A system has been created to search for anomalies in billions of astronomical observations

Given the ever-increasing size of astronomical data sets, even if our telescopes

discover unexpected interestingastronomical phenomena, it is very unlikely that scientists will be able to recognize them among millions or even billions of observations. Astronomers have solved the problem by creating an automated tool specifically designed to recognize unusual behavior hidden among billions of measurements. Some of these tools already exist and are used, for example, to detect credit card fraud among millions of transactions every day. However, adapting them to scientific data is not easy due to the complexities associated with the nature of observations in astronomy. The SNAD team has been working for 3 years on the development and adaptation of such solutions in the context of astronomy.

During their last annual meeting,The team focused its efforts on objects whose brightness changes over time. Their system combines the strengths of machine learning algorithms and the indispensable knowledge of human experts to create a robust anomaly detection tool across billions of astronomical observations. 

The group has also developed a specially designeda web interface to instantly visualize and compare each candidate with existing astronomical catalogs. This was done in order to facilitate the work of experts who need to compare the candidates for anomalies with any other publicly available information about the studied coordinates of the sky.

Quickly and easily separating artifacts from interesting anomaly candidates is critical for current and upcoming next-generation observatories.

Read more

Traces of rocket fuel were found on Saturn's moon Rhea. Where does it come from?

The largest iceberg in the world collapsed, fragments rushed north. Is it dangerous?

Abortion and science: what will happen to the children who will give birth