Sign In

Communications of the ACM

ACM TechNews

Deep Learning at Scale for Construction of Galaxy Catalogs


View as: Print Mobile App Share: Send by email Share on reddit Share on StumbleUpon Share on Hacker News Share on Tweeter Share on Facebook
The teams innovative framework lays the foundations to exploit deep transfer learning at scale, data clustering and recursive training to produce large-scale galaxy catalogs.

A team of scientists is applying the power of artificial intelligence and high-performance supercomputers to accelerate efforts to analyze the increasingly massive datasets produced by cosmological surveys.

Credit: Inside HPC

Researchers at the University of Illinois at Urbana-Champaign's National Center for Supercomputing Applications (NCSA) and the U.S. Department of Energy's Argonne National Laboratory combined artificial intelligence with high-performance computing to expedite analysis of massive datasets generated by cosmological surveys.

The team integrated deep learning techniques to classify hundreds of millions of unlabeled galaxies with high accuracy.

The researchers utilized Sloan Digital Sky Survey (SDSS) datasets produced by the Galaxy Zoo initiative to train neural network models to classify galaxies in the Dark Energy Survey (DES), which intersect with the footprint of both surveys.

The technique was 99.6% accurate in identifying spiral and elliptical galaxies.

The team also developed open source software stacks to mine galactic images from the SDSS and DES surveys at scale, using the NCSA's Blue Waters supercomputer.

From Inside HPC
View Full Article

 

Abstracts Copyright © 2019 SmithBucklin, Washington, DC, USA


 

No entries found