{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,16]],"date-time":"2026-07-16T15:32:45Z","timestamp":1784215965164,"version":"3.55.0"},"reference-count":292,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"2","license":[{"start":{"date-parts":[[2020,1,1]],"date-time":"2020-01-01T00:00:00Z","timestamp":1577836800000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2020,1,1]],"date-time":"2020-01-01T00:00:00Z","timestamp":1577836800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2020,1,1]],"date-time":"2020-01-01T00:00:00Z","timestamp":1577836800000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"DOI":"10.13039\/501100012166","name":"National Key Research and Development Program of China","doi-asserted-by":"crossref","award":["2019YFB2101901"],"award-info":[{"award-number":["2019YFB2101901"]}],"id":[{"id":"10.13039\/501100012166","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/501100012166","name":"National Key Research and Development Program of China","doi-asserted-by":"crossref","award":["2018YFC0809803"],"award-info":[{"award-number":["2018YFC0809803"]}],"id":[{"id":"10.13039\/501100012166","id-type":"DOI","asserted-by":"crossref"}]},{"DOI":"10.13039\/501100001809","name":"National Science Foundation of China","doi-asserted-by":"publisher","award":["61702364"],"award-info":[{"award-number":["61702364"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Science Foundation of China","doi-asserted-by":"publisher","award":["61972432"],"award-info":[{"award-number":["61972432"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"DOI":"10.13039\/501100001809","name":"National Science Foundation of China","doi-asserted-by":"publisher","award":["U1711265"],"award-info":[{"award-number":["U1711265"]}],"id":[{"id":"10.13039\/501100001809","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Program for Guangdong Introducing Innovative and Enterpreneurial Teams","award":["2017ZT07X355"],"award-info":[{"award-number":["2017ZT07X355"]}]},{"name":"Chinese National Engineering Laboratory for Big Data System Computing Technology"},{"DOI":"10.13039\/501100000038","name":"Canadian Natural Sciences and Engineering Research Council","doi-asserted-by":"publisher","id":[{"id":"10.13039\/501100000038","id-type":"DOI","asserted-by":"publisher"}]},{"name":"Singapore NRF National Satellite of Excellence, Design Science and Technology for Secure Critical Infrastructure National Satellite of Excellence","award":["DeST-SCI2019-0007"],"award-info":[{"award-number":["DeST-SCI2019-0007"]}]},{"name":"A*STAR-NTU-SUTD Joint Research Grant Call on Artificial Intelligence for the Future of Manufacturing","award":["RGANS1906"],"award-info":[{"award-number":["RGANS1906"]}]},{"name":"WASP\/NTU","award":["M4082187 (4080)"],"award-info":[{"award-number":["M4082187 (4080)"]}]},{"DOI":"10.13039\/501100001459","name":"Singapore MOE Tier 1","doi-asserted-by":"publisher","award":["2017-T1-002-007 RG122\/17"],"award-info":[{"award-number":["2017-T1-002-007 RG122\/17"]}],"id":[{"id":"10.13039\/501100001459","id-type":"DOI","asserted-by":"publisher"}]},{"name":"MOE Tier 2","award":["MOE2014-T2-2-015 ARC4\/15"],"award-info":[{"award-number":["MOE2014-T2-2-015 ARC4\/15"]}]},{"name":"Singapore","award":["NRF2015-NRF-ISF001-2277"],"award-info":[{"award-number":["NRF2015-NRF-ISF001-2277"]}]},{"DOI":"10.13039\/501100001447","name":"Singapore Energy Market Authority Energy Resilience","doi-asserted-by":"publisher","award":["NRF2017EWT-EP003-041"],"award-info":[{"award-number":["NRF2017EWT-EP003-041"]}],"id":[{"id":"10.13039\/501100001447","id-type":"DOI","asserted-by":"publisher"}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Commun. Surv. Tutorials"],"published-print":{"date-parts":[[2020]]},"DOI":"10.1109\/comst.2020.2970550","type":"journal-article","created":{"date-parts":[[2020,1,30]],"date-time":"2020-01-30T21:31:11Z","timestamp":1580419871000},"page":"869-904","source":"Crossref","is-referenced-by-count":1276,"title":["Convergence of Edge Computing and Deep Learning: A Comprehensive Survey"],"prefix":"10.1109","volume":"22","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-7223-1030","authenticated-orcid":false,"given":"Xiaofei","family":"Wang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-6124-4701","authenticated-orcid":false,"given":"Yiwen","family":"Han","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0003-3529-2640","authenticated-orcid":false,"given":"Victor