DeepZipper: A Novel Deep-learning Architecture for Lensed Supernovae Identification

R. Morgan; B. Nord; K. Bechtol; S. J. González; E. Buckley-Geer; A. Möller; J. W. Park; A. G. Kim; S. Birrer; M. Aguena; J. Annis; S. Bocquet; D. Brooks; A. Carnero Rosell; M. Carrasco Kind; J. Carretero; R. Cawthon; L. N. Da Costa; T. M. Davis; J. De Vicente; P. Doel; I. Ferrero; D. Friedel; J. Frieman; J. García-Bellido; M. Gatti; E. Gaztanaga; G. Giannini; D. Gruen; R. A. Gruendl; G. Gutierrez; D. L. Hollowood; K. Honscheid; D. J. James; K. Kuehn; N. Kuropatkin; M. A.G. Maia; R. Miquel; A. Palmese; F. Paz-Chinchón; M. E.S. Pereira; A. Pieres; A. A. Plazas Malagón; K. Reil; A. Roodman; E. Sanchez; M. Smith; E. Suchyta; M. E.C. Swanson; G. Tarle; C. To

doi:10.3847/1538-4357/ac5178

DeepZipper: A Novel Deep-learning Architecture for Lensed Supernovae Identification

R. Morgan, B. Nord, K. Bechtol, S. J. González, E. Buckley-Geer, A. Möller, J. W. Park, A. G. Kim, S. Birrer, M. Aguena, J. Annis, S. Bocquet, D. Brooks, A. Carnero Rosell, M. Carrasco Kind, J. Carretero, R. Cawthon, L. N. Da Costa, T. M. Davis, J. De VicenteP. Doel, I. Ferrero, D. Friedel, J. Frieman, J. García-Bellido, M. Gatti, E. Gaztanaga, G. Giannini, D. Gruen, R. A. Gruendl, G. Gutierrez, D. L. Hollowood, K. Honscheid, D. J. James, K. Kuehn, N. Kuropatkin, M. A.G. Maia, R. Miquel, A. Palmese, F. Paz-Chinchón, M. E.S. Pereira, A. Pieres, A. A. Plazas Malagón, K. Reil, A. Roodman, E. Sanchez, M. Smith, E. Suchyta, M. E.C. Swanson, G. Tarle, C. To

Workflow Systems

Research output: Contribution to journal › Article › peer-review

7 Scopus citations

Abstract

Large-scale astronomical surveys have the potential to capture data on large numbers of strongly gravitationally lensed supernovae (LSNe). To facilitate timely analysis and spectroscopic follow-up before the supernova fades, an LSN needs to be identified soon after it begins. To quickly identify LSNe in optical survey data sets, we designed ZipperNet, a multibranch deep neural network that combines convolutional layers (traditionally used for images) with long short-term memory layers (traditionally used for time series). We tested ZipperNet on the task of classifying objects from four categories - no lens, galaxy-galaxy lens, lensed Type-Ia supernova, lensed core-collapse supernova - within high-fidelity simulations of three cosmic survey data sets: the Dark Energy Survey, Rubin Observatory's Legacy Survey of Space and Time (LSST), and a Dark Energy Spectroscopic Instrument (DESI) imaging survey. Among our results, we find that for the LSST-like data set, ZipperNet classifies LSNe with a receiver operating characteristic area under the curve of 0.97, predicts the spectroscopic type of the lensed supernovae with 79% accuracy, and demonstrates similarly high performance for LSNe 1-2 epochs after first detection. We anticipate that a model like ZipperNet, which simultaneously incorporates spatial and temporal information, can play a significant role in the rapid identification of lensed transient systems in cosmic survey experiments.

Original language	English
Article number	109
Journal	Astrophysical Journal
Volume	927
Issue number	1
DOIs	https://doi.org/10.3847/1538-4357/ac5178
State	Published - Mar 1 2022

