Skip to main content
OSTI.GOVU.S. Department of Energy Office of Scientific and Technical Information
U.S. Department of Energy
Office of Scientific and Technical Information
 

DeepMerge – II. Building robust deep learning algorithms for merging galaxy identification across domains

Journal Article · · Monthly Notices of the Royal Astronomical Society
; ; ; ; ; ; ; ;

In astronomy, neural networks are often trained on simulation data with the prospect of being used on telescope observations. Unfortunately, training a model on simulation data and then applying it to instrument data leads to a substantial and potentially even detrimental decrease in model accuracy on the new target dataset. Simulated and instrument data represent different data domains, and for an algorithm to work in both, domain-invariant learning is necessary. Here we employ domain adaptation techniques— Maximum Mean Discrepancy (MMD) as an additional transfer loss and Domain Adversarial Neural Networks (DANNs)— and demonstrate their viability to extract domain-invariant features within the astronomical context of classifying merging and non-merging galaxies. Additionally, we explore the use of Fisher loss and entropy minimization to enforce better in-domain class discriminability. We show that the addition of each domain adaptation technique improves the performance of a classifier when compared to conventional deep learning algorithms. We demonstrate this on two examples: between two Illustris-1 simulated datasets of distant merging galaxies, and between Illustris-1 simulated data of nearby merging galaxies and observed data from the Sloan Digital Sky Survey. The use of domain adaptation techniques in our experiments leads to an increase of target domain classification accuracy of up to ~20%. With further development, these techniques will allow astronomers to successfully implement neural network models trained on simulation data to efficiently detect and study astrophysical objects in current and future large-scale astronomical surveys.

Research Organization:
Argonne National Laboratory (ANL), Argonne, IL (United States)
Sponsoring Organization:
Argonne National Laboratory - Argonne Leadership Computing Facility; USDOE Office of Science - Office of High Energy Physics; Fermi Research Alliance, LLC
DOE Contract Number:
AC02-06CH11357
OSTI ID:
1840563
Journal Information:
Monthly Notices of the Royal Astronomical Society, Vol. 506, Issue 1
Publisher:
Royal Astronomical Society
Country of Publication:
United States
Language:
English

References (42)

A Mathematical Theory of Communication July 1948
Towards a consistent framework of comparing galaxy mergers in observations and simulations December 2020
The high-redshift galaxy population in hierarchical galaxy formation models March 2007
The IllustrisTNG simulations: public data release May 2019
A Survey of Unsupervised Deep Domain Adaptation September 2020
DeepHyper: Asynchronous Hyperparameter Search for Deep Neural Networks December 2018
A New Nonparametric Approach to Galaxy Morphological Classification July 2004
Galaxy morphology and star formation in the Illustris Simulation at z  = 0 October 2015
The hierarchical formation of the brightest cluster galaxies February 2007
Scatter Component Analysis: A Unified Framework for Domain Adaptation and Domain Generalization July 2017
Identifying galaxy mergers in observations and simulations with deep learning June 2019
Galaxy Merger Rates up to z ∼ 3 Using a Bayesian Deep Learning Model: A Major-merger Classifier Using IllustrisTNG Simulation Data June 2020
How to Use t-SNE Effectively October 2016
The nature of giant clumps in high- z discs: a deep-learning comparison of simulations and observations December 2020
The EAGLE project: simulating the evolution and assembly of galaxies and their environments November 2014
On Information and Sufficiency March 1951
Cosmic Star-Formation History August 2014
First results from the IllustrisTNG simulations: matter and galaxy clustering December 2017
Grad-CAM: Visual Explanations from Deep Networks via Gradient-Based Localization October 2019
A Direct Measurement of Major Galaxy Mergers at z 3 September 2003
ImageNet: A large-scale hierarchical image database
  • No authors listed
  • 2009 IEEE Computer Society Conference on Computer Vision and Pattern Recognition Workshops (CVPR Workshops), 2009 IEEE Conference on Computer Vision and Pattern Recognition https://doi.org/10.1109/CVPR.2009.5206848
June 2009
Deep Residual Learning for Image Recognition June 2016
Deep Learning Identifies High- z Galaxies in a Central Blue Nugget Phase in a Characteristic Mass Range May 2018
The DEEP2 Galaxy Redshift Survey: Evolution of Close Galaxy Pairs and Major-Merger Rates up to z ~ 1.2 November 2004
Simulating galaxy formation with the IllustrisTNG model October 2017
DeepMerge: Classifying high-redshift merging galaxies with deep neural networks July 2020
Galaxy Zoo: the fraction of merging galaxies in the SDSS and their morphologies January 2010
Automated distant galaxy merger classifications from Space Telescope images using the Illustris simulation April 2019
First results from the IllustrisTNG simulations: the galaxy colour bimodality November 2017
Deep visual domain adaptation: A survey October 2018
Scalable reinforcement-learning-based neural architecture search for cancer deep learning research
  • No authors listed
  • SC '19: The International Conference for High Performance Computing, Networking, Storage, and Analysis, Proceedings of the International Conference for High Performance Computing, Networking, Storage and Analysis https://doi.org/10.1145/3295500.3356202
November 2019
Domain Adaptation via Transfer Component Analysis February 2011
Deep learning predictions of galaxy merger stage and the importance of observational realism October 2019
Deep CORAL: Correlation Alignment for Deep Domain Adaptation January 2016
Introducing the Illustris Project: simulating the coevolution of dark and visible matter in the Universe August 2014
GAMA: towards a physical understanding of galaxy formation October 2009
Galaxy Zoo 1: data release of morphological classifications for nearly 900 000 galaxies★: Galaxy Zoo November 2010
Stellar masses of giant clumps in CANDELS and simulated galaxies using machine learning September 2020
Contrastive Adaptation Network for Unsupervised Domain Adaptation June 2019
Bars formed in galaxy merging and their classification with deep learning September 2020
Galaxy Zoo: morphologies derived from visual inspection of galaxies from the Sloan Digital Sky Survey ★ September 2008
Using transfer learning to detect galaxy mergers May 2018

Cited By (1)

DeepGhostBusters: Using Mask R-CNN to detect and mask ghosting and scattered-light artifacts from optical survey images April 2022

Similar Records

Related Subjects

galaxies: interactions
methods: data analysis
techniques: image processing