Artificial intelligence independently detected a supernova star

Northwestern University

A team of astronomers from Northwestern University (USA) has created the first fully automated AI-based system to search for and identify supernovae stars.

Here's What We Know

Typically, the process of detecting supernovae required a combination of automated instruments and manual inspection by humans. Robotic telescopes would find candidates, then astronomers would collect their spectra for analysis.

The BTSbot system eliminates humans from this process. It was trained on 1.4 million images from 16,000 astronomical sources. After training, the AI independently identified the new supernova SN2023tyk and requested its spectrum for analysis from the robotic telescope at the Palomar Observatory in California.

The developers believe that saving time on manual searches will allow scientists to analyse more data and develop new hypotheses about the origin of cosmic explosions.

Source: Northwestern University