{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,5,12]],"date-time":"2026-05-12T12:37:32Z","timestamp":1778589452386,"version":"3.51.4"},"reference-count":55,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"12","license":[{"start":{"date-parts":[[2022,12,1]],"date-time":"2022-12-01T00:00:00Z","timestamp":1669852800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2022,12,1]],"date-time":"2022-12-01T00:00:00Z","timestamp":1669852800000},"content-version":"am","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2022,12,1]],"date-time":"2022-12-01T00:00:00Z","timestamp":1669852800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2022,12,1]],"date-time":"2022-12-01T00:00:00Z","timestamp":1669852800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/100000015","name":"U.S. Department of Energy","doi-asserted-by":"publisher","id":[{"id":"10.13039\/100000015","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Office of Science, Advanced Scientific Computing Research","award":["DE-AC02-06CH11357"],"award-info":[{"award-number":["DE-AC02-06CH11357"]}]},{"name":"Exascale Computing Project","award":["17-SC-20-SC"],"award-info":[{"award-number":["17-SC-20-SC"]}]},{"DOI":"10.13039\/501100008982","name":"National Science Foundation","doi-asserted-by":"publisher","award":["OAC-1931354"],"award-info":[{"award-number":["OAC-1931354"]}],"id":[{"id":"10.13039\/501100008982","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100008982","name":"National Science Foundation","doi-asserted-by":"publisher","award":["OAC-1818253"],"award-info":[{"award-number":["OAC-1818253"]}],"id":[{"id":"10.13039\/501100008982","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100008982","name":"National Science Foundation","doi-asserted-by":"publisher","award":["OAC-2106661"],"award-info":[{"award-number":["OAC-2106661"]}],"id":[{"id":"10.13039\/501100008982","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100008982","name":"National Science Foundation","doi-asserted-by":"publisher","award":["OAC-2107511"],"award-info":[{"award-number":["OAC-2107511"]}],"id":[{"id":"10.13039\/501100008982","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Parallel Distrib. Syst."],"published-print":{"date-parts":[[2022,12,1]]},"DOI":"10.1109\/tpds.2022.3161187","type":"journal-article","created":{"date-parts":[[2022,3,22]],"date-time":"2022-03-22T19:37:18Z","timestamp":1647977838000},"page":"3616-3627","source":"Crossref","is-referenced-by-count":4,"title":["Deep Neural Network Training With Distributed K-FAC"],"prefix":"10.1109","volume":"33","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-6547-6902","authenticated-orcid":false,"given":"J. Gregory","family":"Pauloski","sequence":"first","affiliation":[{"name":"Department of Computer Science, University of Chicago, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Lei","family":"Huang","sequence":"additional","affiliation":[{"name":"Texas Advanced Computing Center, Austin, TX, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Weijia","family":"Xu","sequence":"additional","affiliation":[{"name":"Texas Advanced Computing Center, Austin, TX, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7370-4805","authenticated-orcid":false,"given":"Kyle","family":"Chard","sequence":"additional","affiliation":[{"name":"Department of Computer Science, University of Chicago, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-2129-5269","authenticated-orcid":false,"given":"Ian T.","family":"Foster","sequence":"additional","affiliation":[{"name":"Department of Computer Science, University of Chicago, Chicago, IL, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-5921-0035","authenticated-orcid":false,"given":"Zhao","family":"Zhang","sequence":"additional","affiliation":[{"name":"Texas Advanced Computing Center, Austin, TX, USA"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref39","first-page":"1","article-title":"On large-batch training for deep learning: Generalization gap and sharp minima","author":"keskar","year":"2017","journal-title":"Proc 5th Int Conf Learn Representations"},{"key":"ref38","article-title":"MXNet: A flexible and efficient machine learning library for heterogeneous distributed systems","author":"chen","year":"2015"},{"key":"ref33","article-title":"How to scale distributed deep learning?","author":"jin","year":"2016"},{"key":"ref32","first-page":"685","article-title":"Deep learning with elastic averaging SGD","author":"zhang","year":"2015","journal-title":"Proc Adv Neural Informat Process Syst"},{"key":"ref31","first-page":"693","article-title":"HOGWILD: A lock-free approach to parallelizing stochastic gradient descent","author":"recht","year":"2011","journal-title":"Proc Adv Neural Informat Process Syst"},{"key":"ref30","article-title":"Large batch training of convolutional networks","author":"you","year":"2017"},{"key":"ref37","article-title":"A Kronecker-factored approximate Fisher matrix for convolution layers","author":"grosse","year":"2016"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1145\/3212734.3212763"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1145\/2640087.2644155"},{"key":"ref34","doi-asserted-by":"publisher","DOI":"10.1109\/ALLERTON.2016.7852343"},{"key":"ref28","first-page":"265","article-title":"TensorFlow: A system for large-scale machine learning","author":"abadi","year":"2016","journal-title":"Proc 