skip to main content
OSTI.GOV title logo U.S. Department of Energy
Office of Scientific and Technical Information

Title: Optimising Automatic Morphological Classification of Galaxies with Machine Learning and Deep Learning using Dark Energy Survey Imaging

Abstract

There are several supervised machine learning methods used for the application of automated morphological classification of galaxies; however, there has not yet been a clear comparison of these different methods using imaging data, or an investigation for maximizing their effectiveness. We carry out a comparison between several common machine learning methods for galaxy classification [Convolutional Neural Network (CNN), K-nearest neighbour, logistic regression, Support Vector Machine, Random Forest, and Neural Networks] by using Dark Energy Survey (DES) data combined with visual classifications from the Galaxy Zoo 1 project (GZ1). Our goal is to determine the optimal machine learning methods when using imaging data for galaxy classification. We show that CNN is the most successful method of these ten methods in our study. Using a sample of ~2800 galaxies with visual classification from GZ1, we reach an accuracy of ~0.99 for the morphological classification of ellipticals and spirals. The further investigation of the galaxies that have a different ML and visual classification but with high predicted probabilities in our CNN usually reveals the incorrect classification provided by GZ1. We further find the galaxies having a low probability of being either spirals or ellipticals are visually lenticulars (S0), demonstrating that supervised learning ismore » able to rediscover that this class of galaxy is distinct from both ellipticals and spirals. We confirm that ~2.5 per cent galaxies are misclassified by GZ1 in our study. After correcting these galaxies’ labels, we improve our CNN performance to an average accuracy of over 0.99 (accuracy of 0.994 is our best result).« less

Authors:
 [1];  [1];  [1];  [1];  [2];  [3];  [4];  [5];  [5];  [6];  [7];  [8];  [4];  [9];  [10];  [11]
  1. Nottingham U.
  2. Cambridge U., KICC
  3. Zurich, ETH
  4. Fermilab
  5. U. Coll. London
  6. Madrid, IFT
  7. Harvard-Smithsonian Ctr. Astrophys.
  8. Macquarie U.
  9. Southhampton U.
  10. LIneA, Rio de Janeiro
  11. Michigan U.
Publication Date:
Research Org.:
Fermi National Accelerator Lab. (FNAL), Batavia, IL (United States)
Sponsoring Org.:
USDOE Office of Science (SC), High Energy Physics (HEP)
OSTI Identifier:
1648538
Report Number(s):
FERMILAB-PUB-20-236-SCD
Journal ID: ISSN 0035-8711; oai:inspirehep.net:1812528
DOE Contract Number:  
AC02-07CH11359
Resource Type:
Journal Article
Journal Name:
Monthly Notices of the Royal Astronomical Society
Additional Journal Information:
Journal Volume: 493; Journal Issue: 3; Journal ID: ISSN 0035-8711
Country of Publication:
United States
Language:
English
Subject:
79 ASTRONOMY AND ASTROPHYSICS