C. M.","family":"Leung","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7442-7416","authenticated-orcid":false,"given":"Dusit","family":"Niyato","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xueqiang","family":"Yan","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"ORCID":"https:\/\/orcid.org\/0000-0001-9943-6020","authenticated-orcid":false,"given":"Xu","family":"Chen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref275","doi-asserted-by":"publisher","DOI":"10.1109\/MCOM.2018.1701233"},{"key":"ref274","doi-asserted-by":"publisher","DOI":"10.1145\/3297858.3304013"},{"key":"ref277","doi-asserted-by":"publisher","DOI":"10.1109\/BigDataCongress.2017.85"},{"key":"ref276","doi-asserted-by":"publisher","DOI":"10.1109\/TII.2019.2896965"},{"key":"ref271","author":"bogale","year":"2018","journal-title":"Machine Intelligence Techniques for Next-Generation Context-Aware Wireless Networks"},{"key":"ref270","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2018.2875926"},{"key":"ref273","doi-asserted-by":"publisher","DOI":"10.1109\/MCOM.2017.1700246"},{"key":"ref170","author":"venkatesan","year":"2016","journal-title":"Diving deeper into mentee networks"},{"key":"ref272","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2014.2371999"},{"key":"ref172","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM.2018.8485905"},{"key":"ref171","author":"biookaghazadeh","year":"2018","journal-title":"Are FPGAs suitable for edge computing?"},{"key":"ref174","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2019.2893250"},{"key":"ref173","doi-asserted-by":"publisher","DOI":"10.1109\/TII.2019.2897001"},{"key":"ref176","doi-asserted-by":"publisher","DOI":"10.1109\/ICNP.2017.8117585"},{"key":"ref175","doi-asserted-by":"publisher","DOI":"10.1145\/3267809.3267828"},{"key":"ref178","doi-asserted-by":"publisher","DOI":"10.1145\/1814433.1814441"},{"key":"ref177","doi-asserted-by":"publisher","DOI":"10.23919\/DATE.2017.7927211"},{"key":"ref168","doi-asserted-by":"publisher","DOI":"10.1145\/3220192.3220460"},{"key":"ref169","author":"simonyan","year":"2014","journal-title":"Very Deep Convolutional Networks for Large-scale Image Recognition"},{"key":"ref39","year":"2019","journal-title":"Intel Xeon Processor D-2100 Product Brief Advanced Intelligence for High-Density Edge Solutions"},{"key":"ref38","doi-asserted-by":"publisher","DOI":"10.1145\/3079856.3080246"},{"key":"ref33","year":"2019","journal-title":"Snapdragon 8 Series Mobile Platforms"},{"key":"ref32","year":"2019","journal-title":"An All-Scenario AI Infrastructure Solution That Bridges &#x2019;Device Edge and Cloud&#x2019; and Delivers Unrivaled Compute Power to Lead You Towards an AI-Fueled Future"},{"key":"ref31","year":"2019","journal-title":"Latest Jetson Products"},{"key":"ref30","year":"2019","journal-title":"Intel movidius neural compute stick"},{"key":"ref267","doi-asserted-by":"publisher","DOI":"10.1109\/JSAC.2015.2393496"},{"key":"ref37","year":"2019","journal-title":"NVIDIA Turing GPU Architecture"},{"key":"ref268","doi-asserted-by":"publisher","DOI":"10.1109\/TWC.2019.2927312"},{"key":"ref36","year":"2019","journal-title":"MediaTek Helio P60"},{"key":"ref269","doi-asserted-by":"publisher","DOI":"10.1109\/MNET.2019.1800458"},{"key":"ref35","year":"2019","journal-title":"The World&#x2019;s First Full-Stack All-Scenario AI Chip"},{"key":"ref34","year":"2019","journal-title":"Kirin"},{"key":"ref288","first-page":"2137","article-title":"Learning to communicate with deep multi-agent reinforcement learning","author":"foerster","year":"2016","journal-title":"Proc 29th Adv Neural Inf Process Syst (NeurIPS)"},{"key":"ref287","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2017.2707140"},{"key":"ref286","doi-asserted-by":"publisher","DOI":"10.1109\/MWC.2016.1600317WC"},{"key":"ref285","doi-asserted-by":"publisher","DOI":"10.1109\/SEC.2016.8"},{"key":"ref284","doi-asserted-by":"publisher","DOI":"10.1145\/3213344.3213351"},{"key":"ref181","author":"lai","year":"2018","journal-title":"Rethinking machine learning development and deployment for edge devices"},{"key":"ref283","author":"nishio","year":"2018","journal-title":"Client selection for federated learning with heterogeneous resources in mobile edge"},{"key":"ref180","author":"polese","year":"2018","journal-title":"Machine Learning at the Edge A Data-Driven Architecture with Applications to 5G Cellular Networks"},{"key":"ref282","author":"wu","year":"2019","journal-title":"SAFA A semi-asynchronous protocol for fast federated learning with low overhead"},{"key":"ref281","author":"xie","year":"2019","journal-title":"Asynchronous federated optimization"},{"key":"ref280","first-page":"4120","article-title":"Asynchronous stochastic gradient descent with delay compensation","author":"zheng","year":"2017","journal-title":"Proc Int Conf Mach Learn (ICML)"},{"key":"ref185","first-page":"1","article-title":"ECO: Harmonizing edge and cloud with ML\/DL orchestration","author":"talagala","year":"2018","journal-title":"Proceedings of USENIX Workshop on Hot Topics in Edge Computing (HotEdge 