12th USENIX Symp Oper Syst Des Implementation"},{"key":"ref27","article-title":"Horovod: Fast and easy distributed deep learning in TensorFlow","author":"sergeev","year":"2018"},{"key":"ref29","year":"0"},{"key":"ref2","article-title":"Scale out for large minibatch SGD: Residual network training on ImageNet-1K with improved accuracy and reduced time to train","author":"codreanu","year":"2017"},{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1145\/3225058.3225069"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.322"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1145\/3458817.3476152"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1109\/SC41405.2020.00098"},{"key":"ref24","year":"0"},{"key":"ref23","first-page":"103","article-title":"GPipe: Efficient training of giant neural networks using pipeline parallelism","author":"huang","year":"2019","journal-title":"Proc 33rd Int Conf Neural Informat Process Syst"},{"key":"ref26","article-title":"Intel machine learning scaling library","year":"2019"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1177\/1094342005051521"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10602-1_48"},{"key":"ref55","year":"0"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2015.11"},{"key":"ref53","year":"0"},{"key":"ref52","year":"0"},{"key":"ref10","article-title":"An empirical model of large-batch training","author":"mccandlish","year":"2018"},{"key":"ref40","article-title":"Accurate, large minibatch SGD: Training ImageNet in 1 hour","author":"goyal","year":"2017"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1137\/16M1080173"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1145\/3295500.3356137"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref14","article-title":"BERT: Pre-training of deep bidirectional transformers for language understanding","author":"devlin","year":"2018"},{"key":"ref15","first-page":"2408","article-title":"Optimizing neural networks with Kronecker-factored approximate curvature","author":"martens","year":"2015","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref16","first-page":"5053","article-title":"Inefficiency of K-FAC for large batch size training","author":"ma","year":"2019","journal-title":"Proc AAAI Conf Artif Intell"},{"key":"ref17","year":"0"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1007\/BF01589116"},{"key":"ref19","first-page":"8024","article-title":"PyTorch: An imperative style, high-performance deep learning library","author":"paszke","year":"2019","journal-title":"Proc Adv Neural Informat Process Syst"},{"key":"ref4","article-title":"Image classification at supercomputer scale","author":"ying","year":"2018"},{"key":"ref3","article-title":"Extremely large minibatch SGD: Training ResNet-50 on ImageNet in 15 minutes","author":"akiba","year":"2017"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.01264"},{"key":"ref5","article-title":"Massively distributed SGD: ImageNet\/ResNet-50 training in a flash","author":"mikami","year":"2018"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1038\/nphys4035"},{"key":"ref49","article-title":"Learning multiple layers of features from tiny images","author":"krizhevsky","year":"2009"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1109\/DLS49591.2019.00007"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1038\/s41586-019-1116-4"},{"key":"ref46","first-page":"1","article-title":"Mixed precision training","author":"micikevicius","year":"2018"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1145\/3394486.3406703"},{"key":"ref48","first-page":"9573","article-title":"Fast approximate natural gradient descent in a Kronecker-factored eigenbasis","author":"george","year":"2018","journal-title":"Proc 32nd Int Conf Neural Informat Process Syst"},{"key":"ref47","first-page":"6566","article-title":"EigenDamage: Structured pruning in the Kronecker-factored eigenbasis","author":"wang","year":"2019","journal-title":"Proc 36th Int Conf Mach Learn"},{"key":"ref42","first-page":"1","article-title":"Kronecker-factored curvature approximations for recurrent neural networks","author":"martens","year":"2018","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref41","first-page":"1","article-title":"Distributed second-order optimization using Kronecker-factored approximations","author":"ba","year":"2017","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2020.3004354"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1145\/3394486.3403265"}],"container-title":["IEEE Transactions on Parallel and Distributed Systems"],"original-title":[],"link":[{"URL":"https:\/\/ieeexplore.ieee.org\/ielam\/71\/9790018\/9739867-aam.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/71\/9790018\/09739867.pdf?arnumber=9739867","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,8,5]],"date-time":"2022-08-05T00:23:38Z","timestamp":1659659018000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/9739867\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,12,1]]},"references-count":55,"journal-issue":{"issue":"12"},"URL":"https:\/\/doi.org\/10.1109\/tpds.2022.3161187","relation":{},"ISSN":["1045-9219","1558-2183","2161-9883"],"issn-type":[{"value":"1045-9219","type":"print"},{"value":"1558-2183","type":"electronic"},{"value":"2161-9883","type":"electronic"}],"subject":[],"published":{"date-parts":[[2022,12,1]]}}}