Using Machine Learning to Determine Morphologies of z < 1 AGN Host Galaxies in the Hyper Suprime-Cam Wide Survey
We present a machine-learning framework to accurately characterize the morphologies of active galactic nucleus (AGN) host galaxies within z < 1. We first use PSFGAN to decouple host galaxy light from the central point source, then we invoke the Galaxy Morphology Network (G a M or N et ) to estima...
| Main Authors: | , , , , , , , , , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
IOP Publishing
2023-01-01
|
| Series: | The Astrophysical Journal |
| Subjects: | |
| Online Access: | https://doi.org/10.3847/1538-4357/acad79 |
| _version_ | 1827826801787797504 |
|---|---|
| author | Chuan Tian C. Megan Urry Aritra Ghosh Ryan Ofman Tonima Tasnim Ananna Connor Auge Nico Cappelluti Meredith C. Powell David B. Sanders Kevin Schawinski Dominic Stark Grant R. Tremblay |
| author_facet | Chuan Tian C. Megan Urry Aritra Ghosh Ryan Ofman Tonima Tasnim Ananna Connor Auge Nico Cappelluti Meredith C. Powell David B. Sanders Kevin Schawinski Dominic Stark Grant R. Tremblay |
| author_sort | Chuan Tian |
| collection | DOAJ |
| description | We present a machine-learning framework to accurately characterize the morphologies of active galactic nucleus (AGN) host galaxies within z < 1. We first use PSFGAN to decouple host galaxy light from the central point source, then we invoke the Galaxy Morphology Network (G a M or N et ) to estimate whether the host galaxy is disk-dominated, bulge-dominated, or indeterminate. Using optical images from five bands of the HSC Wide Survey, we build models independently in three redshift bins: low (0 < z < 0.25), mid (0.25 < z < 0.5), and high (0.5 < z < 1.0). By first training on a large number of simulated galaxies, then fine-tuning using far fewer classified real galaxies, our framework predicts the actual morphology for ∼60%–70% of the host galaxies from test sets, with a classification precision of ∼80%–95%, depending on the redshift bin. Specifically, our models achieve a disk precision of 96%/82%/79% and bulge precision of 90%/90%/80% (for the three redshift bins) at thresholds corresponding to indeterminate fractions of 30%/43%/42%. The classification precision of our models has a noticeable dependency on host galaxy radius and magnitude. No strong dependency is observed on contrast ratio. Comparing classifications of real AGNs, our models agree well with traditional 2D fitting with GALFIT. The PSFGAN+G a M or N et framework does not depend on the choice of fitting functions or galaxy-related input parameters, runs orders of magnitude faster than GALFIT, and is easily generalizable via transfer learning, making it an ideal tool for studying AGN host galaxy morphology in forthcoming large imaging surveys. |
| first_indexed | 2024-03-12T03:17:33Z |
| format | Article |
| id | doaj.art-ef3d62ee02934c349e218cfcb0c03c95 |
| institution | Directory Open Access Journal |
| issn | 1538-4357 |
| language | English |
| last_indexed | 2024-03-12T03:17:33Z |
| publishDate | 2023-01-01 |
| publisher | IOP Publishing |
| record_format | Article |
| series | The Astrophysical Journal |
| spelling | doaj.art-ef3d62ee02934c349e218cfcb0c03c952023-09-03T14:08:25ZengIOP PublishingThe Astrophysical Journal1538-43572023-01-01944212410.3847/1538-4357/acad79Using Machine Learning to Determine Morphologies of z < 1 AGN Host Galaxies in the Hyper Suprime-Cam Wide SurveyChuan Tian0https://orcid.org/0000-0003-4056-7071C. Megan Urry1https://orcid.org/0000-0002-0745-9792Aritra Ghosh2https://orcid.org/0000-0002-2525-9647Ryan Ofman3Tonima Tasnim Ananna4https://orcid.org/0000-0001-8211-3807Connor Auge5https://orcid.org/0000-0002-5504-8752Nico Cappelluti6https://orcid.org/0000-0002-1697-186XMeredith C. Powell7https://orcid.org/0000-0003-2284-8603David B. Sanders8https://orcid.org/0000-0002-1233-9998Kevin Schawinski9https://orcid.org/0000-0001-5464-0888Dominic Stark10Grant R. Tremblay11https://orcid.org/0000-0002-5445-5401Department of Physics, Yale University , New Haven, CT, USA ; chuan.tian@yale.edu, ufsccosmos@gmail.comDepartment of Physics, Yale University , New Haven, CT, USA ; chuan.tian@yale.edu, ufsccosmos@gmail.com; Department of Astronomy, Yale University , New Haven, CT, USA; Yale Center for Astronomy and Astrophysics, Yale University , New Haven, CT, USADepartment of Astronomy, Yale University , New Haven, CT, USA; Yale Center for Astronomy and Astrophysics, Yale University , New Haven, CT, USADepartment of Astronomy, Yale University , New Haven, CT, USADepartment of Physics and Astronomy, Dartmouth College , 6127 Wilder Laboratory, Hanover, NH, USAInstitute for Astronomy, University of Hawaii , Honolulu, HI, USADepartment of Physics, University of Miami , Coral Gables, FL, USA; INAF—Osservatorio di Astrofisica e Scienza dello Spazio di Bologna , Bologna, ItalyKavli Institute for Particle Astrophysics and Cosmology, Stanford University , Stanford, CA, USAInstitute for Astronomy, University of Hawaii , Honolulu, HI, USAModulos AG , Technoparkstr. 