18)"},{"key":"ref184","doi-asserted-by":"publisher","DOI":"10.1145\/3210240.3223570"},{"key":"ref183","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS.2019.00182"},{"key":"ref182","doi-asserted-by":"publisher","DOI":"10.1145\/3285017.3285019"},{"key":"ref189","doi-asserted-by":"publisher","DOI":"10.1109\/GLOCOM.2016.7842181"},{"key":"ref188","doi-asserted-by":"publisher","DOI":"10.1109\/SEC.2018.00010"},{"key":"ref187","first-page":"19","article-title":"openLEON: An end-to-end emulator from the edge data center to the mobile users carlos","author":"ramiro","year":"2018","journal-title":"Proc ACM Workshop Wireless Netw Testbeds Exp Eval Charact (WinTECH)"},{"key":"ref186","first-page":"1","article-title":"pCAMP: Performance comparison of machine learning packages on the edges","author":"zhang","year":"2018","journal-title":"Proceedings of USENIX Workshop on Hot Topics in Edge Computing (HotEdge 18)"},{"key":"ref28","year":"2019","journal-title":"Multi-access edge computing"},{"key":"ref27","first-page":"169","author":"bonomi","year":"2014","journal-title":"ch Fog Computing A Platform for Internet of Things and Analytics"},{"key":"ref179","doi-asserted-by":"publisher","DOI":"10.1038\/s41928-018-0059-3"},{"key":"ref29","year":"2019","journal-title":"What Is Azure Data Box Edge"},{"key":"ref20","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2019.2921977"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2017.2771153"},{"key":"ref21","author":"lim","year":"2019","journal-title":"Federated learning in mobile edge networks A comprehensive survey"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1109\/MPRV.2009.82"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1016\/j.comnet.2017.10.002"},{"key":"ref278","doi-asserted-by":"publisher","DOI":"10.1145\/3285017.3285027"},{"key":"ref26","doi-asserted-by":"publisher","DOI":"10.1145\/2342509.2342513"},{"key":"ref279","author":"wood","year":"2014","journal-title":"Ethereum A Secure Decentralised Generalised Transaction Ledger"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1109\/AINA.2015.254"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1109\/ICCCN.2018.8487352"},{"key":"ref292","author":"zhang","year":"2018","journal-title":"Deep learning on graphs A survey"},{"key":"ref51","doi-asserted-by":"publisher","DOI":"10.1145\/3020078.3021740"},{"key":"ref154","doi-asserted-by":"publisher","DOI":"10.1109\/IPSN.2016.7460664"},{"key":"ref153","doi-asserted-by":"publisher","DOI":"10.1145\/3007787.3001163"},{"key":"ref156","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2018.2858384"},{"key":"ref155","doi-asserted-by":"publisher","DOI":"10.1145\/3275219.3275236"},{"key":"ref150","doi-asserted-by":"publisher","DOI":"10.1145\/3240765.3243473"},{"key":"ref291","author":"zhou","year":"2018","journal-title":"Graph neural networks A review of methods and applications"},{"key":"ref152","first-page":"123","article-title":"MCDNN: An execution framework for deep neural networks on resource-constrained devices","author":"shen","year":"2016","journal-title":"Proc 9th Annu Int Conf Mobile Syst Appl Services (MobiSys)"},{"key":"ref290","first-page":"6379","article-title":"Multi-agent actor&#x2013;critic for mixed cooperative-competitive environments","author":"lowe","year":"2017","journal-title":"Proc 30th Adv Neural Inf Process Syst (NeurIPS)"},{"key":"ref151","doi-asserted-by":"publisher","DOI":"10.1145\/3131672.3131675"},{"key":"ref146","first-page":"1","article-title":"Deep compression: Compressing deep neural networks with pruning, trained quantization and Huffman coding","author":"han","year":"2016","journal-title":"Proc Int Conf Learn Represent (ICLR)ICLR"},{"key":"ref147","doi-asserted-by":"publisher","DOI":"10.1145\/2994551.2994564"},{"key":"ref148","doi-asserted-by":"publisher","DOI":"10.1145\/3211332.3211336"},{"key":"ref149","doi-asserted-by":"publisher","DOI":"10.1145\/3210240.3210337"},{"key":"ref289","first-page":"2681","article-title":"Deep decentralized multi-task multi-agent reinforcement learning under partial observability","author":"omidshafiei","year":"2017","journal-title":"Proc Int Conf Mach Learn (ICML)"},{"key":"ref59","year":"2019","journal-title":"Deeplearning4j Open-source distributed deep learning for the JVM Apache Software Foundation License 2 0"},{"key":"ref58","first-page":"265","article-title":"TensorFlow: A system for large-scale machine learning","author":"abadi","year":"2016","journal-title":"Proc 12th USENIX Conf Oper Syst Design Implement (OSDI)"},{"key":"ref57","first-page":"1","article-title":"Chainer: A next-generation open source framework for deep learning","author":"tokui","year":"2015","journal-title":"Proc Workshop Mach Learn Syst (LearningSys) 29th Annu Conf Neural Inf Process Syst (NeurIPS)"},{"key":"ref56","year":"2019","journal-title":"Microsoft Cognitive Toolkit (CNTK) an Open Source