Funding

R. Morgan thanks the Universities Research Association Fermilab Visiting Scholars Program for funding his work on this project. R. Morgan also thanks the LSSTC Data Science Fellowship Program, which is funded by LSSTC, NSF Cybertraining grant #1829740, the Brinson Foundation, and the Moore Foundation; his participation in the program has benefited this work. This material is based upon work supported by the National Science Foundation Graduate Research Fellowship Program under grant No. 1744555. Any opinions, findings, and conclusions or recommendations expressed in this material are those of the author(s) and do not necessarily reflect the views of the National Science Foundation. Funding for the DES Projects has been provided by the U.S. Department of Energy, the U.S. National Science Foundation, the Ministry of Science and Education of Spain, the Science and Technology Facilities Council of the United Kingdom, the Higher Education Funding Council for England, the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign, the Kavli Institute of Cosmological Physics at the University of Chicago, the Center for Cosmology and Astro-Particle Physics at the Ohio State University, the Mitchell Institute for Fundamental Physics and Astronomy at Texas A&M University, Financiadora de Estudos e Projetos, Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro, Conselho Nacional de Desenvolvimento Científico e Tecnológico and the Ministério da Ciência, Tecnologia e Inovação, the Deutsche Forschungsgemeinschaft and the Collaborating Institutions in the Dark Energy Survey. The DES data management system is supported by the National Science Foundation under grant Numbers AST-1138766 and AST-1536171. The DES participants from Spanish institutions are partially supported by MICINN under grants ESP2017-89838, PGC2018-094773, PGC2018-102021, SEV-2016-0588, SEV-2016-0597, and MDM-2015-0509, some of which include ERDF funds from the European Union. IFAE is partially funded by the CERCA program of the Generalitat de Catalunya. Research leading to these results has received funding from the European Research Council under the European Union’s Seventh Framework Program (FP7/2007-2013) including ERC grant agreements 240672, 291329, and 306478. We acknowledge support from the Brazilian Instituto Nacional de Ciência e Tecnologia (INCT) e-Universe (CNPq grant 465376/2014-2). This manuscript has been authored by Fermi Research Alliance, LLC under Contract No. DE-AC02-07CH11359 with the U.S. Department of Energy, Office of Science, Office of High Energy Physics.

Funders	Funder number
Brazilian Instituto Nacional de Ciência e Tecnologia
Collaborating Institutions in the Dark Energy Survey
Fermi Research Alliance, LLC	DE-AC02-07CH11359
INCT
LSSTC
Mitchell Institute for Fundamental Physics and Astronomy at Texas A&M University
National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign
Science and Technology Facilities Council of the United Kingdom
Universities Research Association
National Science Foundation	AST-1138766, 1829740, AST-1536171, 1744555
U.S. Department of Energy
Brinson Foundation
Blanche Moore Foundation
Office of Science
High Energy Physics
Ohio State University
University of Chicago
Seventh Framework Programme
Higher Education Funding Council for England
Engineering Research Centers	240672, 306478, 291329
European Commission
European Research Council
Deutsche Forschungsgemeinschaft
Generalitat de Catalunya
Ministério da Ciência, Tecnologia e Inovação
Conselho Nacional de Desenvolvimento Científico e Tecnológico	465376/2014-2
Fundação Carlos Chagas Filho de Amparo à Pesquisa do Estado do Rio de Janeiro
Financiadora de Estudos e Projetos
Ministerio de Ciencia e Innovación	SEV-2016-0588, SEV-2016-0597, MDM-2015-0509, PGC2018-094773, PGC2018-102021, ESP2017-89838
Ministry of Education and Science of Ukraine
European Regional Development Fund

Access to Document

10.3847/1538-4357/ac5178

Cite this

Morgan, R., Nord, B., Bechtol, K., González, S. J., Buckley-Geer, E., Möller, A., Park, J. W., Kim, A. G., Birrer, S., Aguena, M., Annis, J., Bocquet, S., Brooks, D., Carnero Rosell, A., Carrasco Kind, M., Carretero, J., Cawthon, R., Da Costa, L. N., Davis, T. M., ... To, C. (2022). DeepZipper: A Novel Deep-learning Architecture for Lensed Supernovae Identification. Astrophysical Journal, 927(1), Article 109. https://doi.org/10.3847/1538-4357/ac5178

