Communications of the ACM
Deep Learning at Scale for Construction of 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