Deep-Learning Toolkit"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1109\/MCC.2014.51"},{"key":"ref54","author":"ignatov","year":"2018","journal-title":"AI benchmark running deep neural networks on Android smartphones"},{"key":"ref53","year":"2019","journal-title":"Qualcomm Neural Processing SDK for AI"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM.2018.8485850"},{"key":"ref40","year":"2019","journal-title":"Exynos 9820 Mobile Processor"},{"key":"ref167","first-page":"29","article-title":"Mainstream: Dynamic stem-sharing for multi-tenant video processing","author":"jiang","year":"2018","journal-title":"Proc of USENIX Annual Technical Conf (USENIX ATC)"},{"key":"ref166","doi-asserted-by":"publisher","DOI":"10.1145\/3241539.3241557"},{"key":"ref165","doi-asserted-by":"publisher","DOI":"10.1145\/3241539.3241563"},{"key":"ref164","doi-asserted-by":"publisher","DOI":"10.1145\/3081333.3081360"},{"key":"ref163","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS.2017.94"},{"key":"ref162","doi-asserted-by":"publisher","DOI":"10.1145\/3229556.3229562"},{"key":"ref161","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS.2017.226"},{"key":"ref160","doi-asserted-by":"publisher","DOI":"10.1109\/ICPR.2016.7900006"},{"key":"ref4","doi-asserted-by":"publisher","DOI":"10.1145\/2967360.2967369"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1109\/INFCOM.2013.6566921"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2016.2579198"},{"key":"ref5","year":"2018","journal-title":"Mobile-Edge Computing&#x2014;Introductory Technical White Paper"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1016\/j.sysarc.2019.02.009"},{"key":"ref159","first-page":"1","article-title":"MODI: Mobile deep inference made efficient by edge computing","author":"ogden","year":"2018","journal-title":"Proceedings of USENIX Workshop on Hot Topics in Edge Computing (HotEdge 18)"},{"key":"ref7","doi-asserted-by":"publisher","DOI":"10.1145\/3195970.3196036"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01418-6_40"},{"key":"ref157","doi-asserted-by":"publisher","DOI":"10.1145\/3274783.3275199"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.91"},{"key":"ref158","doi-asserted-by":"publisher","DOI":"10.1109\/MNET.2018.1700202"},{"key":"ref46","year":"2019","journal-title":"NVIDIA EGX Edge Computing Platform Real-Time AI at the Edge"},{"key":"ref45","year":"2019","journal-title":"Akraino Edge Stack"},{"key":"ref48","year":"2019","journal-title":"Google Cloud IoT Unlock Business Insights From Your Global Device Network With an Intelligent IoT Platform"},{"key":"ref47","year":"2019","journal-title":"AWS IoT Greengrass Bring Local Compute Messaging Data Caching Sync and ML Inference Capabilities to Edge Devices"},{"key":"ref42","year":"2019","journal-title":"OpenEdge Extend Cloud Computing Data and Service Seamlessly to Edge Devices"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1109\/SEC.2018.00048"},{"key":"ref44","year":"2019","journal-title":"EdgeX the Open Platform for the IoT Edge"},{"key":"ref43","year":"2019","journal-title":"Azure IoT Edge Extend Cloud Intelligence and Analytics to Edge Devices"},{"key":"ref73","doi-asserted-by":"publisher","DOI":"10.1109\/MNET.2017.1700037"},{"key":"ref72","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2019.2924377"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1109\/TMC.2018.2871842"},{"key":"ref70","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2017.2704444"},{"key":"ref76","doi-asserted-by":"publisher","DOI":"10.1109\/MWC.2019.1800466"},{"key":"ref77","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2017.2705720"},{"key":"ref74","doi-asserted-by":"publisher","DOI":"10.1109\/MNET.2018.1700394"},{"key":"ref75","year":"2015","journal-title":"Mobile edge computing A key technology towards 5G"},{"key":"ref78","doi-asserted-by":"publisher","DOI":"10.1038\/nature14539"},{"key":"ref79","author":"haykin","year":"2009","journal-title":"Neural Networks and Learning Machines"},{"key":"ref60","year":"2019","journal-title":"Deploy machine learning models on mobile and iot devices"},{"key":"ref62","year":"2019","journal-title":"Pytorch Tensors and Dynamic neural networks in Python with strong GPU acceleration"},{"key":"ref61","author":"chen","year":"2015","journal-title":"Mxnet A flexible and efficient machine learning library for heterogeneous distributed systems"},{"key":"ref63","year":"2019","journal-title":"Core ml Integrate machine learning models into your app"},{"key":"ref64","year":"2019","journal-title":"NCNN Is a High-Performance Neural Network Inference Framework Optimized for the Mobile Platform"},{"key":"ref65","year":"2019","journal-title":"MNN Is a Lightweight Deep Neural Network Inference Engine"},{"key":"ref66","year":"2019","journal-title":"Multi-Platform