{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,2,21]],"date-time":"2025-02-21T23:58:44Z","timestamp":1740182324074,"version":"3.37.3"},"reference-count":50,"publisher":"IOP Publishing","issue":"2","license":[{"start":{"date-parts":[[2021,2,26]],"date-time":"2021-02-26T00:00:00Z","timestamp":1614297600000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2021,2,26]],"date-time":"2021-02-26T00:00:00Z","timestamp":1614297600000},"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":[[2021,6,1]]},"abstract":"<jats:title>Abstract<\/jats:title>\n               <jats:p>Lagrangian methods to solve the inviscid Euler equations produce numerical oscillations near shock waves. A common approach to reducing these oscillations is to add artificial viscosity (AV) to the discrete equations. The AV term acts as a dissipative mechanism that attenuates oscillations by smearing the shock across a finite number of computational cells. However, AV introduces several control parameters that are not determined by the underlying physical model, and hence, in practice are tuned to the characteristics of a given problem. We seek to improve the standard quadratic-linear AV form by replacing it with a learned neural function that reduces oscillations relative to exact solutions of the Euler equations, resulting in a hybrid numerical-neural hydrodynamic solver. Because AV is an artificial construct that exists solely to improve the numerical properties of a hydrodynamic code, there is no offline \u2018viscosity data\u2019 against which a neural network can be trained before inserting into a numerical simulation, thus requiring online training. We achieve this via differentiable programming, i.e. end-to-end backpropagation or adjoint solution through both the neural and differential equation code, using automatic differentiation of the hybrid code in the Julia programming language to calculate the necessary loss function gradients. A novel offline pre-training step accelerates training by initializing the neural network to the default numerical AV scheme, which can be learned rapidly by space-filling sampling over the AV input space. We find that online training over early time steps of simulation is sufficient to learn a neural AV function that reduces numerical oscillations in long-term hydrodynamic shock simulations. These results offer an early proof-of-principle that online differentiable training of hybrid numerical schemes with novel neural network components can improve certain performance aspects existing in purely numerical schemes.<\/jats:p>","DOI":"10.1088\/2632-2153\/abd644","type":"journal-article","created":{"date-parts":[[2020,12,23]],"date-time":"2020-12-23T22:17:59Z","timestamp":1608761879000},"page":"025015","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":0,"title":["Differentiable programming for online training of a neural artificial viscosity function within a staggered grid Lagrangian hydrodynamics scheme"],"prefix":"10.1088","volume":"2","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-1980-9072","authenticated-orcid":false,"given":"Pake","family":"Melland","sequence":"first","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-4099-8990","authenticated-orcid":false,"given":"Jason","family":"Albright","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-2264-3512","authenticated-orcid":false,"given":"Nathan M","family":"Urban","sequence":"additional","affiliation":[],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"266","published-online":{"date-parts":[[2021,2,26]]},"reference":[{"volume":"vol 132","year":"1992","author":"LeVeque","key":"mlstabd644bib1"},{"year":"2013","author":"Toro","key":"mlstabd644bib2"},{"key":"mlstabd644bib3","doi-asserted-by":"publisher","first-page":"232","DOI":"10.1063\/1.1699639","volume":"21","author":"VonNeumann","year":"1950","journal-title":"J. Appl. Phys."},{"key":"mlstabd644bib4","doi-asserted-by":"publisher","first-page":"281","DOI":"10.1016\/0021-9991(80)90161-8","volume":"36","author":"Wilkins","year":"1980","journal-title":"J. Comput. Phys."},{"key":"mlstabd644bib5","doi-asserted-by":"publisher","first-page":"227","DOI":"10.1006\/jcph.1998.6029","volume":"146","author":"Caramana","year":"1998","journal-title":"J. Comput. Phys."