{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,3,25]],"date-time":"2026-03-25T05:59:55Z","timestamp":1774418395213,"version":"3.50.1"},"reference-count":52,"publisher":"IOP Publishing","issue":"3","license":[{"start":{"date-parts":[[2022,7,7]],"date-time":"2022-07-07T00:00:00Z","timestamp":1657152000000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2022,7,7]],"date-time":"2022-07-07T00:00:00Z","timestamp":1657152000000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"content-domain":{"domain":["iopscience.iop.org"],"crossmark-restriction":false},"short-container-title":["Mach. Learn.: Sci. Technol."],"published-print":{"date-parts":[[2022,9,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Theoretical uncertainty limits our ability to extract cosmological information from baryonic fields such as the thermal Sunyaev\u2013Zel\u2019dovich (tSZ) effect. Being sourced by the electron pressure field, the tSZ effect depends on baryonic physics that is usually modeled by expensive hydrodynamic simulations. We train neural networks on the IllustrisTNG-300 cosmological simulation to predict the continuous electron pressure field in galaxy clusters from gravity-only simulations. Modeling clusters is challenging for neural networks as most of the gas pressure is concentrated in a handful of voxels and even the largest hydrodynamical simulations contain only a few hundred clusters that can be used for training. Instead of conventional convolutional neural net (CNN) architectures, we choose to employ a rotationally equivariant DeepSets architecture to operate directly on the set of dark matter particles. We argue that set-based architectures provide distinct advantages over CNNs. For example, we can enforce exact rotational and permutation equivariance, incorporate existing knowledge on the tSZ field, and work with sparse fields as are standard in cosmology. We compose our architecture with separate, physically meaningful modules, making it amenable to interpretation. For example, we can separately study the influence of local and cluster-scale environment, determine that cluster triaxiality has negligible impact, and train a module that corrects for mis-centering. Our model improves by <jats:inline-formula>\n                     <jats:tex-math\/>\n                     <mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\" overflow=\"scroll\">\n                        <mml:mn>70<\/mml:mn>\n                        <mml:mi mathvariant=\"normal\">%<\/mml:mi>\n                     <\/mml:math>\n                     <jats:inline-graphic xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" xlink:href=\"mlstac78c2ieqn1.gif\" xlink:type=\"simple\"\/>\n                  <\/jats:inline-formula> on analytic profiles fit to the same simulation data. We argue that the electron pressure field, viewed as a function of a gravity-only simulation, has inherent stochasticity, and model this property through a conditional-VAE extension to the network. This modification yields further improvement by <jats:inline-formula>\n                     <jats:tex-math\/>\n                     <mml:math xmlns:mml=\"http:\/\/www.w3.org\/1998\/Math\/MathML\" overflow=\"scroll\">\n                        <mml:mn>7<\/mml:mn>\n                        <mml:mi mathvariant=\"normal\">%<\/mml:mi>\n                     <\/mml:math>\n                     <jats:inline-graphic xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" xlink:href=\"mlstac78c2ieqn2.gif\" xlink:type=\"simple\"\/>\n                  <\/jats:inline-formula>, it is limited by our small training set however. We envision that our method will prove useful in problems beyond the specific one considered here<jats:xref ref-type=\"fn\" rid=\"mlstac78c2fn5\">\n                     <jats:sup>5<\/jats:sup>\n                  <\/jats:xref>\n                  <jats:fn id=\"mlstac78c2fn5\">\n                     <jats:label>5<\/jats:label>\n                     <jats:p>We make our code publicly available at <jats:ext-link xmlns:xlink=\"http:\/\/www.w3.org\/1999\/xlink\" ext-link-type=\"uri\" xlink:href=\"https:\/\/github.com\/leanderthiele\/tSZ_DeepSet\" xlink:type=\"simple\">this URL<\/jats:ext-link>. Data products, trained models, and hyperparameter SQL databases will be shared upon reasonable request.<\/jats:p>\n                  <\/jats:fn>.