Embedded Deep Learning Framework"},{"key":"ref67","year":"2019","journal-title":"MACE Is a Deep Learning Inference Framework Optimized for Mobile Heterogeneous Computing Platforms"},{"key":"ref68","author":"wang","year":"2019","journal-title":"FANN-on-MCU An open-source toolkit for energy-efficient neural network inference at the edge of the Internet of Things"},{"key":"ref69","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2019.2927919"},{"key":"ref197","first-page":"265","article-title":"Calibrating noise to sensitivity in private data analysis","author":"dwork","year":"2006","journal-title":"Proceedings of Theoretical Cryptography Conference (TCC)"},{"key":"ref198","first-page":"1273","article-title":"Communication-efficient learning of deep networks from decentralized data","author":"mcmahan","year":"2017","journal-title":"Proc Int Conf Artif Intell Statist (AISTATS)"},{"key":"ref199","author":"bonawitz","year":"2019","journal-title":"Towards Federated Learning at Scale System Design"},{"key":"ref193","first-page":"1488","article-title":"Scalable distributed DNN training using commodity GPU cloud computing","author":"strom","year":"2015","journal-title":"Proc 16th Annu Conf Int Speech Commun Assoc (INTERSPEECH)"},{"key":"ref194","author":"jeong","year":"2018","journal-title":"Communication-efficient on-device machine learning Federated distillation and augmentation under non-iid private data"},{"key":"ref195","doi-asserted-by":"publisher","DOI":"10.1145\/2810103.2813677"},{"key":"ref196","doi-asserted-by":"publisher","DOI":"10.1109\/TBDATA.2018.2829886"},{"key":"ref95","first-page":"1223","article-title":"Large scale distributed deep networks","author":"dean","year":"2012","journal-title":"Proc 25th Adv Neural Inf Process Syst (NeurIPS)"},{"key":"ref94","first-page":"1057","article-title":"Policy gradient methods for reinforcement learning with function approximation","author":"sutton","year":"1999","journal-title":"Proc 12th Int Conf Neural Inf Process Syst (NeurIPS)"},{"key":"ref190","doi-asserted-by":"publisher","DOI":"10.1016\/j.pmcj.2017.07.014"},{"key":"ref93","author":"schulman","year":"2017","journal-title":"Proximal policy optimization algorithms"},{"key":"ref191","author":"lin","year":"2017","journal-title":"Deep Gradient Compression Reducing the Communication Bandwidth for Distributed Training"},{"key":"ref92","first-page":"1928","article-title":"Asynchronous methods for deep reinforcement learning","author":"mnih","year":"2016","journal-title":"Proc Int Conf Mach Learn (ICML)"},{"key":"ref192","first-page":"1","article-title":"eSGD: Communication efficient distributed deep learning on the edge","author":"tao","year":"2018","journal-title":"Proceedings of USENIX Workshop on Hot Topics in Edge Computing (HotEdge 18)"},{"key":"ref91","first-page":"1","article-title":"Continuous control with deep reinforcement learning","author":"lillicrap","year":"2016","journal-title":"Proc Int Conf Learn Represent (ICLR)ICLR"},{"key":"ref90","first-page":"1995","article-title":"Dueling network architectures for deep reinforcement learning","author":"wang","year":"2016","journal-title":"Proc Int Conf Mach Learn (ICML)"},{"key":"ref98","first-page":"1058","article-title":"1-bit stochastic gradient descent and its application to data-parallel distributed training of speech DNNs","author":"seide","year":"2014","journal-title":"Proc 15th Annu Conf Int Speech Commun Assoc (INTERSPEECH)"},{"key":"ref99","first-page":"1337","article-title":"Deep learning with cots HPC systems","author":"coates","year":"2013","journal-title":"Proc of the 30th Int Conf on Machine Learning"},{"key":"ref96","doi-asserted-by":"publisher","DOI":"10.14778\/2733004.2733082"},{"key":"ref97","first-page":"26","article-title":"Pipelined back-propagation for context-dependent deep neural networks","author":"chen","year":"2012","journal-title":"Proc 13th Annu Conf Int Speech Commun Assoc (INTERSPEECH)"},{"key":"ref82","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-10590-1_53"},{"key":"ref81","author":"manning","year":"1999","journal-title":"Foundations of Statistical Natural Language Processing"},{"key":"ref84","doi-asserted-by":"publisher","DOI":"10.1162\/neco.1997.9.8.1735"},{"key":"ref83","first-page":"2672","article-title":"Generative adversarial nets","author":"goodfellow","year":"2014","journal-title":"Proc 27th Adv Neural Inf Process Syst (NeurIPS)"},{"key":"ref80","doi-asserted-by":"publisher","DOI":"10.1145\/1015330.1015415"},{"key":"ref89","first-page":"2094","article-title":"Deep reinforcement learning with double Q-learning","author":"van hasselt","year":"2016","journal-title":"Proc AAAI Conf Artif Intell (AAAI)"},{"key":"ref85","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2009.191"},{"key":"ref86","author":"hinton","year":"2015","journal-title":"Distilling