Citation Formats

Cheng, Ting-Yun, Conselice, Christopher J., Aragon-Salamanca, Alfonso, Li, Nan, Bluck, Asa F.L., Hartley, Will G., Annis, James, Brooks, David, Doel, Peter, Garcia-Bellido, Juan, James, David J., Kuehn, Kyler, Kuropatkin, Nikolay, Smith, Mathew, Sobreira, Flavia, and Tar;e. Gregory, Tar;e. Gregory. Optimising Automatic Morphological Classification of Galaxies with Machine Learning and Deep Learning using Dark Energy Survey Imaging. United States: N. p., 2020. Web. doi:10.1093/mnras/staa501.
Cheng, Ting-Yun, Conselice, Christopher J., Aragon-Salamanca, Alfonso, Li, Nan, Bluck, Asa F.L., Hartley, Will G., Annis, James, Brooks, David, Doel, Peter, Garcia-Bellido, Juan, James, David J., Kuehn, Kyler, Kuropatkin, Nikolay, Smith, Mathew, Sobreira, Flavia, & Tar;e. Gregory, Tar;e. Gregory. Optimising Automatic Morphological Classification of Galaxies with Machine Learning and Deep Learning using Dark Energy Survey Imaging. United States. https://doi.org/10.1093/mnras/staa501
Cheng, Ting-Yun, Conselice, Christopher J., Aragon-Salamanca, Alfonso, Li, Nan, Bluck, Asa F.L., Hartley, Will G., Annis, James, Brooks, David, Doel, Peter, Garcia-Bellido, Juan, James, David J., Kuehn, Kyler, Kuropatkin, Nikolay, Smith, Mathew, Sobreira, Flavia, and Tar;e. Gregory, Tar;e. Gregory. Fri . "Optimising Automatic Morphological Classification of Galaxies with Machine Learning and Deep Learning using Dark Energy Survey Imaging". United States. https://doi.org/10.1093/mnras/staa501. https://www.osti.gov/servlets/purl/1648538.
@article{osti_1648538,
title = {Optimising Automatic Morphological Classification of Galaxies with Machine Learning and Deep Learning using Dark Energy Survey Imaging},
author = {Cheng, Ting-Yun and Conselice, Christopher J. and Aragon-Salamanca, Alfonso and Li, Nan and Bluck, Asa F.L. and Hartley, Will G. and Annis, James and Brooks, David and Doel, Peter and Garcia-Bellido, Juan and James, David J. and Kuehn, Kyler and Kuropatkin, Nikolay and Smith, Mathew and Sobreira, Flavia and Tar;e. Gregory, Tar;e. Gregory},
abstractNote = {There are several supervised machine learning methods used for the application of automated morphological classification of galaxies; however, there has not yet been a clear comparison of these different methods using imaging data, or an investigation for maximizing their effectiveness. We carry out a comparison between several common machine learning methods for galaxy classification [Convolutional Neural Network (CNN), K-nearest neighbour, logistic regression, Support Vector Machine, Random Forest, and Neural Networks] by using Dark Energy Survey (DES) data combined with visual classifications from the Galaxy Zoo 1 project (GZ1). Our goal is to determine the optimal machine learning methods when using imaging data for galaxy classification. We show that CNN is the most successful method of these ten methods in our study. Using a sample of ~2800 galaxies with visual classification from GZ1, we reach an accuracy of ~0.99 for the morphological classification of ellipticals and spirals. The further investigation of the galaxies that have a different ML and visual classification but with high predicted probabilities in our CNN usually reveals the incorrect classification provided by GZ1. We further find the galaxies having a low probability of being either spirals or ellipticals are visually lenticulars (S0), demonstrating that supervised learning is able to rediscover that this class of galaxy is distinct from both ellipticals and spirals. We confirm that ~2.5 per cent galaxies are misclassified by GZ1 in our study. After correcting these galaxies’ labels, we improve our CNN performance to an average accuracy of over 0.99 (accuracy of 0.994 is our best result).},
doi = {10.1093/mnras/staa501},
url = {https://www.osti.gov/biblio/1648538}, journal = {Monthly Notices of the Royal Astronomical Society},
issn = {0035-8711},
number = 3,
volume = 493,
place = {United States},
year = {2020},
month = {2}
}

Works referenced in this record:

A New Approach to Galaxy Morphology. I. Analysis of the Sloan Digital Sky Survey Early Data Release
journal, May 2003


Automated Lensing Learner: Automated Strong Lensing Identification with a Computer Vision Technique
journal, May 2019


Galaxy types in the Sloan Digital Sky Survey using supervised artificial neural networks
journal, March 2004


Robust Machine Learning Applied to Astronomical Data Sets. I. Star‐Galaxy Classification of the Sloan Digital Sky Survey DR3 Using Decision Trees
journal, October 2006


Galaxy Zoo: the dependence of morphology and colour on environment
journal, March 2009


Galaxy Zoo: reproducing galaxy morphologies via machine learning★: Galaxy Zoo: morphology via machine learning
journal, April 2010


Integrating human and machine intelligence in galaxy morphology classification tasks
journal, March 2018


