Detecting outliers in astronomical images with deep generative networks

Abstract

With the advent of future big-data surveys, automated tools for unsupervised discovery are becoming ever more necessary. In this work, we explore the ability of deep generative networks for detecting outliers in astronomical imaging datasets. The main advantage of such generative models is that they are able to learn complex representations directly from the pixel space. Therefore, these methods enable us to look for subtle morphological deviations which are typically missed by more traditional moment-based approaches. We use a generative model to learn a representation of expected data defined by the training set and then look for deviations from the learned representation by looking for the best reconstruction of a given object. In this first proof-of-concept work, we apply our method to two different test cases. We first show that from a set of simulated galaxies, we are able to detect $\sim 90\%$ of merging galaxies if we train our network only with a sample of isolated ones. We then explore how the presented approach can be used to compare observations and hydrodynamic simulations by identifying observed galaxies not well represented in the models.

References in corpus (48)

• Stellar population synthesis at the resolution of 2003
• Galactic Stellar and Substellar Initial Mass Function
• Seven-Year Wilkinson Microwave Anisotropy Probe (WMAP) Observations: Cosmological Interpretation
• Star Formation in Galaxies Along the Hubble Sequence
• CANDELS: The Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey
• Cosmological Hydrodynamics with Adaptive Mesh Refinement: a new high resolution code called RAMSES
• CANDELS: The Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey - The Hubble Space Telescope Observations, Imaging Data Products and Mosaics
• Radiative Transfer in a Clumpy Universe: II. The Utraviolet Extragalactic Background
• Progressive Growing of GANs for Improved Quality, Stability, and Variation
• Improved Training of Wasserstein GANs
• Dancing in the dark: galactic properties trace spin swings along the cosmic web
• The Relationship Between Stellar Light Distributions of Galaxies and their Formation Histories
• A New Non-Parametric Approach to Galaxy Morphological Classification
• The merger rate of galaxies in the Illustris Simulation: a comparison with observations and semi-empirical models
• Wasserstein GAN
• The optical morphologies of galaxies in the IllustrisTNG simulation: a comparison to Pan-STARRS observations
• A New Approach to Galaxy Morphology: I. Analysis of the Sloan Digital Sky Survey Early Data Release
• Improving galaxy morphologies for SDSS with Deep Learning
• A Critical Assessment of Photometric Redshift Methods: A CANDELS Investigation
• Self-regulated growth of supermassive black holes by a dual jet/heating AGN feedback mechanism: methods, tests and implications for cosmological simulations
• The origin and implications of dark matter anisotropic cosmic infall on ~L* haloes
• The evolution of the dust and gas content in galaxies
• A catalog of visual-like morphologies in the 5 CANDELS fields using deep-learning
• Building Merger Trees from Cosmological N-body Simulations
• The Horizon-AGN simulation: evolution of galaxy properties over cosmic time
• Star-galaxy Classification Using Deep Convolutional Neural Networks
• CANDELS Multiwavelength Catalogs: Source Identification and Photometry in the CANDELS Extended Groth Strip
• Photometric redshifts from SDSS images using a Convolutional Neural Network
• The CANDELS/SHARDS multi-wavelength catalog in GOODS-N: Photometry, Photometric Redshifts, Stellar Masses, Emission line fluxes and Star Formation Rates
• CANDELS Multiwavelength Catalogs: Source Identification and Photometry in the CANDELS COSMOS Survey Field
• Finding strong lenses in CFHTLS using convolutional neural networks
• The weirdest SDSS galaxies: results from an outlier detection algorithm
• Deep-HiTS: Rotation Invariant Convolutional Neural Network for Transient Detection
• Deep Learning Identifies High-z Galaxies in a Central Blue Nugget Phase in a Characteristic Mass Range
• Deep learning for galaxy surface brightness profile fitting
• Finding outlier light-curves in catalogs of periodic variable stars
• A catalog of polychromatic bulge-disk decompositions of ~ 17.600 galaxies in CANDELS
• Horizon-AGN virtual observatory - 1. SED-fitting performance and forecasts for future imaging surveys
• Photometry of high-redshift blended galaxies using deep learning
• Star formation rates and stellar masses from machine learning
• Discovering the Unexpected in Astronomical Survey Data
• Unusual quasars from the Sloan Digital Sky Survey selected by means of Kohonen self-organising maps
• Horizon-AGN virtual observatory -- 2: Template-free estimates of galaxy properties from colours
• Galaxy And Mass Assembly: Automatic Morphological Classification of Galaxies Using Statistical Learning
• Automated novelty detection in the WISE survey with one-class support vector machines
• A New Detection of Extragalactic Anomalous Microwave Emission in a Compact, Optically-Faint Region of NGC\,4725
• Identifying Mergers Using Quantitative Morphologies in Zoom Simulations of High-Redshift Galaxies
• Anomaly detection in wide area network mesh using two machine learning anomaly detection algorithms