the knowledge in a neural network"},{"key":"ref87","first-page":"426","article-title":"Deep reinforcement learning: An overview","author":"mousavi","year":"2016","journal-title":"Proc SAI Intell Syst Conf (IntelliSys)"},{"key":"ref88","doi-asserted-by":"crossref","first-page":"529","DOI":"10.1038\/nature14236","article-title":"Human-level control through deep reinforcement learning","volume":"518","author":"mnih","year":"2015","journal-title":"Nature"},{"key":"ref200","doi-asserted-by":"publisher","DOI":"10.1109\/TCOMM.2019.2956472"},{"key":"ref101","year":"2019","journal-title":"Theano Is a Python Library That Allows You to Define Optimize and Evaluate Mathematical Expressions Involving Multi-Dimensional Arrays Efficiently"},{"key":"ref100","author":"moritz","year":"2015","journal-title":"Spark Net Training Deep Networks in Spark"},{"key":"ref209","doi-asserted-by":"publisher","DOI":"10.1109\/JSAC.2019.2904348"},{"key":"ref203","author":"kone?n\u00fd","year":"2016","journal-title":"Federated learning Strategies for improving communication efficiency"},{"key":"ref204","author":"reisizadeh","year":"2019","journal-title":"FedPAQ A communication-efficient federated learning method with periodic averaging and quantization"},{"key":"ref201","author":"xie","year":"2019","journal-title":"Practical distributed learning Secure machine learning with communication-efficient local updates"},{"key":"ref202","author":"abad","year":"2019","journal-title":"Hierarchical federated learning across heterogeneous cellular networks"},{"key":"ref207","author":"jiang","year":"2019","journal-title":"Model pruning enables efficient federated learning on edge devices"},{"key":"ref208","doi-asserted-by":"publisher","DOI":"10.1109\/INFOCOM.2018.8486403"},{"key":"ref205","author":"caldas","year":"2018","journal-title":"Expanding the reach of federated learning by reducing client resource requirements"},{"key":"ref206","first-page":"334","article-title":"Diameters of some finite-dimensional sets and classes of smooth functions","volume":"41","author":"kashin","year":"1977","journal-title":"Izvestiya Rossiiskoi Akademii Nauk Seriya Matematicheskaya"},{"key":"ref211","author":"hu","year":"2019","journal-title":"Distributed machine learning through heterogeneous edge systems"},{"key":"ref210","doi-asserted-by":"publisher","DOI":"10.1109\/INFCOMW.2018.8406837"},{"key":"ref212","doi-asserted-by":"publisher","DOI":"10.1109\/ICCD46524.2019.00038"},{"key":"ref213","doi-asserted-by":"publisher","DOI":"10.1214\/aoms\/1177729694"},{"key":"ref214","author":"yang","year":"2018","journal-title":"Federated learning via over-the-air computation"},{"key":"ref215","doi-asserted-by":"publisher","DOI":"10.1109\/TIT.2007.904785"},{"key":"ref216","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2018.2843321"},{"key":"ref217","author":"zhu","year":"2018","journal-title":"Broadband analog aggregation for low-latency federated edge learning (extended version)"},{"key":"ref218","author":"xu","year":"2019","journal-title":"ELFISH Resource-aware federated learning on heterogeneous edge devices"},{"key":"ref219","author":"dinh","year":"2019","journal-title":"Federated learning over wireless networks Convergence analysis and resource allocation"},{"key":"ref220","author":"chen","year":"2019","journal-title":"A joint learning and communications framework for federated learning over wireless networks"},{"key":"ref222","doi-asserted-by":"publisher","DOI":"10.1145\/3133956.3133982"},{"key":"ref221","author":"li","year":"2019","journal-title":"Fair resource allocation in federated learning"},{"key":"ref229","doi-asserted-by":"publisher","DOI":"10.1109\/MCOM.2016.7565185"},{"key":"ref228","doi-asserted-by":"publisher","DOI":"10.1109\/MCOM.2014.6736753"},{"key":"ref227","doi-asserted-by":"publisher","DOI":"10.1016\/B978-155860834-4\/50022-6"},{"key":"ref226","doi-asserted-by":"crossref","first-page":"5","DOI":"10.1109\/SURV.2013.050113.00015","article-title":"Fairness in wireless networks: Issues, measures and challenges","volume":"16","author":"shi","year":"2014","journal-title":"IEEE Commun Surveys Tuts"},{"key":"ref225","doi-asserted-by":"publisher","DOI":"10.1002\/nav.3800020109"},{"key":"ref224","doi-asserted-by":"publisher","DOI":"10.1016\/0047-2727(82)90020-2"},{"key":"ref223","author":"kim","year":"2018","journal-title":"On-device federated learning via blockchain and its latency analysis"},{"key":"ref127","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref126","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2015.7298594"},{"key":"ref125","first-page":"2285","article-title":"Compressing neural networks with the hashing trick","author":"chen","year":"2015","journal-title":"Proc Int Conf Int Conf Mach Learn (ICML)"},{"key":"ref124","first-page":"1269","article-title":"Exploiting linear structure within convolutional networks for efficient evaluation","author":"denton","year":"2014","journal-title":"Proc 27th Adv Neural Inf Process Syst (NeurIPS)"},{"key":"ref129","first-page":"1135","article-title":"Learning both weights and connections for efficient neural networks","author":"han","year":"2015","journal-title":"Proc 28th Adv Neural Inf Process Syst (NeurIPS)"},{"key":"ref128","author":"cheng","year":"2017","journal-title":"A survey of model compression and acceleration for deep neural networks"},{"key":"ref130","doi-asserted-by":"publisher","DOI":"10.1109\/MICRO.2016.7783725"},{"key":"ref133","first-page":"168","article-title":"Embedded binarized neural networks","author":"mcdanel","year":"2017","journal-title":"Proc Int Conf Embedded Wireless Syst Netw (EWSN)"},{"key":"ref134","author":"iandola","year":"2016","journal-title":"Squeezenet Alexnet-level accuracy with 50x fewer parameters and< 0 5 mb model size"},{"key":"ref131","first-page":"3123","article-title":"BinaryConnect: Training deep neural networks with binary weights during propagations","author":"courbariaux","year":"2015","journal-title":"Proc 28th Adv Neural Inf Process Syst (NeurIPS)"},{"key":"ref132","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-319-46493-0_32"},{"key":"ref232","doi-asserted-by":"publisher","DOI":"10.1109\/MNET.2019.1800058"},{"key":"ref233","doi-asserted-by":"publisher","DOI":"10.1109\/MWC.2018.1700317"},{"key":"ref230","doi-asserted-by":"publisher","DOI":"10.1109\/TCOMM.2017.2713384"},{"key":"ref231","doi-asserted-by":"publisher","DOI":"10.1109\/JSAC.2018.2844658"},{"key":"ref239","doi-asserted-by":"publisher","DOI":"10.1109\/MNET.2018.1800109"},{"key":"ref238","doi-asserted-by":"publisher","DOI":"10.1287\/opre.1070.0445"},{"key":"ref235","author":"ndikumana","year":"2018","journal-title":"Deep learning based caching for self-driving car in multi-access edge computing"},{"key":"ref234","doi-asserted-by":"publisher","DOI":"10.1007\/s00521-019-04040-z"},{"key":"ref237","doi-asserted-by":"publisher","DOI":"10.1016\/j.future.2018.08.019"},{"key":"ref236","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2002.1017616"},{"key":"ref136","doi-asserted-by":"publisher","DOI":"10.1145\/3220192.3220459"},{"key":"ref135","author":"howard","year":"2017","journal-title":"Mobilenets Efficient convolutional neural networks for mobile vision applications"},{"key":"ref138","doi-asserted-by":"publisher","DOI":"10.1145\/3230543.3230574"},{"key":"ref137","doi-asserted-by":"publisher","DOI":"10.1145\/3225058.3225103"},{"key":"ref139","doi-asserted-by":"publisher","DOI":"10.1109\/EDGE.2018.00025"},{"key":"ref140","doi-asserted-by":"publisher","DOI":"10.1145\/3300061.3300116"},{"key":"ref141","year":"2019","journal-title":"Homer Simpson"},{"key":"ref142","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00716"},{"key":"ref143","doi-asserted-by":"publisher","DOI":"10.1109\/TCSI.2017.2735490"},{"key":"ref2","year":"2019","journal-title":"Cisco Global Cloud Index Forecast and Methodology"},{"key":"ref144","doi-asserted-by":"publisher","DOI":"10.14778\/3137628.3137664"},{"key":"ref1","year":"2019","journal-title":"Fog Computing and the Internet of Things Extend the Cloud to Where the Things Are"},{"key":"ref145","doi-asserted-by":"publisher","DOI":"10.1145\/3020078.3021745"},{"key":"ref241","doi-asserted-by":"publisher","DOI":"10.1109\/CISS.2018.8362276"},{"key":"ref242","author":"dulac-arnold","year":"2015","journal-title":"Deep reinforcement learning in large discrete action spaces"},{"key":"ref243","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2017.2682318"},{"key":"ref244","doi-asserted-by":"publisher","DOI":"10.1109\/TNET.2015.2487344"},{"key":"ref240","doi-asserted-by":"publisher","DOI":"10.1109\/WCNC.2017.7925848"},{"key":"ref248","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2019.2903191"},{"key":"ref247","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2018.2839563"},{"key":"ref246","doi-asserted-by":"publisher","DOI":"10.1109\/TCOMM.2018.2866572"},{"key":"ref245","doi-asserted-by":"publisher","DOI":"10.1109\/TCCN.2017.2725277"},{"key":"ref249","doi-asserted-by":"publisher","DOI":"10.1109\/PIMRC.2017.8292514"},{"key":"ref109","doi-asserted-by":"publisher","DOI":"10.1109\/SEC.2018.00016"},{"key":"ref108","doi-asserted-by":"publisher","DOI":"10.1145\/3213344.3213345"},{"key":"ref107","doi-asserted-by":"publisher","DOI":"10.1109\/CIC.2018.00042"},{"key":"ref106","doi-asserted-by":"publisher","DOI":"10.1145\/3132211.3134459"},{"key":"ref105","doi-asserted-by":"publisher","DOI":"10.1145\/3241539.3241559"},{"key":"ref104","doi-asserted-by":"publisher","DOI":"10.1109\/SEC.2016.38"},{"key":"ref103","doi-asserted-by":"publisher","DOI":"10.1109\/TSC.2017.2662008"},{"key":"ref102","doi-asserted-by":"crossref","first-page":"137","DOI":"10.1109\/MNET.2018.1700415","article-title":"Distributed