Big Universe, Big Data Machine Learning, and Image Analysis for Astronomy

Abstract

Today astronomy produces an enormous amount of data from the new generation of telescopes that come fitted with digital cameras, which can image large fractions of the sky much faster than ever. In some instances, the volume of data gathered in one night is equivalent to that obtained in an entire survey from ten years ago. In fact, to tackle this enormous influx of information, astronomers have to resort to advanced techniques from data science, particularly from machine learning and image analysis. These techniques, therefore, have a number of advantages besides accelerating data processing. First, they give room for task automation, such as object detection and classification, which saves time and human effort otherwise wasted in manual analysis. Second, predictive modelling will help the astronomer to predict celestial events or find trends that would possibly not be easily evident from more classical methods. However, working with such huge datasets is not all that has to be faced.