14 July 2020

The European Southern Observatory (ESO) is the foremost intergovernmental astronomy organisation in Europe and the world’s most productive astronomical observatory. This European astronomy organisation provides state-of-the-art research facilities to astronomers and astrophysicists to carry out front-line science and to access the southern sky. ESO has built and operated some of the largest and most technologically-advanced telescopes (e.g. Very Large Telescope) and is now building the Extremely Large Telescope (begin of operations expected in 2025) which will become the world's largest optical reflecting telescope. It will allow to address a wide variety of scientific questions ranging from our own solar system to extra-solar planets, from nearby galaxies to the furthest observable objects at the edge of the visible Universe, from fundamental physics to cosmology.

ESO telescopes generate large amounts of data at a high rate which are stored in a permanent archive facility known as the ESO Science Archive Facility, containing about 2PB of data, all available to the public with an input rate of over 130 TB per year. Unsupervised analysis of this massive amount of data unleashes its potential for science discoveries. This is where Deep Learning, a technology enabling computers to understand images at a useful level and a reasonable cost by discovering patterns in the digital representation of data, can come into play and is at the heart of ESO’s collaboration with ESCAPE.

ESCAPE is deploying a Deep Learning Analysis on ESO Science Archive enhance the ability to decipher the content of images

Deep Learning Classification & analytics to provide advanced archive content

With ESCAPE’s support, ESO has access to a precious network of experts with whom ESO can establish collaboration for new ideas to thrive, as well as direct access to resources that allow ESO to better analyse its archive of astronomical data. This enhancement of data management will be made possible through the five services that ESCAPE is making available to ESO. ESCAPE Data Infrastructure for Open Science (DIOS) will provide potential solutions to streamline and optimize ESO’s data management and data access needs. The ESCAPE Virtual Observatory (VO) will help build a common data stewardship culture for ESO data, for an optimal data exploitation. ESCAPE Open-source scientific Software and Service Repository (OSSR) will support ESO in sharing high quality data based on FAIR data principles for dissemination and reuse, while the ESCAPE Science Analysis Platform (SAP) will support the combination ESO’s different data to boost science discovery.

Even with advanced Deep Learning techniques, there will always be data that may confuse the computer algorithms and, in this case, ESCAPE Citizen Science (CS) comes into play. A dedicated experiment may be created to validate and complement the results from the implemented Deep Learning analysis, thanks to the support from volunteers eager to contribute to science.

All in all, thanks to ESCAPE, ESO will access resources to perform Deep Learning Analysis to its Science Archive Facility and analyse their astronomical spectral thanks to a platform that will interactively the results.


Views 1,881