},{"key":"mlstabd644bib6","doi-asserted-by":"publisher","first-page":"78","DOI":"10.1016\/0021-9991(87)90074-X","volume":"72","author":"Noh","year":"1987","journal-title":"J. Comput. Phys."},{"article-title":"Machine learning-based optimization strategies for artificial viscosity part II","year":"2019","author":"Urban","key":"mlstabd644bib7"},{"year":"2020","author":"Rackauckas","key":"mlstabd644bib8"},{"key":"mlstabd644bib9","first-page":"3424","article-title":"Accelerating Eulerian fluid simulation with convolutional networks","author":"Tompson","year":"2017"},{"key":"mlstabd644bib10","first-page":"59","volume":"38","author":"Kim","year":"2019","journal-title":"Computer Graphics Forum"},{"key":"mlstabd644bib11","doi-asserted-by":"publisher","first-page":"1339","DOI":"10.1016\/j.jcp.2018.08.029","volume":"375","author":"Sirignano","year":"2018","journal-title":"J. Comput. Phys."},{"key":"mlstabd644bib12","doi-asserted-by":"publisher","first-page":"8505","DOI":"10.1073\/pnas.1718942115","volume":"115","author":"Han","year":"2018","journal-title":"Proc. Natl Acad. Sci."},{"year":"2017","author":"de Bezenac","key":"mlstabd644bib13"},{"key":"mlstabd644bib14","doi-asserted-by":"publisher","first-page":"56","DOI":"10.1016\/j.jcp.2019.05.024","volume":"394","author":"Zhu","year":"2019","journal-title":"J. Comput. Phys."},{"key":"mlstabd644bib15","doi-asserted-by":"publisher","first-page":"15344","DOI":"10.1073\/pnas.1814058116","volume":"116","author":"Bar-Sinai","year":"2019","journal-title":"Proc. Natl Acad. Sci."},{"year":"2018","author":"Mishra","key":"mlstabd644bib16"},{"key":"mlstabd644bib17","doi-asserted-by":"publisher","first-page":"122","DOI":"10.1017\/jfm.2018.770","volume":"858","author":"Maulik","year":"2019","journal-title":"J. Fluid Mech."},{"article-title":"Embedding hard physical constraints in convolutional neural networks for 3d turbulence","year":"2020","author":"Mohan","key":"mlstabd644bib18"},{"key":"mlstabd644bib19","doi-asserted-by":"publisher","first-page":"9684","DOI":"10.1073\/pnas.1810286115","volume":"115","author":"Rasp","year":"2018","journal-title":"Proc. Natl Acad. Sci."},{"key":"mlstabd644bib20","doi-asserted-by":"publisher","first-page":"6289","DOI":"10.1029\/2018GL078510","volume":"45","author":"Brenowitz","year":"2018","journal-title":"Geophys. Res. Lett."},{"key":"mlstabd644bib21","doi-asserted-by":"publisher","first-page":"2548","DOI":"10.1029\/2018MS001351","volume":"10","author":"O\u2019Gorman","year":"2018","journal-title":"J. Adv. Model. Earth Syst."},{"key":"mlstabd644bib22","doi-asserted-by":"publisher","first-page":"363","DOI":"10.1029\/94RG01872","volume":"32","author":"Large","year":"1994","journal-title":"Rev. Geophys."},{"key":"mlstabd644bib23","doi-asserted-by":"publisher","first-page":"589","DOI":"10.1175\/BAMS-D-15-00135.1","volume":"98","author":"Hourdin","year":"2017","journal-title":"Bull. Am. Meteorol. Soc."},{"key":"mlstabd644bib24","doi-asserted-by":"publisher","first-page":"3207","DOI":"10.5194\/gmd-10-3207-2017","volume":"10","author":"Schmidt","year":"2017","journal-title":"Geosci. Model Dev."},{"volume":"vol 1","year":"2003","author":"Santner","key":"mlstabd644bib25"},{"key":"mlstabd644bib26","doi-asserted-by":"publisher","first-page":"681","DOI":"10.1007\/s11222-011-9242-3","volume":"22","author":"Pronzato","year":"2012","journal-title":"Stat. Comput."},{"key":"mlstabd644bib27","doi-asserted-by":"publisher","first-page":"533","DOI":"10.1038\/323533a0","volume":"323","author":"Rumelhart","year":"1986","journal-title":"Nature"},{"first-page":"65","year":"1992","author":"Hecht-Nielsen","key":"mlstabd644bib28"},{"key":"mlstabd644bib29","doi-asserted-by":"publisher","first-page":"265","DOI":"10.1007\/s11222-011-9242-3","article-title":"Tensorflow: a system for large-scale machine learning","author":"Abadi","year":"2016"},{"key":"mlstabd644bib30","first-page":"8026","article-title":"Pytorch: an imperative style, high-performance deep learning library","author":"Paszke","year":"2019"},{"key":"mlstabd644bib31","first-page":"1","volume":"18","author":"Baydin","year":"2018","journal-title":"J. Mach. Learn. Res."