Transfer learning for galaxy morphology from one survey to another

Abstract

Deep Learning (DL) algorithms for morphological classification of galaxies have proven very successful, mimicking (or even improving) visual classifications. However, these algorithms rely on large training samples of labelled galaxies (typically thousands of them). A key question for using DL classifications in future Big Data surveys is how much of the knowledge acquired from an existing survey can be exported to a new dataset, i.e. if the features learned by the machines are meaningful for different data. We test the performance of DL models, trained with Sloan Digital Sky Survey (SDSS) data, on Dark Energy survey (DES) using images for a sample of $\sim$5000 galaxies with a similar redshift distribution to SDSS. Applying the models directly to DES data provides a reasonable global accuracy ($\sim$ 90%), but small completeness and purity values. A fast domain adaptation step, consisting in a further training with a small DES sample of galaxies ($\sim$500-300), is enough for obtaining an accuracy > 95% and a significant improvement in the completeness and purity values. This demonstrates that, once trained with a particular dataset, machines can quickly adapt to new instrument characteristics (e.g., PSF, seeing, depth), reducing by almost one order of magnitude the necessary training sample for morphological classification. Redshift evolution effects or significant depth differences are not taken into account in this study.

arXiv commentAccepted for publication in MNRAS

References in corpus (33)

• The Seventh Data Release of the Sloan Digital Sky Survey
• Convolutional Neural Networks for Medical Image Analysis: Full Training or Fine Tuning?
• Detailed Structural Decomposition of Galaxy Images
• SDSS-III: Massive Spectroscopic Surveys of the Distant Universe, the Milky Way Galaxy, and Extra-Solar Planetary Systems
• CANDELS: The Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey
• The Cosmic Evolution Survey (COSMOS) -- Overview
• The Dark Energy Camera
• The Dark Energy Survey: more than dark energy - an overview
• The Relationship Between Stellar Light Distributions of Galaxies and their Formation Histories
• Cluster Lensing And Supernova survey with Hubble (CLASH): An Overview
• Rotation-invariant convolutional neural networks for galaxy morphology prediction
• Galaxy Zoo 1 : Data Release of Morphological Classifications for nearly 900,000 galaxies
• The DEEP2 Galaxy Redshift Survey: Design, Observations, Data Reduction, and Redshifts
• Galaxy Zoo 2: detailed morphological classifications for 304,122 galaxies from the Sloan Digital Sky Survey
• The Horizon-AGN simulation: morphological diversity of galaxies promoted by AGN feedback
• A Catalog of Detailed Visual Morphological Classifications for 14034 Galaxies in the Sloan Digital Sky Survey
• Improving galaxy morphologies for SDSS with Deep Learning
• A catalog of visual-like morphologies in the 5 CANDELS fields using deep-learning
• The Horizon-AGN simulation: evolution of galaxy properties over cosmic time
• The long lives of giant clumps and the birth of outflows in gas-rich galaxies at high redshift
• A catalogue of 2D photometric decompositions in the SDSS-DR7 spectroscopic main galaxy sample: preferred models and systematics
• Galaxy Zoo: Reproducing Galaxy Morphologies Via Machine Learning
• The importance of minor-merger-driven star formation and black-hole growth in disk galaxies
• Using transfer learning to detect galaxy mergers
• Neural computation as a tool for galaxy classification : methods and examples
• P-MaNGA: Emission Lines Properties - Gas Ionisation and Chemical Abundances from Prototype Observations
• Galaxies, Human Eyes and Artificial Neural Networks
• The Euclid mission design
• Deep learning for galaxy surface brightness profile fitting
• A Catalogue of Two-Dimensional Photometric Decompositions in the SDSS-DR7 Spectroscopic Main Galaxy Sample: Extension to g- and i-Bands
• SpotTune: Transfer Learning through Adaptive Fine-tuning
• Improving galaxy morphology with machine learning
• Transfer Learning Using Classification Layer Features of CNN

Cited by in corpus (61)