@article{cae34ea450c14f76ae03403342713019,

title = "DeepZipper: A Novel Deep-learning Architecture for Lensed Supernovae Identification",

abstract = "Large-scale astronomical surveys have the potential to capture data on large numbers of strongly gravitationally lensed supernovae (LSNe). To facilitate timely analysis and spectroscopic follow-up before the supernova fades, an LSN needs to be identified soon after it begins. To quickly identify LSNe in optical survey data sets, we designed ZipperNet, a multibranch deep neural network that combines convolutional layers (traditionally used for images) with long short-term memory layers (traditionally used for time series). We tested ZipperNet on the task of classifying objects from four categories - no lens, galaxy-galaxy lens, lensed Type-Ia supernova, lensed core-collapse supernova - within high-fidelity simulations of three cosmic survey data sets: the Dark Energy Survey, Rubin Observatory's Legacy Survey of Space and Time (LSST), and a Dark Energy Spectroscopic Instrument (DESI) imaging survey. Among our results, we find that for the LSST-like data set, ZipperNet classifies LSNe with a receiver operating characteristic area under the curve of 0.97, predicts the spectroscopic type of the lensed supernovae with 79% accuracy, and demonstrates similarly high performance for LSNe 1-2 epochs after first detection. We anticipate that a model like ZipperNet, which simultaneously incorporates spatial and temporal information, can play a significant role in the rapid identification of lensed transient systems in cosmic survey experiments.",

author = "R. Morgan and B. Nord and K. Bechtol and Gonz{\'a}lez, {S. J.} and E. Buckley-Geer and A. M{\"o}ller and Park, {J. W.} and Kim, {A. G.} and S. Birrer and M. Aguena and J. Annis and S. Bocquet and D. Brooks and {Carnero Rosell}, A. and {Carrasco Kind}, M. and J. Carretero and R. Cawthon and {Da Costa}, {L. N.} and Davis, {T. M.} and {De Vicente}, J. and P. Doel and I. Ferrero and D. Friedel and J. Frieman and J. Garc{\'i}a-Bellido and M. Gatti and E. Gaztanaga and G. Giannini and D. Gruen and Gruendl, {R. A.} and G. Gutierrez and Hollowood, {D. L.} and K. Honscheid and James, {D. J.} and K. Kuehn and N. Kuropatkin and Maia, {M. A.G.} and R. Miquel and A. Palmese and F. Paz-Chinch{\'o}n and Pereira, {M. E.S.} and A. Pieres and {Plazas Malag{\'o}n}, {A. A.} and K. Reil and A. Roodman and E. Sanchez and M. Smith and E. Suchyta and Swanson, {M. E.C.} and G. Tarle and C. To",

note = "Publisher Copyright: {\textcopyright} 2022. The Author(s). Published by the American Astronomical Society.",

year = "2022",

month = mar,

day = "1",

doi = "10.3847/1538-4357/ac5178",

language = "English",

volume = "927",

journal = "Astrophysical Journal",

issn = "0004-637X",

publisher = "American Astronomical Society",

number = "1",

}

Morgan, R, Nord, B, Bechtol, K, González, SJ, Buckley-Geer, E, Möller, A, Park, JW, Kim, AG, Birrer, S, Aguena, M, Annis, J, Bocquet, S, Brooks, D, Carnero Rosell, A, Carrasco Kind, M, Carretero, J, Cawthon, R, Da Costa, LN, Davis, TM, De Vicente, J, Doel, P, Ferrero, I, Friedel, D, Frieman, J, García-Bellido, J, Gatti, M, Gaztanaga, E, Giannini, G, Gruen, D, Gruendl, RA, Gutierrez, G, Hollowood, DL, Honscheid, K, James, DJ, Kuehn, K, Kuropatkin, N, Maia, MAG, Miquel, R, Palmese, A, Paz-Chinchón, F, Pereira, MES, Pieres, A, Plazas Malagón, AA, Reil, K, Roodman, A, Sanchez, E, Smith, M, Suchyta, E, Swanson, MEC, Tarle, G & To, C 2022, 'DeepZipper: A Novel Deep-learning Architecture for Lensed Supernovae Identification', Astrophysical Journal, vol. 927, no. 1, 109. https://doi.org/10.3847/1538-4357/ac5178

TY - JOUR

T1 - DeepZipper

T2 - A Novel Deep-learning Architecture for Lensed Supernovae Identification

AU - Morgan, R.

AU - Nord, B.

AU - Bechtol, K.

AU - González, S. J.

AU - Buckley-Geer, E.

AU - Möller, A.

AU - Park, J. W.

AU - Kim, A. G.

AU - Birrer, S.

AU - Aguena, M.

AU - Annis, J.

AU - Bocquet, S.

AU - Brooks, D.

AU - Carnero Rosell, A.

