{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,11,3]],"date-time":"2025-11-03T13:42:43Z","timestamp":1762177363807,"version":"3.37.3"},"reference-count":30,"publisher":"IOP Publishing","issue":"2","license":[{"start":{"date-parts":[[2021,3,2]],"date-time":"2021-03-02T00:00:00Z","timestamp":1614643200000},"content-version":"vor","delay-in-days":0,"URL":"http:\/\/creativecommons.org\/licenses\/by\/4.0"},{"start":{"date-parts":[[2021,3,2]],"date-time":"2021-03-02T00:00:00Z","timestamp":1614643200000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/iopscience.iop.org\/info\/page\/text-and-data-mining"}],"funder":[{"name":"Quantum Reservoir Impact AI, LLC"}],"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>Is a deep learning model capable of representing systems governed by certain first principle physics laws by only observing the system\u2019s output? In an effort to simulate two-dimensional subsurface fluid dynamics in porous media, we found that an accurate deep-learning-based proxy model can be taught efficiently by a computationally expensive finite-volume-based simulator. We pose the problem as an image-to-image regression, running the simulator with different input parameters to furnish a synthetic training dataset upon which we fit the deep learning models. Since the data is spatiotemporal, we compare the performance of three alternative treatments of time; a convolutional LSTM, an autoencoder network that treats time as a direct input and an echo state network. Adversarial methods are adopted to address the sharp spatial gradient in the fluid dynamics problem. Compared to traditional simulation, the proposed deep learning approach enables much faster forward computation, which allows us to explore more scenarios with a much larger parameter space given the same time. It is shown that the improved forward computation efficiency is particularly valuable in solving inversion problems, where the physics model has unknown parameters to be determined by history matching. By computing the pixel-level attention of the trained model, we quantify the sensitivity of the deep learning model to key physical parameters and hence demonstrate that the inverse problem can be solved with great acceleration. We assess the efficacy of the machine learning surrogate in terms of its training speed and accuracy. The network can be trained within minutes using limited training data and achieve accuracy that scales desirably with the amount of training data supplied.<\/jats:p>","DOI":"10.1088\/2632-2153\/abd1cf","type":"journal-article","created":{"date-parts":[[2020,12,9]],"date-time":"2020-12-09T06:46:07Z","timestamp":1607496367000},"page":"025022","update-policy":"https:\/\/doi.org\/10.1088\/crossmark-policy","source":"Crossref","is-referenced-by-count":5,"title":["Fast modeling and understanding fluid dynamics systems with encoder\u2013decoder networks"],"prefix":"10.1088","volume":"2","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-0360-602X","authenticated-orcid":false,"given":"Rohan","family":"Thavarajah","sequence":"first","affiliation":[]},{"given":"Xiang","family":"Zhai","sequence":"additional","affiliation":[]},{"given":"Zheren","family":"Ma","sequence":"additional","affiliation":[]},{"given":"David","family":"Castineira","sequence":"additional","affiliation":[]}],"member":"266","published-online":{"date-parts":[[2021,3,2]]},"reference":[{"key":"mlstabd1cfbib1","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1371\/journal.pone.0130140","article-title":"On pixel-wise explanations for non-linear classifier decisions by layer-wise relevance propagation","volume":"10","author":"Bach","year":"2015","journal-title":"PloS One"},{"year":"2015","author":"Badrinarayanan","key":"mlstabd1cfbib2"},{"key":"mlstabd1cfbib3","doi-asserted-by":"publisher","first-page":"32","DOI":"10.1007\/BF01448839","article-title":"\u00dcber die partiellen differenzengleichungen der mathematischen physik","volume":"100","author":"Courant","year":"1928","journal-title":"Math Ann"},{"year":"2009","author":"Deng","key":"mlstabd1cfbib4"},{"year":"2017","author":"Farimani","key":"mlstabd1cfbib5"},{"first-page":"2672","year":"2014","author":"Goodfellow","key":"mlstabd1cfbib6"},{"key":"mlstabd1cfbib7","first-page":"481","article-title":"Convolutional neural networks for steady flow approximation","author":"Guo","year":"2016"},{"key":"mlstabd1cfbib8","article-title":"Deep residual learning for image recognition","author":"He","year":"2015","journal-title":"CoRR"},{"year":"2019","author":"Heim","key":"mlstabd1cfbib9"},{"key":"mlstabd1cfbib10","doi-asserted-by":"publisher","first-page":"1735","DOI":"10.1162\/neco.1997.9.8.1735","article-title":"Long short-term memory","volume":"9","author":"Hochreiter","year":"1997","journal-title":"Neural Comput."