Strong Lens Discoveries in DESI Legacy Imaging Surveys DR10 with Two Deep Learning Architectures
Jose Carlos Inchausti, Christopher J. Storfer, Xiaosheng Huang, Yuan-Ming Hsu, Brandt Kaufmann, Chaitanya Pasupala, S. Banka, A. Dey, D. Lang, A. Meisner, J. Moustakas, A.D. Myers, E. F. Schlafly, D.J. Schlegel · 2025
Read Strong Lens Discoveries in DESI Legacy Imaging Surveys DR10 with Two Deep Learning Architectures by Jose Carlos Inchausti, Christopher J. Storfer, Xiaosheng Huang, Yuan-Ming Hsu, Brandt Kaufmann, Chaitanya Pasupala, S. Banka, A. Dey, D. Lang, A. Meisner, J. Moustakas, A.D. Myers, E. F. Schlafly, D.J. Schlegel (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2508-20087.txt.
Introduction
Jose Carlos Inchausti and a team of co-authors report hundreds of new strong gravitational lens candidates. Their twenty twenty-five paper searches the DESI Legacy Imaging Surveys, tenth data release. A strong lens is a galaxy whose gravity bends the light of a farther galaxy into arcs, a rare and prized sight.
Two deep learning networks, a residual network and EfficientNet, were trained on known lenses and non-lenses, and a meta-learner merged their calls. The team then eyeballed only the top sliver of ranked images. They found over eight hundred new candidates across a vast southern sky area. With past papers in the series, the haul nears four thousand.
Catalog record
| CREATOR | Jose Carlos Inchausti, Christopher J. Storfer, Xiaosheng Huang, Yuan-Ming Hsu, Brandt Kaufmann, Chaitanya Pasupala, S. Banka, A. Dey, D. Lang, A. Meisner, J. Moustakas, A.D. Myers, E. F. Schlafly, D.J. Schlegel |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 44 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2508-20087
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Strong Lens Discoveries in DESI Legacy Imaging Surveys DR10 with Two Deep Learning Architectures Jose Carlos Inchausti, Christopher J. Storfer, Xiaosheng Huang, Yuan-Ming Hsu, Brandt Kaufmann, Chaitanya Pasupala, S. Banka, A. Dey, D. Lang, A. Meisner, J. Moustakas, A.D. Myers, E. F. Schlafly, D.J. Schlegel Abstract We have conducted a search for strong gravitational lensing systems in the Dark Energy Spectroscopic Instrument (DESI) Legacy Imaging Surveys Data Release 10 (DR10). This paper is the fourth in a series of searches (following Huang et al. 2020 ; Huang et al. 2021 ; Storfer et al. 2024 , Paper I, II, & III respectively) . This is the first catalog of lens candidates covering nearly the entirety of the extragalactic sky south of declination δ ≈ + 32 ∘ \delta\penalty\ \approx+32^{\circ} , all of it observed by the DECam, covering ∼ \sim 14,000 d e g 2 deg^{2} . We impose a z z -band magnitude cut of < < 20 in AB magnitude. We deploy a Residual Neural Network and EfficientNet as an ensemble trained on a compilation of known lensing systems and high-grade candidates as well as nonlenses in the same footprint. The predictions from these two base models are aggregated using a meta-learner. After applying our ensemble to the survey data, we exclude known candidates and systems, and use our own visual inspection portal to rank images in the top 0.01 percentile of all neural network recommendations. We have found 811 new lens candidates. These include 484 new candidates in the Legacy Surveys DR9 footprint, all parts of which have been searched for strong lenses at least once before, either by our group or others. Combining the discoveries from this work with those from Paper I (335), II (1210), and III (1512), we have discovered a total of 3868 new candidates in the DESI Legacy Surveys. I Introduction Strong gravitational lensing systems are a powerful tool for cosmology. They have been used to study dark matter in galaxies and clusters ( Kochanek 1991 ; Broadhurst et al. 2000 ; Koopmans & Treu 2002 ; Bolton et al. 2006 ; Koopmans et al. 2006 ; Massey et al. 2010 ; Grillo et al. 2015 ; Shu et al. 2017 , e.g.,) , and are uniquely suited to probe substructure in cluster and galaxy scale lenses, as well as line-of-sight low-mass halos and test the predictions of the cold dark matter (CDM) model beyond the local universe ( Vegetti & Koopmans 2009 ; Huang et al. 2009 ; Vegetti et al. 2010 ; Meneghetti et al. 2020 ; Meneghetti et al. 2023 ; Çağan Şengül et al. 2021 ; Wagner-Carena et al. 2022 ; Fagin et al. 2024 , e.g.,) . Searches and spectroscopic follow-up of lensed quasars ( Dawes et al. 2023 ; Sheu et al. 2024 ; He et al. 2023 , e.g.,) have enabled time-delay cosmography, offering an independent route to measure H_{0} . Multiply lensed supernovae (SNe) are also ideal for measuring time delays and H_{0} because of their well-characterized light curves, and in the case of Type Ia, with the added benefit of standardizable luminosity ( Refsdal 1964 ; Treu 2010 ; Oguri & Marshall 2010 , e.g.,) , provided microlensing can be accurately characterized ( Yahalomi et al. 2017 ) . Furthermore, lens models can be constructed after the SNe have faded ( Ding et al. 2021 ) . In the last decade, strongly lensed supernovae, both core-collapse ( Rodney et al. 2016 ; Chen et al. 2022 ) and Type Ia ( Pierel et al. 2023 ; Pierel et al. 2024 ; Pascale et al. 2024 , e.g.,) have been discovered. Sheu et al. 2023 conducted a retrospective search for strongly lensed supernovae in the Dark Energy Spectroscopic Instrument (DESI) Legacy Imaging Surveys ( Dey et al. 2019 ) , and found seven promising candidates. Time-delay H_{0} measurements from multiply imaged supernovae ( Pierel & Rodney 2019 ; Suyu et al. 2020 ; Huber et al. 2021 ; Pascale et al. 2025 , e.g.,) , combined with measurements from distance ladders ( Freedman et al. 2020 ; Riess et al. 2021 , e.g.,) and lensed quasars ( Wong et al. 2019 ; Millon et al. 2020 ; Birrer et al. 2020 , e.g.,) , can be an important test of the tension between H_{0} measured locally and the value inferred from the Cosmic Microwave Background ( Planck Collaboration et al. 2020 , CMB;) . In addition, magnified (but not multiply-lensed) SNe Ia were identified ( Patel et al. 2014 ; Nordin et al. 2014 ; Rubin et al. 2018 ) and used to test the lens models. Finally strong lensing systems can be used to constrain the properties of dark energy ( Jullo et al. 2010 ; Li et al. 2024 , e.g.,) , especially using systems with multiple source planes ( Collett & Auger 2014 ; Linder 2016 ; Sharma & Linder 2022 ; Caminha et al. 2022 , e.g.,) . The introduction of neural networks to identify gravitational lens candidates in imaging surveys has been transformative and continues to show success in identifying lensing candidates ( Jacobs et al. 2017 ; Metcalf et al. 2018 ; Jacobs et al. 2019a ; Jacobs et al. 2019b ; Cañameras et al. 2020 ; Cañameras et al. 2021 ; Rojas et al. 2022 ; Shu et al. 2022 ; Zaborowski et al. 2023 ; González et al. 2025 , e.g.,) . In our recent work, we discovered over 3000 new strong lenses ( Huang et al. 2020 ; Huang et al. 2021 ; Storfer et al. 2024 , hereafter referred to as Papers I, II, and III, respectively) in the DESI Legacy Imaging Surveys ( Dey et al. 2019 ) by using residual neural networks trained on observed images. In this paper, we present results from our fourth search for strong lenses in the Legacy Surveys, now using DR10. We provide an overview of the observations in § II . In § III , we describe the construction of the training sample, our neural network ensemble, and our optimized visual inspection method. New lens candidates are presented in § IV . We discuss our discoveries in § V and conclude in § VI . II Data II.1 The Tractor First, we briefly describe The Tractor ( Lang et al. 2016 ) , a forward modeling algorithm that performs probabilistic astronomical source detection and typing, and constructs the source catalogs for the Legacy Surveys. Source extraction is done on pixel-level data, taking as input the individual images from multiple exposures in multiple bands, with different seeing in each. The Tractor treats the fitting process as a \chi^{2} minimization problem. A detected source is retained if the initial penalized \chi^{2} is improved by 25 1 1 1 For more details, see https://www.legacysurvey.org/dr10/description/ . The Tractor models detected sources as the better of either a point source (“ PSF ”) or round exponential (“ REX ”) galaxy. A detected source can be further classified as de Vaucouleurs ( DEV ; Sérsic index n=4 ) or exponential ( EXP ; n=1 ) profile over REX / PSF if such a fit improves the \chi^{2} by 9. The