DeepMerge II: Building Robust Deep Learning Algorithms for Merging Galaxy Identification Across Domains

Abstract

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 $\sim 20\%$. 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.

arXiv commentSubmitted to MNRAS; 21 pages, 9 figures, 9 tables

References in corpus (33)

• Very Deep Convolutional Networks for Large-Scale Image Recognition
• Grad-CAM: Visual Explanations from Deep Networks via Gradient-based Localization
• Deep Residual Learning for Image Recognition
• Cosmic Star Formation History
• The EAGLE project: Simulating the evolution and assembly of galaxies and their environments
• Introducing the Illustris Project: Simulating the coevolution of dark and visible matter in the Universe
• Deep Visual Domain Adaptation: A Survey
• Simulating Galaxy Formation with the IllustrisTNG Model
• First results from the IllustrisTNG simulations: matter and galaxy clustering
• The hierarchical formation of the brightest cluster galaxies
• First results from the IllustrisTNG simulations: the galaxy color bimodality
• Galaxy Zoo : Morphologies derived from visual inspection of galaxies from the Sloan Digital Sky Survey
• A New Non-Parametric Approach to Galaxy Morphological Classification
• Galaxy Zoo 1 : Data Release of Morphological Classifications for nearly 900,000 galaxies
• A Direct Measurement of Major Galaxy Mergers at z < 3
• GAMA: towards a physical understanding of galaxy formation
• The high redshift galaxy population in hierarchical galaxy formation models
• The Deep2 Galaxy Redshift Survey: Evolution of Close Galaxy Pairs and Major-Merger Rates Up to z ~ 1.2
• Galaxy Zoo: the fraction of merging galaxies in the SDSS and their morphologies
• Galaxy Morphology and Star Formation in the Illustris Simulation at z=0
• Using transfer learning to detect galaxy mergers
• Deep learning predictions of galaxy merger stage and the importance of observational realism
• Contrastive Adaptation Network for Unsupervised Domain Adaptation
• Galaxy Merger Rates up to z $\sim$ 3 using a Bayesian Deep Learning Model $-$ A Major-Merger classifier using IllustrisTNG Simulation data
• Deep Learning Identifies High-z Galaxies in a Central Blue Nugget Phase in a Characteristic Mass Range
• Identifying Galaxy Mergers in Observations and Simulations with Deep Learning
• Automated Distant Galaxy Merger Classifications from Space Telescope Images using the Illustris Simulation
• Stellar Masses of Giant Clumps in CANDELS and Simulated Galaxies Using Machine Learning
• Scalable Reinforcement-Learning-Based Neural Architecture Search for Cancer Deep Learning Research
• Towards a consistent framework of comparing galaxy mergers in observations and simulations
• Bars formed in galaxy merging and their classification with deep learning
• The nature of giant clumps in high-z discs: a deep-learning comparison of simulations and observations
• Fisher Deep Domain Adaptation

Cited by in corpus (28)

• The DAWES review 10: The impact of deep learning for the analysis of galaxy surveys
• IllustrisTNG in the HSC-SSP: image data release and the major role of mini mergers as drivers of asymmetry and star formation
• ERGO-ML I: Inferring the assembly histories of IllustrisTNG galaxies from integral observable properties via invertible neural networks
• The combined and respective roles of imaging and stellar kinematics in identifying galaxy merger remnants
• The scatter in the galaxy-halo connection: a machine learning analysis
• A complete catalogue of merger fractions in AGN hosts: No evidence for an increase in detected merger fraction with AGN luminosity
• Dust and Power: Unravelling the merger - active galactic nucleus connection in the second half of cosmic history
• Astronomaly at scale: searching for anomalies amongst 4 million galaxies
• DIGS: Deep Inference of Galaxy Spectra with Neural Posterior Estimation
• Machine Learning and Cosmology
• ERGO-ML: Comparing IllustrisTNG and HSC galaxy images via contrastive learning
• From Images to Features: Unbiased Morphology Classification via Variational Auto-Encoders and Domain Adaptation
• Target Detection Framework for Lobster Eye X-Ray Telescopes with Machine Learning Algorithms
• Machine learning the gap between real and simulated nebulae: A domain-adaptation approach to classify ionised nebulae in nearby galaxies
• Ask The Machine: Systematic detection of wind-type outflows in low-mass X-ray binaries
• SDSS-IV MaNGA: Unveiling Galaxy Interaction by Merger Stages with Machine Learning
• Machine learning technique for morphological classification of galaxies from the SDSS. III. Image-based inference of detailed features
• Inferring redshift and galaxy properties via a multi-task neural net with probabilistic outputs: An application to simulated MOONS spectra
• A post-merger enhancement only in star-forming Type 2 Seyfert galaxies: the deep learning view
• Automated classification of Chandra X-ray point sources using machine learning methods
• Classifying merger stages with adaptive deep learning and cosmological hydrodynamical simulations
• Using AI for Wavefront Estimation with the Rubin Observatory Active Optics System
• LenNet: Direct Detection and Localization of Strong Gravitational Lenses in Wide-Field Sky Survey Images
• Flexible Simulation Based Inference for Galaxy Photometric Fitting with Synthesizer
• The major merger-active galactic nucleus connection up to the cosmic noon
• A Simulation Driven Deep Learning Approach for Separating Mergers and Star Forming Galaxies: The Formation Histories of Clumpy Galaxies in all the CANDELS Fields
• Merger fraction in galaxy groups and clusters at z < 0.2: A non-parametric morphological study with Subaru Hyper Suprime-Cam
• Timing the last major merger of galaxy clusters with large halo sparsity