1, CH-8005, Zurich, SwitzerlandModulos AG , Technoparkstr. 1, CH-8005, Zurich, SwitzerlandCenter for Astrophysics ∣ Harvard & Smithsonian , 60 Garden St., Cambridge, MA 02138, USAWe present a machine-learning framework to accurately characterize the morphologies of active galactic nucleus (AGN) host galaxies within z < 1. We first use PSFGAN to decouple host galaxy light from the central point source, then we invoke the Galaxy Morphology Network (G a M or N et ) to estimate whether the host galaxy is disk-dominated, bulge-dominated, or indeterminate. Using optical images from five bands of the HSC Wide Survey, we build models independently in three redshift bins: low (0 < z < 0.25), mid (0.25 < z < 0.5), and high (0.5 < z < 1.0). By first training on a large number of simulated galaxies, then fine-tuning using far fewer classified real galaxies, our framework predicts the actual morphology for ∼60%–70% of the host galaxies from test sets, with a classification precision of ∼80%–95%, depending on the redshift bin. Specifically, our models achieve a disk precision of 96%/82%/79% and bulge precision of 90%/90%/80% (for the three redshift bins) at thresholds corresponding to indeterminate fractions of 30%/43%/42%. The classification precision of our models has a noticeable dependency on host galaxy radius and magnitude. No strong dependency is observed on contrast ratio. Comparing classifications of real AGNs, our models agree well with traditional 2D fitting with GALFIT. The PSFGAN+G a M or N et framework does not depend on the choice of fitting functions or galaxy-related input parameters, runs orders of magnitude faster than GALFIT, and is easily generalizable via transfer learning, making it an ideal tool for studying AGN host galaxy morphology in forthcoming large imaging surveys.https://doi.org/10.3847/1538-4357/acad79Active galactic nucleiGalaxiesGalaxy classification systemsAstronomy data analysisNeural networksConvolutional neural networks |
| spellingShingle | Chuan Tian C. Megan Urry Aritra Ghosh Ryan Ofman Tonima Tasnim Ananna Connor Auge Nico Cappelluti Meredith C. Powell David B. Sanders Kevin Schawinski Dominic Stark Grant R. Tremblay Using Machine Learning to Determine Morphologies of z < 1 AGN Host Galaxies in the Hyper Suprime-Cam Wide Survey The Astrophysical Journal Active galactic nuclei Galaxies Galaxy classification systems Astronomy data analysis Neural networks Convolutional neural networks |
| title | Using Machine Learning to Determine Morphologies of z < 1 AGN Host Galaxies in the Hyper Suprime-Cam Wide Survey |
| title_full | Using Machine Learning to Determine Morphologies of z < 1 AGN Host Galaxies in the Hyper Suprime-Cam Wide Survey |
| title_fullStr | Using Machine Learning to Determine Morphologies of z < 1 AGN Host Galaxies in the Hyper Suprime-Cam Wide Survey |
| title_full_unstemmed | Using Machine Learning to Determine Morphologies of z < 1 AGN Host Galaxies in the Hyper Suprime-Cam Wide Survey |
| title_short | Using Machine Learning to Determine Morphologies of z < 1 AGN Host Galaxies in the Hyper Suprime-Cam Wide Survey |
| title_sort | using machine learning to determine morphologies of z 1 agn host galaxies in the hyper suprime cam wide survey |
| topic | Active galactic nuclei Galaxies Galaxy classification systems Astronomy data analysis Neural networks Convolutional neural networks |
| url | https://doi.org/10.3847/1538-4357/acad79 |
| work_keys_str_mv | AT chuantian usingmachinelearningtodeterminemorphologiesofz1agnhostgalaxiesinthehypersuprimecamwidesurvey AT cmeganurry usingmachinelearningtodeterminemorphologiesofz1agnhostgalaxiesinthehypersuprimecamwidesurvey AT aritraghosh usingmachinelearningtodeterminemorphologiesofz1agnhostgalaxiesinthehypersuprimecamwidesurvey AT ryanofman usingmachinelearningtodeterminemorphologiesofz1agnhostgalaxiesinthehypersuprimecamwidesurvey AT tonimatasnimananna usingmachinelearningtodeterminemorphologiesofz1agnhostgalaxiesinthehypersuprimecamwidesurvey AT connorauge usingmachinelearningtodeterminemorphologiesofz1agnhostgalaxiesinthehypersuprimecamwidesurvey AT nicocappelluti usingmachinelearningtodeterminemorphologiesofz1agnhostgalaxiesinthehypersuprimecamwidesurvey AT meredithcpowell usingmachinelearningtodeterminemorphologiesofz1agnhostgalaxiesinthehypersuprimecamwidesurvey AT davidbsanders usingmachinelearningtodeterminemorphologiesofz1agnhostgalaxiesinthehypersuprimecamwidesurvey AT kevinschawinski usingmachinelearningtodeterminemorphologiesofz1agnhostgalaxiesinthehypersuprimecamwidesurvey AT dominicstark usingmachinelearningtodeterminemorphologiesofz1agnhostgalaxiesinthehypersuprimecamwidesurvey AT grantrtremblay usingmachinelearningtodeterminemorphologiesofz1agnhostgalaxiesinthehypersuprimecamwidesurvey |