Transfer learning for galaxy morphology from one survey to another

H. Dominguez Sanchez; M. Huertas-Company; M. Bernardi; S. Kaviraj; J. L. Fischer; T. M. C. Abbott; F. B. Abdalla; J. Annis; S. Avila; D. Brooks; E. Buckley-Geer; A. Carnero Rosell; M. Carrasco Kind; J. Carretero; C. E. Cunha; C. B. D'Andrea; L. N. da Costa; C. Davis; J. De Vicente; P. Doel; A. E. Evrard; P. Fosalba; J. Frieman; J. Garcia-Bellido; E. Gaztanaga; D. W. Gerdes; D. Gruen; R. A. Gruendl; J. Gschwend; G. Gutierrez; W. G. Hartley; D. L. Hollowood; K. Honscheid; B. Hoyle; D. J. James; K. Kuehn; N. Kuropatkin; O. Lahav; M. A. G. Maia; M. March; P. Melchior; F. Menanteau; R. Miquel; B. Nord; A. A. Plazas; E. Sanchez; V. Scarpine; R. Schindler; M. Schubnell; M. Smith; R. C. Smith; M. Soares-Santos; F. Sobreira; E. Suchyta; M. E. C. Swanson; G. Tarle; D. Thomas; A. R. Walker; J. Zuntz

doi:10.1093/mnras/sty3497

Back

Transfer learning for galaxy morphology from one survey to another

Journal article

Open access

Peer reviewed

Transfer learning for galaxy morphology from one survey to another

H. Dominguez Sanchez, M. Huertas-Company, M. Bernardi, S. Kaviraj, J. L. Fischer, T. M. C. Abbott, F. B. Abdalla, J. Annis, S. Avila, D. Brooks, …

Monthly notices of the Royal Astronomical Society, Vol.484(1), pp.93-100

03/21/2019

DOI: https://doi.org/10.1093/mnras/sty3497

Abstract

Astronomy & Astrophysics

Science & Technology

Physical Sciences

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 data set, 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 similar to 5000 galaxies with a similar redshift distribution to SDSS. Applying the models directly to DES data provides a reasonable global accuracy (similar to 90 per cent), but small completeness and purity values. A fast domain adaptation step, consisting of a further training with a small DES sample of galaxies (similar to 500-300), is enough for obtaining an accuracy >95 per cent and a significant improvement in the completeness and purity values. This demonstrates that, once trained with a particular data set, 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.

Files and links (1)

url

https://doi.org/10.1093/mnras/sty3497View

Published (Version of record) Open

Metrics

1 Record Views

See more details

Details

Title: Transfer learning for galaxy morphology from one survey to another
Creators: H. Dominguez Sanchez - University of Pennsylvania
M. Huertas-Company - Instituto de Astrofísica de Canarias
M. Bernardi - University of Pennsylvania
S. Kaviraj - University of Hertfordshire
J. L. Fischer - University of Pennsylvania
T. M. C. Abbott - Cerro Tololo Inter-American Observatory
F. B. Abdalla - Rhodes University
J. Annis - Fermi National Accelerator Laboratory
S. Avila - University of Portsmouth
D. Brooks - University College London
E. Buckley-Geer - Fermi National Accelerator Laboratory
A. Carnero Rosell - Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas
M. Carrasco Kind - University of Illinois Urbana-Champaign
J. Carretero - Institute for High Energy Physics
C. E. Cunha - Fermi National Accelerator Laboratory
C. B. D'Andrea - Univ Penn, Dept Phys & Astron, Philadelphia, PA 19104 USA
L. N. da Costa - Laboratório Interinstitucional de e-Astronomia
C. Davis - University of Louisville
J. De Vicente - Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas
P. Doel - University College London
A. E. Evrard - University of Michigan
P. Fosalba - Institut d'Estudis Espacials de Catalunya
J. Frieman - Fermi National Accelerator Laboratory
J. Garcia-Bellido - Universidad Autónoma de Madrid
E. Gaztanaga - S&N
D. W. Gerdes - Institute of Space Science - INFLPR Subsidiary
D. Gruen - Kavli Institute for Particle Astrophysics and Cosmology
R. A. Gruendl - National Center for Supercomputing Applications
J. Gschwend - Laboratório Interinstitucional de e-Astronomia
G. Gutierrez - Fermi National Accelerator Laboratory
W. G. Hartley - University College London
D. L. Hollowood - University of California, Santa Cruz
K. Honscheid - The Ohio State University
B. Hoyle - Max Planck Institute for Extraterrestrial Physics
D. J. James - Center for Astrophysics Harvard & Smithsonian
K. Kuehn - Australian Astronomical Observatory
N. Kuropatkin - Fermi National Accelerator Laboratory
O. Lahav - University College London
M. A. G. Maia - Laboratório Interinstitucional de e-Astronomia
M. March - University of Pennsylvania
P. Melchior - Princeton University
F. Menanteau - National Center for Supercomputing Applications
R. Miquel - Institució Catalana de Recerca i Estudis Avançats
B. Nord - Fermi National Accelerator Laboratory
A. A. Plazas - Jet Propulsion Laboratory
E. Sanchez - Centro de Investigaciones Energéticas, Medioambientales y Tecnológicas
V. Scarpine - Fermi National Accelerator Laboratory
R. Schindler - SLAC National Accelerator Laboratory
M. Schubnell - University of Michigan
M. Smith - University of Southampton
R. C. Smith - Cerro Tololo Inter-American Observatory
M. Soares-Santos - Brandeis University, Martin A. Fisher School of Physics
F. Sobreira - Universidade Estadual de Campinas (UNICAMP)
E. Suchyta - Oak Ridge National Laboratory
M. E. C. Swanson - National Center for Supercomputing Applications
G. Tarle - Institute for High Energy Physics
D. Thomas - University of Portsmouth
A. R. Walker - Cerro Tololo Inter-American Observatory
J. Zuntz - UK Astronomy Technology Centre
Publication Details: Monthly notices of the Royal Astronomical Society, Vol.484(1), pp.93-100
Publisher: Oxford Univ Press
Number of pages: 8
Grant note: NCSA(UIUC) CNPQ; Conselho Nacional de Desenvolvimento Cientifico e Tecnologico (CNPQ) Dark Energy Survey KICP(U. Chicago) NSF(USA); National Science Foundation (NSF) HEFCE(UK); UK Research & Innovation (UKRI); Higher Education Funding Council for England (HEFCE) MEC/MICINN/MINECO(Spain); Spanish Government CE110001020 / Australian Research Council Centre of Excellence for All-sky Astrophysics (CAAS-TRO); Australian Research Council AST-1138766; AST-1536171; AST-1816330 / NSF; National Science Foundation (NSF) MIFPA(Texas AM)
Identifiers: 9924552835601921
Academic Unit: Martin A. Fisher School of Physics
Language: English
Resource Type: Journal article

Transfer learning for galaxy morphology from one survey to another

Abstract

Files and links (1)

Metrics

Details

Brandeis University Social media