Cited by in corpus (35)

• Galaxy Zoo DECaLS: Detailed Visual Morphology Measurements from Volunteers and Deep Learning for 314,000 Galaxies
• The DAWES review 10: The impact of deep learning for the analysis of galaxy surveys
• Astronomaly: Personalised Active Anomaly Detection in Astronomical Data
• Self-Supervised Representation Learning for Astronomical Images
• HOLISMOKES -- II. Identifying galaxy-scale strong gravitational lenses in Pan-STARRS using convolutional neural networks
• The relationship between fine galaxy stellar morphology and star formation activity in cosmological simulations: a deep learning view
• Beyond the Hubble Sequence -- Exploring Galaxy Morphology with Unsupervised Machine Learning
• AstroCLIP: A Cross-Modal Foundation Model for Galaxies
• IllustrisTNG in the HSC-SSP: image data release and the major role of mini mergers as drivers of asymmetry and star formation
• Anomaly detection in Hyper Suprime-Cam galaxy images with generative adversarial networks
• Data mining techniques on astronomical spectra data. II : Classification Analysis
• Euclid preparation: XIII. Forecasts for galaxy morphology with the Euclid Survey using Deep Generative Models
• Learning to detect RFI in radio astronomy without seeing it
• A method for finding anomalous astronomical light curves and their analogs
• Encoding large scale cosmological structure with Generative Adversarial Networks
• A Robust Study of High-Redshift Galaxies: Unsupervised Machine Learning for Characterising morphology with JWST up to z ~ 8
• The nature of giant clumps in high-z discs: a deep-learning comparison of simulations and observations
• The ROAD to discovery: machine learning-driven anomaly detection in radio astronomy spectrograms
• Galaxy merger challenge: A comparison study between machine learning-based detection methods
• The Gaia-ESO Survey: Preparing the ground for 4MOST & WEAVE galactic surveys. Chemical evolution of lithium with machine learning
• Star formation and morphological properties of galaxies in the Pan-STARRS $3\pi$ survey- I. A machine learning approach to galaxy and supernova classification
• Identifying outliers in astronomical images with unsupervised machine learning
• ERGO-ML: Comparing IllustrisTNG and HSC galaxy images via contrastive learning
• HOLISMOKES -- XI. Evaluation of supervised neural networks for strong-lens searches in ground-based imaging surveys
• COSMOS-Web: A history of galaxy migrations over the stellar mass-star formation rate plane
• Searching for Anomalies in the ZTF Catalog of Periodic Variable Stars
• 3D Detection and Characterisation of ALMA Sources through Deep Learning
• Automatic identification of outliers in Hubble Space Telescope galaxy images
• The emptiness inside: Finding gaps, valleys, and lacunae with geometric data analysis
• AGN -- host galaxy photometric decomposition using a fast, accurate and precise deep learning approach
• Quantitatively rating galaxy simulations against real observations with anomaly detection
• Bridging Simulations and Observations: New Insights into Galaxy Formation Simulations via Out-of-Distribution Detection and Bayesian Model Comparison
• Identification of Anomalous E+A Galaxies in GAMA Using an Isolation Forest
• Where's Swimmy?: Mining unique color features buried in galaxies by deep anomaly detection using Subaru Hyper Suprime-Cam data
• J-PLUS: Understanding outlier white dwarfs in the third data release via dimensionality reduction