AU - Carrasco Kind, M.

AU - Carretero, J.

AU - Cawthon, R.

AU - Da Costa, L. N.

AU - Davis, T. M.

AU - De Vicente, J.

AU - Doel, P.

AU - Ferrero, I.

AU - Friedel, D.

AU - Frieman, J.

AU - García-Bellido, J.

AU - Gatti, M.

AU - Gaztanaga, E.

AU - Giannini, G.

AU - Gruen, D.

AU - Gruendl, R. A.

AU - Gutierrez, G.

AU - Hollowood, D. L.

AU - Honscheid, K.

AU - James, D. J.

AU - Kuehn, K.

AU - Kuropatkin, N.

AU - Maia, M. A.G.

AU - Miquel, R.

AU - Palmese, A.

AU - Paz-Chinchón, F.

AU - Pereira, M. E.S.

AU - Pieres, A.

AU - Plazas Malagón, A. A.

AU - Reil, K.

AU - Roodman, A.

AU - Sanchez, E.

AU - Smith, M.

AU - Suchyta, E.

AU - Swanson, M. E.C.

AU - Tarle, G.

AU - To, C.

PY - 2022/3/1

Y1 - 2022/3/1

N2 - Large-scale astronomical surveys have the potential to capture data on large numbers of strongly gravitationally lensed supernovae (LSNe). To facilitate timely analysis and spectroscopic follow-up before the supernova fades, an LSN needs to be identified soon after it begins. To quickly identify LSNe in optical survey data sets, we designed ZipperNet, a multibranch deep neural network that combines convolutional layers (traditionally used for images) with long short-term memory layers (traditionally used for time series). We tested ZipperNet on the task of classifying objects from four categories - no lens, galaxy-galaxy lens, lensed Type-Ia supernova, lensed core-collapse supernova - within high-fidelity simulations of three cosmic survey data sets: the Dark Energy Survey, Rubin Observatory's Legacy Survey of Space and Time (LSST), and a Dark Energy Spectroscopic Instrument (DESI) imaging survey. Among our results, we find that for the LSST-like data set, ZipperNet classifies LSNe with a receiver operating characteristic area under the curve of 0.97, predicts the spectroscopic type of the lensed supernovae with 79% accuracy, and demonstrates similarly high performance for LSNe 1-2 epochs after first detection. We anticipate that a model like ZipperNet, which simultaneously incorporates spatial and temporal information, can play a significant role in the rapid identification of lensed transient systems in cosmic survey experiments.

AB - Large-scale astronomical surveys have the potential to capture data on large numbers of strongly gravitationally lensed supernovae (LSNe). To facilitate timely analysis and spectroscopic follow-up before the supernova fades, an LSN needs to be identified soon after it begins. To quickly identify LSNe in optical survey data sets, we designed ZipperNet, a multibranch deep neural network that combines convolutional layers (traditionally used for images) with long short-term memory layers (traditionally used for time series). We tested ZipperNet on the task of classifying objects from four categories - no lens, galaxy-galaxy lens, lensed Type-Ia supernova, lensed core-collapse supernova - within high-fidelity simulations of three cosmic survey data sets: the Dark Energy Survey, Rubin Observatory's Legacy Survey of Space and Time (LSST), and a Dark Energy Spectroscopic Instrument (DESI) imaging survey. Among our results, we find that for the LSST-like data set, ZipperNet classifies LSNe with a receiver operating characteristic area under the curve of 0.97, predicts the spectroscopic type of the lensed supernovae with 79% accuracy, and demonstrates similarly high performance for LSNe 1-2 epochs after first detection. We anticipate that a model like ZipperNet, which simultaneously incorporates spatial and temporal information, can play a significant role in the rapid identification of lensed transient systems in cosmic survey experiments.

UR - http://www.scopus.com/inward/record.url?scp=85126624477&partnerID=8YFLogxK

U2 - 10.3847/1538-4357/ac5178

DO - 10.3847/1538-4357/ac5178

M3 - Article

AN - SCOPUS:85126624477

SN - 0004-637X

VL - 927

JO - Astrophysical Journal

JF - Astrophysical Journal

IS - 1

M1 - 109

ER -

DeepZipper: A Novel Deep-learning Architecture for Lensed Supernovae Identification

Abstract

Funding

Access to Document

Other files and links

Fingerprint

Cite this