The use of the area under the ROC curve in the evaluation of machine learning algorithms
journal, July 1997


Random Forests
journal, January 2001


Restricted Boltzmann machine and softmax regression for fault detection and classification
journal, August 2017


The Relationship between Stellar Light Distributions of Galaxies and Their Formation Histories
journal, July 2003


Nearest neighbor pattern classification
journal, January 1967


Machine learning and image analysis for morphological galaxy classification
journal, March 2004


The Dark Energy Survey: Data Release 1
journal, November 2018


Rotation-invariant convolutional neural networks for galaxy morphology prediction
journal, April 2015


Improving galaxy morphologies for SDSS with Deep Learning
journal, February 2018


Dark Energy Survey Year 1 Results: The Photometric Data Set for Cosmology
journal, April 2018


Random forest automated supervised classification of Hipparcos periodic variable stars: Classification of Hipparcos periodic stars
journal, May 2011


Random forests: from early developments to recent advancements
journal, January 2014


An introduction to ROC analysis
journal, June 2006


The dark Energy Camera
journal, October 2015


Cognitron: A self-organizing multilayered neural network
journal, January 1975


Neocognitron: A neural network model for a mechanism of visual pattern recognition
journal, September 1983


Support vector machines and kd-tree for separating quasars from large survey data bases
journal, May 2008


Training Products of Experts by Minimizing Contrastive Divergence
journal, August 2002


An automatic taxonomy of galaxy morphology using unsupervised machine learning
journal, September 2017


Extragalactic nebulae.
journal, October 1926


A Catalog of Visual-Like Morphologies in the 5 Candels Fields Using deep Learning
journal, October 2015


Deep Learning Identifies High- z Galaxies in a Central Blue Nugget Phase in a Characteristic Mass Range
journal, May 2018


Revisiting the Hubble sequence in the SDSS DR7 spectroscopic sample: a publicly available Bayesian automated classification
journal, December 2010


Using machine learning to explore the long-term evolution of GRS 1915+105
journal, December 2016


Handwritten Marathi Character Recognition Using R-HOG Feature
journal, January 2015


Star–galaxy separation strategies for WISE-2MASS all-sky infrared galaxy catalogues
journal, February 2015


Neural computation as a tool for galaxy classification: methods and examples
journal, October 1996


Gradient-based learning applied to document recognition
journal, January 1998


Galaxy Zoo: morphologies derived from visual inspection of galaxies from the Sloan Digital Sky Survey
journal, September 2008


Galaxy Zoo 1: data release of morphological classifications for nearly 900 000 galaxies★: Galaxy Zoo
journal, November 2010


Automated Source Classification Using a Kohonen Network
journal, October 1995


The strong gravitational lens finding challenge
journal, May 2019


Automated morphological classification of APM galaxies by supervised artificial neural networks
journal, August 1995


Automated star/galaxy discrimination with neural networks
journal, January 1992


Learning representations by back-propagating errors
journal, October 1986


Restricted Boltzmann machines for collaborative filtering
conference, January 2007


Automatic morphological classification of galaxy images
journal, November 2009


The optimal distance measure for nearest neighbor classification
journal, September 1981


Precision spectroscopy of kaonic [sup 3]He X-rays at J-PARC
conference, January 2011

  • Sato, M.; Bhang, H.; Cargnelli, M.
  • 12TH INTERNATIONAL CONFERENCE ON MESON-NUCLEON PHYSICS AND THE STRUCTURE OF THE NUCLEON (MENU 2010), AIP Conference Proceedings
  • https://doi.org/10.1063/1.3647129

Galaxy And Mass Assembly: automatic morphological classification of galaxies using statistical learning
journal, November 2017


Morphological Classification of galaxies by Artificial Neural Networks
journal, November 1992


Automated Star/Galaxy Classification for Digitized Poss-II
journal, June 1995


Galaxy Zoo 2: detailed morphological classifications for 304 122 galaxies from the Sloan Digital Sky Survey
journal, September 2013


Support Vector Machines (SVMs) versus Multilayer Perception (MLP) in data classification
journal, November 2012