{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,18]],"date-time":"2026-08-18T00:35:42Z","timestamp":1787013342448,"version":"build-2736575974"},"reference-count":31,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"7","license":[{"start":{"date-parts":[[2022,7,1]],"date-time":"2022-07-01T00:00:00Z","timestamp":1656633600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2022,7,1]],"date-time":"2022-07-01T00:00:00Z","timestamp":1656633600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2022,7,1]],"date-time":"2022-07-01T00:00:00Z","timestamp":1656633600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"Hong Kong RGC Research Impact Fund","award":["R5060-19"],"award-info":[{"award-number":["R5060-19"]}]},{"name":"RGC Collaborative Research Fund","award":["C6021-19EF"],"award-info":[{"award-number":["C6021-19EF"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Parallel Distrib. Syst."],"published-print":{"date-parts":[[2022,7,1]]},"DOI":"10.1109\/tpds.2021.3118609","type":"journal-article","created":{"date-parts":[[2021,10,10]],"date-time":"2021-10-10T12:06:26Z","timestamp":1633867586000},"page":"1569-1580","source":"Crossref","is-referenced-by-count":13,"title":["SaPus: Self-Adaptive Parameter Update Strategy for DNN Training on Multi-GPU Clusters"],"prefix":"10.1109","volume":"33","author":[{"ORCID":"https:\/\/orcid.org\/0000-0001-7717-4306","authenticated-orcid":false,"given":"Zhaorui","family":"Zhang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Choli","family":"Wang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref31","first-page":"1","article-title":"Estimate sequences for stochastic composite optimization: Variance reduction, acceleration, and robustness to noise","volume":"21","author":"kulunchakov","year":"2020","journal-title":"J Mach Learn Res"},{"key":"ref30","doi-asserted-by":"crossref","first-page":"245","DOI":"10.1016\/j.automatica.2012.08.042","article-title":"Event-based broadcasting for multi-agent average consensus","volume":"49","author":"georg","year":"2013","journal-title":"Automatica"},{"key":"ref10","first-page":"4120","article-title":"Asynchronous stochastic gradient descent with delay compensation","author":"zheng","year":"2017","journal-title":"Proc Int Conf Mach Learn"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1145\/3404397.3404432"},{"key":"ref12","first-page":"4229","article-title":"Communication-efficient federated deep learning with layerwise asynchronous model update and temporally weighted aggregation","author":"chen","year":"2019","journal-title":"IEEE Trans Neural Netw Learn Syst"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1145\/3369583.3392681"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM41043.2020.9155269"},{"key":"ref15","first-page":"3817","article-title":"On the discrepancy between the theoretical analysis and practical implementations of compressed communication for distributed deep learning","volume":"34","author":"aritra","year":"2020","journal-title":"Proc AAAI Conf Artif Intell"},{"key":"ref16","author":"han","year":"2012","journal-title":"Data Mining Concepts and Techniques"},{"key":"ref17","author":"jorge","year":"2006","journal-title":"Numerical Optimization"},{"key":"ref18","doi-asserted-by":"crossref","first-page":"838","DOI":"10.1137\/0330046","article-title":"Acceleration of stochastic approximation by averaging","volume":"30","author":"boris","year":"1992","journal-title":"SIAM J Control Optim"},{"key":"ref19","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1561\/2200000016","article-title":"Distributed optimization and statistical learning via the alternating direction method of multipliers","volume":"3","author":"stephen","year":"2010","journal-title":"Found Trends Mach Learn"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/TCOMM.2020.3026398"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1145\/3363554"},{"key":"ref27","first-page":"440","article-title":"Sparse communication for distributed gradient descent","author":"fikri","year":"2017","journal-title":"Proc Conf Empirical Methods Nat Lang Process"},{"key":"ref3","first-page":"418","article-title":"TicTac: Accelerating distributed deep learning with communication scheduling","author":"hashemi","year":"2019","journal-title":"Proc 2nd SysML Conf"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2020.3047894"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP40776.2020.9053834"},{"key":"ref5","first-page":"212","article-title":"Adaptive communication strategies to achieve the best error-runtime trade-off in local-update SGD","author":"wang","year":"2019","journal-title":"Proc 2nd SysML Conf"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v33i01.33015693"},{"key":"ref7","first-page":"685","article-title":"Deep learning with elastic averaging SGD","author":"zhang","year":"2015","journal-title":"Proc Int Conf Neural Inf Process"},{"key":"ref2","first-page":"463","article-title":"A unified architecture for accelerating distributed DNN training in heterogeneous GPU\/CPU clusters","author":"jiang","year":"2020","journal-title":"Proc 14th USENIX Conf Operating Syst Des Implementation"},{"key":"ref9","article-title":"Local SGD converges fast and communicates little","author":"stich","year":"2018","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref1","first-page":"265","article-title":"Tensorflow: A system for large-scale machine learning","author":"abadi","year":"2016","journal-title":"Proc 12th USENIX Conf Operating Syst Des Implementation"},{"key":"ref20","author":"hestenes","year":"1975","journal-title":"Optimization Theory The Finite Dimensional Case"},{"key":"ref22","first-page":"1223","article-title":"Large scale distributed deep networks","author":"jeffrey","year":"2012","journal-title":"Proc Int Conf Neural Inf Process"},{"key":"ref21","first-page":"1","article-title":"ADMM consensus for deep LSTM networks","author":"antonello","year":"2020","journal-title":"Proc Int Joint Conf Neural Netw"},{"key":"ref24","article-title":"Efficient-Adam: Communication-efficient distributed Adam with complexity analysis","author":"chen","year":"2020","journal-title":"Proc Int Conf Learn Representations"},{"key":"ref23","first-page":"1223","article-title":"More effective distributed ML via a stale synchronous parallel parameter server","author":"ho","year":"2013","journal-title":"Proc Int Conf Neural Inf Process"},{"key":"ref26","first-page":"1508","article-title":"TernGrad: Ternary gradients to reduce communication in distributed deep learning","author":"wei","year":"2017","journal-title":"Proc Int Conf Neural Inf Process"},{"key":"ref25","article-title":"Revisiting distributed synchronous SGD","author":"chen","year":"2016"}],"container-title":["IEEE Transactions on Parallel and Distributed Systems"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/71\/9603995\/09563229.pdf?arnumber=9563229","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,4,8]],"date-time":"2022-04-08T13:55:51Z","timestamp":1649426151000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/9563229\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2022,7,1]]},"references-count":31,"journal-issue":{"issue":"7"},"URL":"https:\/\/doi.org\/10.1109\/tpds.2021.3118609","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":[[2022,7,1]]}}}