{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,9,3]],"date-time":"2026-09-03T04:28:26Z","timestamp":1788409706402,"version":"build-2803163510"},"reference-count":55,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"4","license":[{"start":{"date-parts":[[2023,4,1]],"date-time":"2023-04-01T00:00:00Z","timestamp":1680307200000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2023,4,1]],"date-time":"2023-04-01T00:00:00Z","timestamp":1680307200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2023,4,1]],"date-time":"2023-04-01T00:00:00Z","timestamp":1680307200000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"Alibaba Group"},{"name":"Hong Kong RGC","award":["17204619"],"award-info":[{"award-number":["17204619"]}]},{"name":"Hong Kong RGC","award":["17208920"],"award-info":[{"award-number":["17208920"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Trans. Parallel Distrib. Syst."],"published-print":{"date-parts":[[2023,4]]},"DOI":"10.1109\/tpds.2023.3243261","type":"journal-article","created":{"date-parts":[[2023,2,8]],"date-time":"2023-02-08T18:34:41Z","timestamp":1675881281000},"page":"1281-1293","source":"Crossref","is-referenced-by-count":10,"title":["Expediting Distributed DNN Training With Device Topology-Aware Graph Deployment"],"prefix":"10.1109","volume":"34","author":[{"ORCID":"https:\/\/orcid.org\/0000-0003-0838-1883","authenticated-orcid":false,"given":"Shiwei","family":"Zhang","sequence":"first","affiliation":[{"name":"Department of Computer Science, University of Hong Kong, Pokfulam, Hong Kong"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-7860-9543","authenticated-orcid":false,"given":"Xiaodong","family":"Yi","sequence":"additional","affiliation":[{"name":"Department of Computer Science, University of Hong Kong, Pokfulam, Hong Kong"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Lansong","family":"Diao","sequence":"additional","affiliation":[{"name":"Alibaba Group, Hangzhou, Zhejiang, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-3144-4398","authenticated-orcid":false,"given":"Chuan","family":"Wu","sequence":"additional","affiliation":[{"name":"Department of Computer Science, University of Hong Kong, Pokfulam, Hong Kong"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Siyu","family":"Wang","sequence":"additional","affiliation":[{"name":"Alibaba Group, Hangzhou, Zhejiang, China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Wei","family":"Lin","sequence":"additional","affiliation":[{"name":"Alibaba Group, Hangzhou, Zhejiang, China"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.308"},{"key":"ref3","article-title":"Language models are few-shot learners","author":"Brown","year":"2020"},{"key":"ref4","article-title":"PaLM: Scaling language modeling with pathways","author":"Chowdhery","year":"2022"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1145\/3219819.3219869"},{"key":"ref6","article-title":"BERT: Pre-training of deep bidirectional transformers for language understanding","author":"Devlin","year":"2018"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1093\/bioinformatics\/btz682"},{"key":"ref8","article-title":"Placeto: Learning generalizable device placement algorithms for distributed machine learning","author":"Addanki","year":"2019"},{"key":"ref9","article-title":"GDP: Generalized device placement for dataflow graphs","author":"Zhou","year":"2019"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1145\/3386367.3432728"},{"key":"ref11","first-page":"906","article-title":"AutoSync: Learning to synchronize for data-parallel distributed deep learning","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"33","author":"Zhang"},{"key":"ref12","article-title":"Horovod: Fast and easy distributed deep learning in tensorflow","author":"Sergeev","year":"2018"},{"key":"ref13","article-title":"Nccl 2.0","volume-title":"Proc. GPU Technol. Conf.","author":"Jeaugey"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.5555\/2685048.2685095"},{"key":"ref15","first-page":"463","article-title":"A unified architecture for accelerating distributed {DNN} training in heterogeneous GPU\/CPU clusters","volume-title":"Proc. 14th {USENIX} Symp. Operating Syst. Des. Implementation","author":"Jiang"},{"key":"ref16","article-title":"Deep gradient compression: Reducing the communication bandwidth for distributed training","author":"Lin","year":"2047"},{"key":"ref17","article-title":"Parallax: Sparsity-aware data parallel training of deep neural networks","author":"Kim","year":"2018"},{"key":"ref18","article-title":"A hierarchical model for device placement","volume-title":"Proc. Int. Conf. Learn. Representations","author":"Mirhoseini"},{"key":"ref19","first-page":"2430","article-title":"Device placement optimization with reinforcement learning","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Mirhoseini"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1145\/3292500.3330961"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1007\/11871842_29"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-642-40988-2_13"},{"key":"ref23","article-title":"Distributed