},{"year":"2016","author":"Huang","key":"mlstabd1cfbib11"},{"year":"2016","author":"Isola","key":"mlstabd1cfbib12"},{"key":"mlstabd1cfbib13","doi-asserted-by":"publisher","first-page":"199:1","DOI":"10.1145\/2816795.2818129","article-title":"Data-driven fluid simulations using regression forests","volume":"34","author":"Ladick\u00fd","year":"2015","journal-title":"ACM Trans. Graph."},{"key":"mlstabd1cfbib14","first-page":"3431","article-title":"Fully convolutional networks for semantic segmentation","author":"Long","year":"2014"},{"first-page":"4765","year":"2017","author":"Lundberg","key":"mlstabd1cfbib15"},{"year":"2019a","author":"Mo","key":"mlstabd1cfbib16"},{"key":"mlstabd1cfbib17","doi-asserted-by":"publisher","first-page":"703","DOI":"10.1029\/2018WR023528","article-title":"Deep convolutional encoder\u2013decoder networks for uncertainty quantification of dynamic multiphase flow in heterogeneous media","volume":"55","author":"Mo","year":"2019b","journal-title":"Water Resour. Res."},{"key":"mlstabd1cfbib18","doi-asserted-by":"publisher","DOI":"10.1103\/PhysRevLett.120.042003","article-title":"Accelerating science with generative adversarial networks: an application to 3D particle showers in multi-layer calorimeters","volume":"120","author":"Paganini","year":"2018","journal-title":"Phys. Rev. Lett."},{"key":"mlstabd1cfbib19","first-page":"120","article-title":"Model-free prediction of large spatiotemporally chaotic systems from data: a reservoir computing approach","author":"Pathak","year":"2018","journal-title":"Phys. Rev. Lett."},{"year":"2017a","author":"Raissi","key":"mlstabd1cfbib20"},{"year":"2017b","author":"Raissi","key":"mlstabd1cfbib21"},{"first-page":"p 9","year":"2015","author":"Shi","key":"mlstabd1cfbib22"},{"year":"2017","author":"Smilkov","key":"mlstabd1cfbib23"},{"year":"2016","author":"Tompson","key":"mlstabd1cfbib24"},{"first-page":"879","year":"2017","author":"Wang","key":"mlstabd1cfbib25"},{"year":"2018","author":"Wiewel","key":"mlstabd1cfbib26"},{"year":"2016","author":"Zagoruyko","key":"mlstabd1cfbib27"},{"key":"mlstabd1cfbib28","doi-asserted-by":"publisher","first-page":"415","DOI":"10.1016\/j.jcp.2018.04.018","article-title":"Bayesian deep convolutional encoder\u2013decoder networks for surrogate modeling and uncertainty quantification","volume":"366","author":"Zhu","year":"2018","journal-title":"J. Comput. Phys."},{"key":"mlstabd1cfbib29","doi-asserted-by":"publisher","first-page":"56","DOI":"10.1016\/j.jcp.2019.05.024","article-title":"Physics-constrained deep learning for high-dimensional surrogate modeling and uncertainty quantification without labeled data","volume":"394","author":"Zhu","year":"2019","journal-title":"J. Comput. Phys."},{"key":"mlstabd1cfbib30","doi-asserted-by":"publisher","DOI":"10.1063\/1.5022276","article-title":"Observing spatio-temporal dynamics of excitable media using reservoir computing","volume":"28","author":"Zimmermann","year":"2018","journal-title":"Chaos"}],"container-title":["Machine Learning: Science and Technology"],"original-title":[],"link":[{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd1cf","content-type":"text\/html","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd1cf\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd1cf","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd1cf\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd1cf\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd1cf\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"syndication"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd1cf\/pdf","content-type":"application\/pdf","content-version":"am","intended-application":"similarity-checking"},{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd1cf\/pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,2,9]],"date-time":"2022-02-09T02:36:39Z","timestamp":1644374199000},"score":1,"resource":{"primary":{"URL":"https:\/\/iopscience.iop.org\/article\/10.1088\/2632-2153\/abd1cf"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,3,2]]},"references-count":30,"journal-issue":{"issue":"2","published-online":{"date-parts":[[2021,3,2]]},"published-print":{"date-parts":[[2021,6,1]]}},"URL":"https:\/\/doi.org\/10.1088\/2632-2153\/abd1cf","relation":{},"ISSN":["2632-2153"],"issn-type":[{"type":"electronic","value":"2632-2153"}],"subject":[],"published":{"date-parts":[[2021,3,2]]},"assertion":[{"value":"Fast modeling and understanding fluid dynamics systems with encoder\u2013decoder 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 2021 The Author(s). Published by IOP Publishing Ltd","name":"copyright_information","label":"Copyright Information"},{"value":"2020-06-22","name":"date_received","label":"Date Received","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2020-12-08","name":"date_accepted","label":"Date Accepted","group":{"name":"publication_dates","label":"Publication dates"}},{"value":"2021-03-02","name":"date_epub","label":"Online publication date","group":{"name":"publication_dates","label":"Publication dates"}}]}}