“NASA wants to properties of stars to find out machine learning”

Astronomers from the NASA have machine learning used to set the properties of thousands of stars in the Milky way to study. The scientists hope to create new patterns that people will not be noticed.

The astronomers use self-learning computer systems in order certain basic properties of stars. This can be photos of the universe. Normally, data on the spectrum of a star is required for using the different wavelengths of light, the properties of the object to be established, a time-consuming task for astronomers. The machine learning systems of the NASA, however, take pleasure with collections, because they patterns be able to recognise from which the properties of the stars are traced.

According to the NASA create the algorithms of the machine learning system, it is possible to in a relatively short period of time billions of stars to analyze. For example, the size of a star are determined, as well as what metals are present. In addition to the speedups provides the software, also a cost savings.

The software algorithms were first ‘trained’ to be in the recognize of stars. This used the NASA a file in which 9000 stars were included. By the algorithms, as also to feed with data files in which the lichtcurven of the stars were included, could the software after a while, patterns to recognize. The NASA-astronomers have their findings with the system so far in a research report is logged.