<\/jats:p>","DOI":"10.1088\/2632-2153\/ac78c2","type":"journal-article","created":{"date-parts":[[2022,6,14]],"date-time":"2022-06-14T22:28:42Z","timestamp":1655245722000},"page":"035002","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":5,"title":["Predicting the thermal Sunyaev\u2013Zel\u2019dovich field using modular and equivariant set-based neural networks"],"prefix":"10.1088","volume":"3","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-2911-9163","authenticated-orcid":true,"given":"Leander","family":"Thiele","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Miles","family":"Cranmer","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"William","family":"Coulton","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Shirley","family":"Ho","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"David N","family":"Spergel","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"266","published-online":{"date-parts":[[2022,7,7]]},"reference":[{"key":"mlstac78c2bib1","author":"Akiba","year":"2019"},{"key":"mlstac78c2bib2","article-title":"Layer normalization","author":"Ba","year":"2016"},{"key":"mlstac78c2bib3","doi-asserted-by":"publisher","first-page":"75","DOI":"10.1088\/0004-637X\/758\/2\/75","volume":"758","author":"Battaglia","year":"2012","journal-title":"Astrophys. J."},{"key":"mlstac78c2bib4","first-page":"pp 4502","author":"Battaglia","year":"2016"},{"key":"mlstac78c2bib5","author":"Battaglia","year":"2018"},{"key":"mlstac78c2bib6","article-title":"Rockstar: phase-space halo finder","author":"Behroozi","year":"2012"},{"key":"mlstac78c2bib7","doi-asserted-by":"publisher","first-page":"109","DOI":"10.1088\/0004-637X\/762\/2\/109","volume":"762","author":"Behroozi","year":"2013","journal-title":"Astrophys. J."},{"key":"mlstac78c2bib8","author":"Bronstein","year":"2021"},{"key":"mlstac78c2bib9","doi-asserted-by":"publisher","first-page":"18","DOI":"10.1109\/MSP.2017.2693418","volume":"34","author":"Bronstein","year":"2017","journal-title":"IEEE Signal Process. Mag."},{"key":"mlstac78c2bib10","doi-asserted-by":"publisher","first-page":"643","DOI":"10.1146\/annurev.astro.40.060401.093803","volume":"40","author":"Carlstrom","year":"2002","journal-title":"Annu. Rev. Astron. Astrophys."},{"key":"mlstac78c2bib11","first-page":"p 10","author":"Cranmer","year":"2021a"},{"key":"mlstac78c2bib12","first-page":"p 1","author":"Cranmer","year":"2021b"},{"key":"mlstac78c2bib13","author":"Cranmer","year":"2020","journal-title":"NeurIPS"},{"key":"mlstac78c2bib14","author":"Esser","year":"2018"},{"key":"mlstac78c2bib15","author":"Fey","year":"2019"},{"key":"mlstac78c2bib16","author":"Gatti","year":"2021"},{"key":"mlstac78c2bib17","doi-asserted-by":"publisher","first-page":"357","DOI":"10.1038\/s41586-020-2649-2","volume":"585","author":"Harris","year":"2020","journal-title":"Nature"},{"key":"mlstac78c2bib18","article-title":"CoRR","author":"Hinton","year":"2012"},{"key":"mlstac78c2bib19","author":"Horowitz","year":"2021"},{"key":"mlstac78c2bib20","doi-asserted-by":"publisher","first-page":"90","DOI":"10.1109\/MCSE.2007.55","volume":"9","author":"Hunter","year":"2007","journal-title":"Comput. Sci. Eng."},{"key":"mlstac78c2bib21","author":"Johnson","year":"2016"},{"key":"mlstac78c2bib22","author":"Kasmanoff","year":"2020"},{"key":"mlstac78c2bib23","author":"Kingma","year":"2015"},{"key":"mlstac78c2bib24","first-page":"pp 87","author":"Kluyver","year":"2016"},{"key":"mlstac78c2bib25","doi-asserted-by":"publisher","DOI":"10.1002\/advs.202100707","volume":"8","author":"Lannelongue","year":"2021","journal-title":"Adv. Sci."},{"key":"mlstac78c2bib26","doi-asserted-by":"publisher","first-page":"5543","DOI":"10.1093\/mnras\/stab1214","volume":"504","author":"Lanusse","year":"2021","journal-title":"Mon. Not. R. Astron. Soc."},{"key":"mlstac78c2bib27","doi-asserted-by":"publisher","first-page":"5113","DOI":"10.1093\/mnras\/sty2206","volume":"480","author":"Marinacci","year":"2018","journal-title":"Mon. Not. R. Astron. Soc."},{"key":"mlstac78c2bib28","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1086\/521328","volume":"668","author":"Nagai","year":"2007","journal-title":"Astrophys. J."},{"key":"mlstac78c2bib29","doi-asserted-by":"publisher","first-page":"1206","DOI":"10.1093\/mnras\/sty618","volume":"477","author":"Naiman","year":"2018","journal-title":"Mon. Not. R. Astron. Soc."},{"key":"mlstac78c2bib30","doi-asserted-by":"publisher","first-page":"493","DOI":"10.1086\/304888","volume":"490","author":"Navarro","year":"1997","journal-title":"Astrophys. J."},{"key":"mlstac78c2bib31","doi-asserted-by":"publisher","first-page":"624","DOI":"10.1093\/mnras\/stx3040","volume":"475","author":"Nelson","year":"2018","journal-title":"Mon. Not. R. Astron. Soc."},{"key":"mlstac78c2bib32","doi-asserted-by":"publisher","first-page":"2","DOI":"10.1186\/s40668-019-0028-x","volume":"6","author":"Nelson","year":"2019","journal-title":"Computat. Astrophys. Cosmol."