{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2024,9,8]],"date-time":"2024-09-08T11:11:37Z","timestamp":1725793897635},"reference-count":33,"publisher":"IEEE","license":[{"start":{"date-parts":[[2024,5,8]],"date-time":"2024-05-08T00:00:00Z","timestamp":1715126400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2024,5,8]],"date-time":"2024-05-08T00:00:00Z","timestamp":1715126400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100001809","name":"National Natural Science Foundation of China","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100018601","name":"Guangdong Special Support Plan","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100018601","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2024,5,8]]},"DOI":"10.1109\/cscwd61410.2024.10580052","type":"proceedings-article","created":{"date-parts":[[2024,7,10]],"date-time":"2024-07-10T17:21:49Z","timestamp":1720632109000},"page":"1376-1381","source":"Crossref","is-referenced-by-count":0,"title":["EINS: Edge-Cloud Deep Model Inference with Network-Efficiency Schedule in Serverless"],"prefix":"10.1109","author":[{"given":"Shijie","family":"Peng","sequence":"first","affiliation":[{"name":"Chinese Academy of Sciences,Shenzhen Institute of Advanced Technology"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yanying","family":"Lin","sequence":"additional","affiliation":[{"name":"Chinese Academy of Sciences,Shenzhen Institute of Advanced Technology"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Wenyan","family":"Chen","sequence":"additional","affiliation":[{"name":"Chinese Academy of Sciences,Shenzhen Institute of Advanced Technology"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Yingfei","family":"Tang","sequence":"additional","affiliation":[{"name":"Chinese Academy of Sciences,Shenzhen Institute of Advanced Technology"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Xu","family":"Duan","sequence":"additional","affiliation":[{"name":"Chinese Academy of Sciences,Shenzhen Institute of Advanced Technology"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Kejiang","family":"Ye","sequence":"additional","affiliation":[{"name":"Chinese Academy of Sciences,Shenzhen Institute of Advanced Technology"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"article-title":"Bert: Pre-training of deep bidirectional transformers for language understanding","year":"2019","author":"Devlin","key":"ref1"},{"article-title":"Llama: Open and efficient foundation language models","year":"2023","author":"Touvron","key":"ref2"},{"key":"ref3","first-page":"28492","article-title":"Robust speech recognition via large-scale weak supervision","volume-title":"International Conference on Machine Learning","author":"Radford"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1016\/j.ins.2021.06.039"},{"article-title":"Lora: Low-rank adaptation of large language models","year":"2021","author":"Hu","key":"ref5"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2022.3166101"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1145\/3341301.3359655"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1145\/3297858.3304011"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1145\/3318216.3363305"},{"key":"ref10","first-page":"35","article-title":"Starlight: Fast container provisioning on the edge and over the wan","volume-title":"19th USENIX Symposium on Networked Systems Design and Implementation (NSDI 22)","author":"Chen"},{"key":"ref11","first-page":"1267","article-title":"Rdc:energy-efficient data center network congestion relief with topological reconfigurability at the edge","volume-title":"19th USENIX Symposium on Networked Systems Design and Implementation (NSDI 22)","author":"Wang"},{"key":"ref12","first-page":"977","article-title":"Swarmmap: Scaling up real-time collaborative visual slam at the edge","volume-title":"19th USENIX Symposium on Networked Systems Design and Implementation (NSDI 22)","author":"Xu"},{"key":"ref13","doi-asserted-by":"publisher","DOI":"10.1145\/3503222.3507713"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1145\/3472883.3487014"},{"article-title":"Tetris: Memory-efficient serverless inference through tensor sharing","volume-title":"2022 USENIX Annual Technical Conference, USENIX ATC 2022, Carlsbad, CA, USA, July 11-13, 2022","author":"Li","key":"ref15"},{"key":"ref16","first-page":"1049","article-title":"Mark: Exploiting cloud services for cost-effective, slo-aware machine learning inference serving","volume-title":"Proceedings of the 2019 USENIX Conference on Usenix Annual Technical Conference","author":"Zhang"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1145\/3466752.3480095"},{"key":"ref18","first-page":"201","article-title":"Cost-effective cloud edge traffic engineering with cascara","volume-title":"18th USENIX Symposium on Networked Systems Design and Implementation (NSDI 21)","author":"Singh"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/TGCN.2021.3111731"},{"key":"ref20","first-page":"56","article-title":"Bisone: A lightweight and high-performance accelerator for narrow integer linear algebra computing on the edge","volume-title":"ASPLOS \u201922: 27th ACM International Conference on Architectural Support for Programming Languages and Operating Systems","author":"Reggiani"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO50266.2020.00090"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/TWC.2019.2946140"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1145\/3229556.3229562"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1145\/3397315"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS47774.2020.00101"},{"article-title":"Lottery aware sparsity hunting: Enabling federated learning on resource-limited edge","year":"2022","author":"Babakniya","key":"ref26"},{"key":"ref27","doi-asserted-by":"publisher","DOI":"10.1109\/TNSM.2020.3023305"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2022.3167417"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.1109\/CCGrid54584.2022.00028"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1109\/TII.2022.3217477"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1145\/3470496.3527407"},{"key":"ref32","doi-asserted-by":"publisher","DOI":"10.1109\/MNET.125.2200294"},{"first-page":"15","article-title":"Serverless in the wild: Characterizing and optimizing the serverless workload at a large cloud provider","author":"Shahrad","key":"ref33"}],"event":{"name":"2024 27th International Conference on Computer Supported Cooperative Work in Design (CSCWD)","start":{"date-parts":[[2024,5,8]]},"location":"Tianjin, China","end":{"date-parts":[[2024,5,10]]}},"container-title":["2024 27th International Conference on Computer Supported Cooperative Work in Design (CSCWD)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/10579968\/10579982\/10580052.pdf?arnumber=10580052","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2024,7,18]],"date-time":"2024-07-18T05:14:45Z","timestamp":1721279685000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10580052\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2024,5,8]]},"references-count":33,"URL":"https:\/\/doi.org\/10.1109\/cscwd61410.2024.10580052","relation":{},"subject":[],"published":{"date-parts":[[2024,5,8]]}}}