machine learning via sufficient factor broadcasting","author":"Xie","year":"2015"},{"issue":"2016","key":"ref24","first-page":"265","article-title":"TensorFlow: A system for large-scale machine learning","volume-title":"Proc. 12th USENIX Symp. Operating Syst. Des. Implementation","volume":"16","author":"Abadi"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.2307\/j.ctvcm4g18.8"},{"key":"ref26","article-title":"MXNet: A flexible and efficient machine learning library for heterogeneous distributed systems","author":"Chen","year":"2015"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1109\/MM.2019.2935967"},{"key":"ref28","article-title":"Transferable graph optimizers for ML compilers","author":"Zhou","year":"2020"},{"key":"ref29","first-page":"1676","article-title":"Spotlight: Optimizing device placement for training deep neural networks","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Gao"},{"key":"ref30","first-page":"1","article-title":"Beyond data and model parallelism for deep neural networks","volume-title":"Proc. Mach. Learn. Syst.","volume":"1","author":"Jia"},{"key":"ref31","article-title":"Reinforced genetic algorithm learning for optimizing computation graphs","author":"Paliwal","year":"2019"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1145\/3464298.3476132"},{"key":"ref33","first-page":"515","article-title":"Hived: Sharing a {GPU} cluster for deep learning with guarantees","volume-title":"Proc. 14th {USENIX} Symp. Operating Syst. Des. Implementation","author":"Zhao","year":"2020"},{"key":"ref34","first-page":"945","article-title":"MLaaS in the wild: Workload analysis and scheduling in large-scale heterogeneous GPU clusters","volume-title":"Proc. 19th USENIX Symp. Networked Syst. Des. Implementation","author":"Weng"},{"key":"ref35","first-page":"571","article-title":"Project adam: Building an efficient and scalable deep learning training system","volume-title":"Proc. 11th USENIX Symp. Operating Syst. Des. Implementation","author":"Chilimbi"},{"key":"ref36","first-page":"181","article-title":"Poseidon: An efficient communication architecture for distributed deep learning on {GPU} clusters","volume-title":"Proc. {USENIX} Annu. Tech. Conf.","author":"Zhang","year":"2017"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2014-274"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2015-354"},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO50266.2020.00058"},{"key":"ref40","article-title":"A study on overfitting in deep reinforcement learning","author":"Zhang","year":"2018"},{"key":"ref41","article-title":"Observational overfitting in reinforcement learning","author":"Song","year":"2019"},{"key":"ref42","article-title":"TensorFlow-Slim: A lightweight library for defining, training and evaluating complex models in tensorflow","author":"Guadarrama","year":"2016"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1137\/s1064827595287997"},{"key":"ref44","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM41043.2020.9155446"},{"key":"ref45","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v34i04.5747"},{"key":"ref46","article-title":"A note on over-smoothing for graph neural networks","author":"Cai","year":"2020"},{"key":"ref47","article-title":"Graph attention networks","author":"Veli\u010dkovi\u0107","year":"2017"},{"key":"ref48","article-title":"COIN-OR\/CBC: Version 2.99","volume-title":"URL","volume":"1317566","author":"Forrest","year":"2018"},{"key":"ref49","article-title":"Deep graph library: A graph-centric, highly-performant package for graph neural networks","author":"Wang","year":"2019"},{"key":"ref50","article-title":"Adam: A method for stochastic optimization","author":"Kingma","year":"2014"},{"key":"ref51","article-title":"Very deep convolutional networks for large-scale image recognition","author":"Simonyan","year":"2014"},{"key":"ref52","article-title":"Attention is all you need","author":"Vaswani","year":"2017"},{"key":"ref53","first-page":"9971","article-title":"Post: Device placement with cross-entropy minimization and proximal policy optimization","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Gao"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1145\/3419111.3421302"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1109\/SC.2012.71"}],"container-title":["IEEE Transactions on Parallel and Distributed Systems"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/71\/10043602\/10040900.pdf?arnumber=10040900","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,2,13]],"date-time":"2024-02-13T13:11:51Z","timestamp":1707829911000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10040900\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,4]]},"references-count":55,"journal-issue":{"issue":"4"},"URL":"https:\/\/doi.org\/10.1109\/tpds.2023.3243261","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":[[2023,4]]}}}