and efficient object detection in edge computing: Challenges and solutions","volume":"32","author":"guo","year":"2018","journal-title":"IEEE Netw"},{"key":"ref111","author":"qi","year":"2018","journal-title":"Vehicular Edge Computing via Deep Reinforcement Learning"},{"key":"ref112","doi-asserted-by":"publisher","DOI":"10.1109\/TVT.2018.2867191"},{"key":"ref110","doi-asserted-by":"publisher","DOI":"10.1109\/TVT.2017.2760281"},{"key":"ref250","doi-asserted-by":"publisher","DOI":"10.1109\/ISWCS.2018.8491089"},{"key":"ref251","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2018.2876279"},{"key":"ref254","doi-asserted-by":"publisher","DOI":"10.1109\/TVT.2018.2890685"},{"key":"ref255","author":"chen","year":"2018","journal-title":"Decentralized Computation Offloading for Multi-User Mobile Edge Computing A Deep Reinforcement Learning Approach"},{"key":"ref252","doi-asserted-by":"publisher","DOI":"10.1109\/ICC.2018.8422743"},{"key":"ref253","doi-asserted-by":"publisher","DOI":"10.1109\/WCNC.2018.8377343"},{"key":"ref257","doi-asserted-by":"publisher","DOI":"10.1109\/TSC.2018.2867482"},{"key":"ref256","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2018.8461641"},{"key":"ref259","doi-asserted-by":"publisher","DOI":"10.1145\/3297280.3297300"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.1016\/j.neunet.2014.09.003"},{"key":"ref258","author":"huang","year":"2018","journal-title":"Deep Reinforcement Learning for Online Offloading in Wireless Powered Mobile-Edge Computing Networks"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/LCOMM.2018.2875978"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1145\/3037697.3037698"},{"key":"ref13","year":"2019","journal-title":"Democratizing AI"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1038\/s41928-018-0195-9"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/3154815"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2017.2685434"},{"key":"ref118","doi-asserted-by":"publisher","DOI":"10.1109\/WCNC.2017.7925564"},{"key":"ref17","doi-asserted-by":"publisher","DOI":"10.1109\/MCOM.2017.1600863"},{"key":"ref117","doi-asserted-by":"publisher","DOI":"10.1109\/MNET.2019.1800235"},{"key":"ref18","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2019.2941458"},{"key":"ref19","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2019.2918951"},{"key":"ref119","doi-asserted-by":"publisher","DOI":"10.1109\/ICNP.2017.8117594"},{"key":"ref114","doi-asserted-by":"publisher","DOI":"10.1016\/j.future.2018.08.006"},{"key":"ref113","doi-asserted-by":"publisher","DOI":"10.1109\/TII.2018.2842821"},{"key":"ref116","doi-asserted-by":"publisher","DOI":"10.1109\/ASAP.2018.8445118"},{"key":"ref115","doi-asserted-by":"publisher","DOI":"10.1145\/2991561.2991575"},{"key":"ref120","author":"tian","year":"2018","journal-title":"A fog robotic system for dynamic visual servoing"},{"key":"ref121","doi-asserted-by":"publisher","DOI":"10.1109\/TCSS.2018.2871694"},{"key":"ref122","doi-asserted-by":"publisher","DOI":"10.1109\/TII.2017.2679740"},{"key":"ref123","doi-asserted-by":"publisher","DOI":"10.1109\/JSYST.2018.2873430"},{"key":"ref260","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2018.2871020"},{"key":"ref261","doi-asserted-by":"publisher","DOI":"10.1109\/MWC.2018.1700291"},{"key":"ref262","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2018.2878435"},{"key":"ref263","author":"nguyen","year":"2018","journal-title":"Secure computation offloading in blockchain based IoT networks with deep reinforcement learning"},{"key":"ref264","first-page":"1","article-title":"Mobile edge computing platform deployment in 4G LTE networks: A middlebox approach","author":"li","year":"2018","journal-title":"Proceedings of USENIX Workshop on Hot Topics in Edge Computing (HotEdge 18)"},{"key":"ref265","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2018.2846401"},{"key":"ref266","doi-asserted-by":"publisher","DOI":"10.1109\/MWC.2017.1600304WC"}],"container-title":["IEEE Communications Surveys &amp; Tutorials"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/9739\/9102343\/08976180.pdf?arnumber=8976180","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,4,27]],"date-time":"2022-04-27T16:58:57Z","timestamp":1651078737000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/8976180\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2020]]},"references-count":292,"journal-issue":{"issue":"2"},"URL":"https:\/\/doi.org\/10.1109\/comst.2020.2970550","relation":{},"ISSN":["1553-877X","2373-745X"],"issn-type":[{"value":"1553-877X","type":"electronic"},{"value":"2373-745X","type":"electronic"}],"subject":[],"published":{"date-parts":[[2020]]}}}