},{"key":"mlstac78c2bib33","author":"Oladosu","year":"2021"},{"key":"mlstac78c2bib34","first-page":"pp 8024","author":"Paszke","year":"2019"},{"key":"mlstac78c2bib35","doi-asserted-by":"publisher","first-page":"2825","DOI":"10.5555\/1953048.2078195","volume":"12","author":"Pedregosa","year":"2011","journal-title":"J. Mach. Learn. Res."},{"key":"mlstac78c2bib36","doi-asserted-by":"publisher","first-page":"648","DOI":"10.1093\/mnras\/stx3112","volume":"475","author":"Pillepich","year":"2018","journal-title":"Mon. Not. R. Astron. Soc."},{"key":"mlstac78c2bib37","doi-asserted-by":"publisher","first-page":"123","DOI":"10.3847\/1538-3881\/aabc4f","volume":"156","author":"Price-Whelan","year":"2018","journal-title":"Astron. J."},{"key":"mlstac78c2bib38","article-title":"Emulating Sunyaev-Zeldovich images of galaxy clusters using auto-encoders","author":"Rothschild","year":"2021"},{"key":"mlstac78c2bib39","article-title":"CoRR","author":"Smith","year":"2017"},{"key":"mlstac78c2bib40","doi-asserted-by":"publisher","first-page":"676","DOI":"10.1093\/mnras\/stx3304","volume":"475","author":"Springel","year":"2018"},{"key":"mlstac78c2bib41","doi-asserted-by":"publisher","first-page":"537","DOI":"10.1146\/annurev.aa.18.090180.002541","volume":"18","author":"Sunyaev","year":"1980","journal-title":"Annu. Rev. Astron. Astrophys."},{"key":"mlstac78c2bib42","doi-asserted-by":"publisher","first-page":"3","DOI":"10.1007\/BF00653471","volume":"7","author":"Sunyaev","year":"1970","journal-title":"Astrophys. Space Sci."},{"key":"mlstac78c2bib43","doi-asserted-by":"publisher","first-page":"129","DOI":"10.3847\/1538-4357\/abb80f","volume":"902","author":"Thiele","year":"2020","journal-title":"Astrophys. J."},{"key":"mlstac78c2bib44","doi-asserted-by":"publisher","first-page":"L24","DOI":"10.1093\/mnrasl\/slz075","volume":"487","author":"Tr\u00f6ster","year":"2019","journal-title":"Mon. Not. R. Astron. Soc."},{"key":"mlstac78c2bib45","author":"Villaescusa-Navarro","year":"2021"},{"key":"mlstac78c2bib46","author":"Villar","year":"2021"},{"key":"mlstac78c2bib47","doi-asserted-by":"publisher","first-page":"261","DOI":"10.1038\/s41592-019-0686-2","volume":"17","author":"Virtanen","year":"2020","journal-title":"Nat. Methods"},{"key":"mlstac78c2bib48","doi-asserted-by":"publisher","first-page":"42","DOI":"10.3847\/1538-4357\/ac033a","volume":"916","author":"Wadekar","year":"2021","journal-title":"Astrophys. J."},{"key":"mlstac78c2bib49","author":"Yip","year":"2019"},{"key":"mlstac78c2bib50","author":"Zaheer","year":"2017"},{"key":"mlstac78c2bib51","doi-asserted-by":"publisher","first-page":"301","DOI":"10.1007\/BF00661821","volume":"4","author":"Zeldovich","year":"1969","journal-title":"Astrophys. Space Sci."},{"key":"mlstac78c2bib52","author":"Zhang","year":"2019"}],"container-title":["Machine Learning: Science and Technology"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ac78c2","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ac78c2\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ac78c2","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ac78c2\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ac78c2\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ac78c2\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ac78c2\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ac78c2\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,7,7]],"date-time":"2022-07-07T07:10:24Z","timestamp":1657177824000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/ac78c2"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,7,7]]},"references-count":52,"journal-issue":{"issue":"3","published-online":{"date-parts":[[2022,7,7]]},"published-print":{"date-parts":[[2022,9,1]]}},"URL":"https:\/\/doi.org\/10.1088\/2632-2153\/ac78c2","relation":{},"ISSN":["2632-2153"],"issn-type":[{"value":"2632-2153","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,7,7]]},"assertion":[{"value":"Predicting the thermal Sunyaev\u2013Zel\u2019dovich field using modular and equivariant set-based neural networks","name":"article_title","label":"Article Title"},{"value":"Machine Learning: Science and Technology","name":"journal_title","label":"Journal Title"},{"value":"paper","name":"article_type","label":"Article Type"},{"value":"\u00a9 2022 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2022-02-28","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2022-06-14","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2022-07-07","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}