thousand supernovae have been spotted in the sky. Also, since the instrument scans the night sky every night, there is a lot of data to work with. Astronomers can only work for so much, so it sometimes gets overwhelming to go through all the data. To make things easier, they developed novel machine-learning algorithms that make life much easier for astronomers. The algorithm they developed is called SNIascore, and its task is to classify candidate supernovae.
SNIascore can do more
Supernovae come in two distinct classes. Type One, and Type two. In Type I supernovae, hydrogen is absent, while in Type II supernovae, it is abundant. Type I supernovae are produced when massive stars steal matter from their stellar neighbors, resulting in a thermonuclear explosion. Whenever a massive star collapses under its own gravity, it produces a Type II supernovae. SNIascore currently has the ability to classify so-called la supernova. These dying stars go out with a thermonuclear explosion of consistent strength. This allows scientists to measure the expansion rate of the universe. The best part is that SNIascore is further improving.