• Galaxy Zoo DECaLS: Detailed Visual Morphology Measurements from Volunteers and Deep Learning for 314,000 Galaxies
• Galaxy Zoo: Probabilistic Morphology through Bayesian CNNs and Active Learning
• Effect of galaxy mergers on star formation rates
• Deep learning predictions of galaxy merger stage and the importance of observational realism
• LinKS: Discovering galaxy-scale strong lenses in the Kilo-Degree Survey using Convolutional Neural Networks
• Morpheus: A Deep Learning Framework For Pixel-Level Analysis of Astronomical Image Data
• Convolutional neural network identification of galaxy post-mergers in UNIONS using IllustrisTNG
• Self-Supervised Representation Learning for Astronomical Images
• Identifying Exoplanets with Deep Learning II: Two New Super-Earths Uncovered by a Neural Network in K2 Data
• Practical Galaxy Morphology Tools from Deep Supervised Representation Learning
• Using Convolutional Neural Networks to identify Gravitational Lenses in Astronomical images
• Deep Transfer Learning for Star Cluster Classification: I. Application to the PHANGS-HST Survey
• Morphological classification of galaxies with deep learning: comparing 3-way and 4-way CNNs
• Classifying the unknown: discovering novel gravitational-wave detector glitches using similarity learning
• Deblending galaxies with Variational Autoencoders: a joint multi-band, multi-instrument approach
• Pushing automated morphological classifications to their limits with the Dark Energy Survey
• Detecting solar system objects with convolutional neural networks
• Transfer learning for radio galaxy classification
• Star formation characteristics of CNN-identified post-mergers in the Ultraviolet Near Infrared Optical Northern Survey (UNIONS)
• Automatic Classification of Galaxy Morphology: a rotationally invariant supervised machine learning method based on the UML-dataset
• The evolution of barred galaxies in the EAGLE simulations
• Identification of tidal features in deep optical galaxy images with Convolutional Neural Networks
• Explaining deep learning of galaxy morphology with saliency mapping
• AGN in post-mergers from the Ultraviolet Near Infrared Optical Northern Survey
• Cataloging Accreted Stars within Gaia DR2 using Deep Learning
• Unsupervised Galaxy Morphological Visual Representation with Deep Contrastive Learning
• Deep Learning for Cardiologist-level Myocardial Infarction Detection in Electrocardiograms
• Towards robust determination of non-parametric morphologies in marginal astronomical data: resolving uncertainties with cosmological hydrodynamical simulations
• On Improving the Performance of Glitch Classification for Gravitational Wave Detection by using Generative Adversarial Networks
• A catalog of broad morphology of Pan-STARRS galaxies based on deep learning
• Deep learning for Sunyaev-Zel'dovich detection in Planck
• Connecting optical morphology, environment, and HI mass fraction for low-redshift galaxies using deep learning
• The PAU survey: Estimating galaxy photometry with deep learning
• Euclid preparation. XLIII. Measuring detailed galaxy morphologies for Euclid with machine learning
• Protostellar classification using supervised machine learning algorithms
• An extended catalog of galaxy-galaxy strong gravitational lenses discovered in DES using convolutional neural networks
• Euclid Preparation XXXIII. Characterization of convolutional neural networks for the identification of galaxy-galaxy strong lensing events
• The Systematics of Galaxy Morphology in the Comprehensive de Vaucouleurs revised Hubble-Sandage Classification System: Application to the EFIGI Sample
• Subaru Hyper Suprime-Cam revisits the large-scale environmental dependence on galaxy morphology over 360 deg${}^2$ at z=0.3-0.6
• The redshift evolution of the S0 fraction for $z<1$ in COSMOS
• Deep Transfer Learning for Classification of Variable Sources
• Accelerating the search for Axion-Like Particles with machine learning
• Machine learning technique for morphological classification of galaxies from the SDSS. III. Image-based inference of detailed features
• Galaxy Zoo JWST: Up to 75% of discs are featureless at $3<z<7$
• Uncertainty-aware transfer across tasks using hybrid model-based successor feature reinforcement learning
• A morphological segmentation approach to determining bar lengths
• A Machine Learning Approach to Enhancing eROSITA Observations
• Using Machine Learning to Determine Morphologies of $z<1$ AGN Host Galaxies in the Hyper Suprime-Cam Wide Survey
• DES to HSC: Detecting low surface brightness galaxies in the Abell 194 cluster using transfer learning
• The ATLAS Virtual Research Assistant
• Using machine learning to study the kinematics of cold gas in galaxies
• A self-supervised, physics-aware, Bayesian neural network architecture for modelling galaxy emission-line kinematics
• COUGS-DESI: A Catalog of Unusual Galaxies with Polar Structures in the DESI Legacy Imaging Surveys
• Identification of molecular line emission using Convolutional Neural Networks
• Automatic Machine Learning Framework to Study Morphological Parameters of AGN Host Galaxies within $z<1.4$ in the Hyper Supreme-Cam Wide Survey
• J-PAS: A Neural Network Approach to Single Stellar Population Characterization
• DeepMerge: Classifying High-redshift Merging Galaxies with Deep Neural Networks
• The Next Generation Virgo Cluster Survey (NGVS). XL. The Morphological Classification of Virgo Cluster Galaxies
• Dark Matter profiles of "in silico" galaxies: deep learning inference
• Quantifying the Poor Purity and Completeness of Morphological Samples Selected by Galaxy Colour
• On The Transferability of Deep-Q Networks