},{"key":"mlstabd644bib32","doi-asserted-by":"crossref","DOI":"10.2172\/6683420","article-title":"Gress (gradient enhanced software system) version 0. 0 user\u2019s manual","author":"Horwedel","year":"1988"},{"key":"mlstabd644bib33","first-page":"90","article-title":"PADRE2, version 1\u2014user\u2019s manual","author":"Kubota","year":"1990","journal-title":"Research memorandum RMI"},{"year":"2019","author":"Innes","key":"mlstabd644bib34"},{"key":"mlstabd644bib35","first-page":"10180","article-title":"Backpropagation with callbacks: foundations for efficient and expressive differentiable programming","volume":"31","author":"Wang","year":"2018"},{"key":"mlstabd644bib36","doi-asserted-by":"publisher","first-page":"602","DOI":"10.21105\/joss.00602","volume":"3","author":"Innes","year":"2018","journal-title":"J. Open Source Softw."},{"year":"2018","author":"Innes","key":"mlstabd644bib37"},{"key":"mlstabd644bib38","doi-asserted-by":"publisher","first-page":"249","DOI":"10.1023\/A:1011455900396","volume":"65","author":"Gunzburger","year":"2000","journal-title":"Flow Turbul. Combust."},{"year":"2016","author":"Asch","key":"mlstabd644bib39"},{"key":"mlstabd644bib40","doi-asserted-by":"publisher","first-page":"785","DOI":"10.1017\/S0022112077001463","volume":"83","author":"Sod","year":"1977","journal-title":"J. Fluid Mech."},{"key":"mlstabd644bib41","doi-asserted-by":"publisher","DOI":"10.1016\/j.compfluid.2020.104580","author":"Albright","year":"2020","journal-title":"Comput. Fluids"},{"key":"mlstabd644bib42","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1016\/0021-9991(78)90023-2","volume":"27","author":"Sod","year":"1978","journal-title":"J. Comput. Phys."},{"key":"mlstabd644bib43","doi-asserted-by":"publisher","first-page":"70","DOI":"10.1006\/jcph.1998.5989","volume":"144","author":"Caramana","year":"1998","journal-title":"J. Comput. Phys."},{"key":"mlstabd644bib44","doi-asserted-by":"publisher","first-page":"303","DOI":"10.1007\/BF02551274","volume":"2","author":"Cybenko","year":"1989","journal-title":"Math. Control Signals Syst."},{"year":"2019","author":"Strang","key":"mlstabd644bib45"},{"article-title":"Validation test case suite for compressible hydrodynamics computation","year":"2005","author":"Loub\u00e9re","key":"mlstabd644bib46"},{"year":"2014","author":"Kingma","key":"mlstabd644bib47"},{"key":"mlstabd644bib48","doi-asserted-by":"publisher","first-page":"65","DOI":"10.1137\/141000671","volume":"59","author":"Bezanson","year":"2017","journal-title":"SIAM Rev."},{"year":"2018","author":"Rackauckas","key":"mlstabd644bib49"},{"year":"2019","author":"Rackauckas","key":"mlstabd644bib50"}],"container-title":["Machine Learning: Science and Technology"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd644","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd644\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd644","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd644\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd644\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd644\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd644\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd644\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,2,12]],"date-time":"2022-02-12T00:24:53Z","timestamp":1644625493000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd644"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,2,26]]},"references-count":50,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2021,2,26]]},"published-print":{"date-parts":[[2021,6,1]]}},"URL":"https:\/\/doi.org\/10.1088\/2632-2153\/abd644","relation":{},"ISSN":["2632-2153"],"issn-type":[{"type":"electronic","value":"2632-2153"}],"subject":[],"published":{"date-parts":[[2021,2,26]]},"assertion":[{"value":"Differentiable programming for online training of a neural artificial viscosity function within a staggered grid Lagrangian hydrodynamics scheme","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 2021 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2020-08-27","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2020-12-23","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2021-02-26","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}