{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,6,30]],"date-time":"2026-06-30T15:44:24Z","timestamp":1782834264227,"version":"3.54.5"},"reference-count":35,"publisher":"IEEE","license":[{"start":{"date-parts":[[2023,6,19]],"date-time":"2023-06-19T00:00:00Z","timestamp":1687132800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2023,6,19]],"date-time":"2023-06-19T00:00:00Z","timestamp":1687132800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Science Foundation of China (NSFC)","doi-asserted-by":"publisher","award":["62102392,62132019"],"award-info":[{"award-number":["62102392,62132019"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100012226","name":"Fundamental Research Funds for the Central Universities","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100012226","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2023,6,19]]},"DOI":"10.1109\/iwqos57198.2023.10188783","type":"proceedings-article","created":{"date-parts":[[2023,7,27]],"date-time":"2023-07-27T17:21:00Z","timestamp":1690478460000},"page":"1-10","source":"Crossref","is-referenced-by-count":17,"title":["GOAT: Gradient Scheduling with Collaborative In-Network Aggregation for Distributed Training"],"prefix":"10.1109","author":[{"given":"Jin","family":"Fang","sequence":"first","affiliation":[{"name":"School of Computer Science and Technology, University of Science and Technology of China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Gongming","family":"Zhao","sequence":"additional","affiliation":[{"name":"School of Computer Science and Technology, University of Science and Technology of China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Hongli","family":"Xu","sequence":"additional","affiliation":[{"name":"School of Computer Science and Technology, University of Science and Technology of China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhuolong","family":"Yu","sequence":"additional","affiliation":[{"name":"Johns Hopkins University,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Bingchen","family":"Shen","sequence":"additional","affiliation":[{"name":"School of Computer Science and Technology, University of Science and Technology of China"}],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Liguang","family":"Xie","sequence":"additional","affiliation":[{"name":"Virginia Tech,USA"}],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1145\/3387514.3405855"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1145\/3127479.3132018"},{"key":"ref12","year":"0"},{"key":"ref34","year":"0","journal-title":"Behavioral model version 2 (bmv2)"},{"key":"ref15","first-page":"51","article-title":"{PowerTCP}: Pushing the performance limits of datacenter networks","author":"addanki","year":"0","journal-title":"19th USENIX Symposium on Networked Systems Design and Implementation (NSDI 22)"},{"key":"ref14","article-title":"Efficient data-plane memory scheduling for in-network aggregation","author":"wang","year":"2022","journal-title":"ArXiv Preprint"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1007\/s11263-015-0816-y"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1145\/1402946.1402967"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/TNET.2023.3244794"},{"key":"ref33","year":"0","journal-title":"An Instant Virtual Network on Your Laptop"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1145\/3307650.3322259"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1145\/2934872.2934893"},{"key":"ref2","article-title":"Bert: Pre-training of deep bidirectional transformers for language understanding","volume":"abs 1810 4805","author":"devlin","year":"2019","journal-title":"ArXiv"},{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2016.2582748"},{"key":"ref16","article-title":"Alpa: Automating inter- and intra-operator parallelism for distributed deep learning","author":"zheng","year":"2022","journal-title":"ArXiv Preprint"},{"key":"ref19","year":"0","journal-title":"Intel tofino 2"},{"key":"ref18","article-title":"Elastic-df: Scaling performance of dnn inference in fpga clouds through automatic partitioning","author":"alonso","year":"2021","journal-title":"ACM Trans Reconfigurable Technol Syst"},{"key":"ref24","first-page":"8024","article-title":"Pytorch: An imperative style, high-performance deep learning library","author":"paszke","year":"2019","journal-title":"Advances in Neural IInformation Processing Systems"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1109\/TNET.2018.2853587"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM41043.2020.9155282"},{"key":"ref25","year":"0"},{"key":"ref20","author":"wen","year":"2017","journal-title":"Terngrad Ternary gradients to reduce communication in distributed deep learning"},{"key":"ref22","year":"0"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1007\/BF01585178"},{"key":"ref28","article-title":"Densenet: Implementing efficient convnet descriptor pyramids","author":"iandola","year":"2014","journal-title":"ArXiv Preprint"},{"key":"ref27","author":"krizhevsky","year":"0","journal-title":"Cifar-100 (canadian institute for advanced research)"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1145\/2829988.2787507"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1145\/3452296.3472904"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1145\/3267809.3267840"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1145\/3387514.3405895"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1145\/2640087.2644155"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1145\/3285029"},{"key":"ref6","first-page":"785","article-title":"Scaling distributed machine learning with In-Network aggregation","author":"sapio","year":"2021","journal-title":"18th USENIX Symposium on Networked Systems Design and Implementation (NSDI 21)"},{"key":"ref5","first-page":"741","article-title":"ATP: In-network aggregation for multi-tenant learning","author":"lao","year":"2021","journal-title":"18th USENIX Symposium on Networked Systems Design and Implementation (NSDI 21)"}],"event":{"name":"2023 IEEE\/ACM 31st International Symposium on Quality of Service (IWQoS)","location":"Orlando, FL, USA","start":{"date-parts":[[2023,6,19]]},"end":{"date-parts":[[2023,6,21]]}},"container-title":["2023 IEEE\/ACM 31st International Symposium on Quality of Service (IWQoS)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/10188660\/10188690\/10188783.pdf?arnumber=10188783","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2023,8,14]],"date-time":"2023-08-14T17:39:16Z","timestamp":1692034756000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10188783\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2023,6,19]]},"references-count":35,"URL":"https:\/\/doi.org\/10.1109\/iwqos57198.2023.10188783","relation":{},"subject":[],"published":{"date-parts":[[2023,6,19]]}}}