Domain Adaptation in Application to Gravitational Lens Finding

The next decade is expected to see a tenfold increase in the number of strong gravitational lenses, driven by new wide-field imaging surveys. To discover these rare objects, efficient automated detection methods need to be developed. In this work, we assess the performance of three domain adaptation...

Full description

Saved in:
Bibliographic Details
Main Authors: Hanna Parul, Sergei Gleyzer, Pranath Reddy, Michael W. Toomey
Format: Article
Language:English
Published: IOP Publishing 2025-01-01
Series:The Astrophysical Journal
Subjects:
Online Access:https://doi.org/10.3847/1538-4357/adee16
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849225203181158400
author Hanna Parul
Sergei Gleyzer
Pranath Reddy
Michael W. Toomey
author_facet Hanna Parul
Sergei Gleyzer
Pranath Reddy
Michael W. Toomey
author_sort Hanna Parul
collection DOAJ
description The next decade is expected to see a tenfold increase in the number of strong gravitational lenses, driven by new wide-field imaging surveys. To discover these rare objects, efficient automated detection methods need to be developed. In this work, we assess the performance of three domain adaptation (DA) techniques—adversarial discriminative DA, Wasserstein distance guided representation learning (WDGRL), and supervised domain adaptation (SDA)—in enhancing lens-finding algorithms trained on simulated data when applied to observations from the Hyper Suprime-Cam Subaru Strategic Program. We find that WDGRL combined with an equivariant-neural-network-based encoder provides the best performance in an unsupervised setting and that SDA is able to enhance the model’s ability to distinguish between lenses and common similar-looking false positives, such as spiral galaxies, which is crucial for future lens surveys.
format Article
id doaj-art-b83b65bfd416471eb4dce5a22762fec1
institution Kabale University
issn 1538-4357
language English
publishDate 2025-01-01
publisher IOP Publishing
record_format Article
series The Astrophysical Journal
spelling doaj-art-b83b65bfd416471eb4dce5a22762fec12025-08-25T05:51:24ZengIOP PublishingThe Astrophysical Journal1538-43572025-01-0199014710.3847/1538-4357/adee16Domain Adaptation in Application to Gravitational Lens FindingHanna Parul0https://orcid.org/0009-0007-3431-4269Sergei Gleyzer1https://orcid.org/0000-0002-6222-8102Pranath Reddy2https://orcid.org/0000-0002-6609-3495Michael W. Toomey3https://orcid.org/0000-0003-1205-4033LIRA, Observatoire de Paris, Université PSL, Sorbonne Université, Université Paris Cité , CY Cergy Paris Université, CNRS, 92190 Meudon, France ; hparul@crimson.ua.edu; Department of Physics & Astronomy, University of Alabama, Tuscaloosa, AL 35401, USADepartment of Physics & Astronomy, University of Alabama, Tuscaloosa, AL 35401, USAUniversity of Florida , Gainesville, FL 32611, USACenter for Theoretical Physics – a Leinweber Institute, Massachusetts Institute of Technology , Cambridge, MA 02139, USAThe next decade is expected to see a tenfold increase in the number of strong gravitational lenses, driven by new wide-field imaging surveys. To discover these rare objects, efficient automated detection methods need to be developed. In this work, we assess the performance of three domain adaptation (DA) techniques—adversarial discriminative DA, Wasserstein distance guided representation learning (WDGRL), and supervised domain adaptation (SDA)—in enhancing lens-finding algorithms trained on simulated data when applied to observations from the Hyper Suprime-Cam Subaru Strategic Program. We find that WDGRL combined with an equivariant-neural-network-based encoder provides the best performance in an unsupervised setting and that SDA is able to enhance the model’s ability to distinguish between lenses and common similar-looking false positives, such as spiral galaxies, which is crucial for future lens surveys.https://doi.org/10.3847/1538-4357/adee16Strong gravitational lensingAstronomy data analysis
spellingShingle Hanna Parul
Sergei Gleyzer
Pranath Reddy
Michael W. Toomey
Domain Adaptation in Application to Gravitational Lens Finding
The Astrophysical Journal
Strong gravitational lensing
Astronomy data analysis
title Domain Adaptation in Application to Gravitational Lens Finding
title_full Domain Adaptation in Application to Gravitational Lens Finding
title_fullStr Domain Adaptation in Application to Gravitational Lens Finding
title_full_unstemmed Domain Adaptation in Application to Gravitational Lens Finding
title_short Domain Adaptation in Application to Gravitational Lens Finding
title_sort domain adaptation in application to gravitational lens finding
topic Strong gravitational lensing
Astronomy data analysis
url https://doi.org/10.3847/1538-4357/adee16
work_keys_str_mv AT hannaparul domainadaptationinapplicationtogravitationallensfinding
AT sergeigleyzer domainadaptationinapplicationtogravitationallensfinding
AT pranathreddy domainadaptationinapplicationtogravitationallensfinding
AT michaelwtoomey domainadaptationinapplicationtogravitationallensfinding