{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,7,15]],"date-time":"2026-07-15T07:14:58Z","timestamp":1784099698436,"version":"3.55.0"},"reference-count":145,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","issue":"4","license":[{"start":{"date-parts":[[2021,8,1]],"date-time":"2021-08-01T00:00:00Z","timestamp":1627776000000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2021,8,1]],"date-time":"2021-08-01T00:00:00Z","timestamp":1627776000000},"content-version":"am","delay-in-days":0,"URL":"https:\/\/ieeexplore.ieee.org\/Xplorehelp\/downloads\/license-information\/IEEE.html"},{"start":{"date-parts":[[2021,8,1]],"date-time":"2021-08-01T00:00:00Z","timestamp":1627776000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2021,8,1]],"date-time":"2021-08-01T00:00:00Z","timestamp":1627776000000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"funder":[{"name":"NSF","award":["#2007832"],"award-info":[{"award-number":["#2007832"]}]},{"name":"IBMIllinois Center for Cognitive Computing System Research C3SR"},{"name":"Semiconductor Research Corporation SRC","award":["Task 2803.001\/2804.001"],"award-info":[{"award-number":["Task 2803.001\/2804.001"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Des. Test"],"published-print":{"date-parts":[[2021,8]]},"DOI":"10.1109\/mdat.2021.3069952","type":"journal-article","created":{"date-parts":[[2021,3,31]],"date-time":"2021-03-31T19:35:20Z","timestamp":1617219320000},"page":"7-26","source":"Crossref","is-referenced-by-count":43,"title":["Enabling Design Methodologies and Future Trends for Edge AI: Specialization and Codesign"],"prefix":"10.1109","volume":"38","author":[{"given":"Cong","family":"Hao","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jordan","family":"Dotzel","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Jinjun","family":"Xiong","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Luca","family":"Benini","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Zhiru","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Deming","family":"Chen","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"263","reference":[{"key":"ref39","first-page":"1","article-title":"TinyTL: Reduce memory, not parameters for efficient on-device learning","author":"cai","year":"2020","journal-title":"Proc Annu Conf Neural Inf Process Syst (NeurIPS)"},{"key":"ref38","first-page":"500","article-title":"HALO: Hardware-aware learning to optimize","author":"li","year":"2020","journal-title":"Proc Eur Conf Comput Vis"},{"key":"ref33","article-title":"Why compute matters for UAV energy efficiency?","author":"boroujerdian","year":"2018","journal-title":"Proc 2nd Int Symp Aerial Robot"},{"key":"ref32","first-page":"1","article-title":"Challenges for building a cloud native scalable and trustable multi-tenant AIoT platform","author":"xiong","year":"2020","journal-title":"Proc 39th Int Conf Comput -Aided Design (ICCAD)"},{"key":"ref31","year":"0","journal-title":"Apache t"},{"key":"ref30","year":"0","journal-title":"Xilinx Edge AI platform"},{"key":"ref37","article-title":"Ultra-low precision 4-bit training of deep neural networks","volume":"33","author":"sun","year":"2020","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref36","first-page":"4900","article-title":"Hybrid 8-bit floating point (HFP8) training and inference for deep neural networks","author":"sun","year":"2019","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref35","article-title":"FracTrain: Fractionally squeezing bit savings both temporally and spatially for efficient DNN training","volume":"33","author":"fu","year":"2020","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref34","first-page":"5138","article-title":"E2-train: Training state-of-the-art CNNs with over 80% energy savings","author":"wang","year":"2019","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref28","year":"0","journal-title":"Nvidia tensorrt"},{"key":"ref27","year":"0","journal-title":"Arm nn sdk"},{"key":"ref29","year":"0","journal-title":"Intel OpenVINO Toolkit"},{"key":"ref20","year":"0","journal-title":"Lattice iCE40 UltraPlus FPGAs"},{"key":"ref22","year":"0","journal-title":"Ascend 310 AI Processor"},{"key":"ref21","year":"0","journal-title":"Google Edge TPU"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1109\/ISCAS45731.2020.9181003"},{"key":"ref23","first-page":"494","article-title":"A 65nm 1Mb nonvolatile computing-in-memory ReRAM macro with sub-16ns multiply-and-accumulate for binary DNN AI edge processors","author":"chen","year":"2018","journal-title":"IEEE Int Solid-State Circuits Conf (ISSCC) Dig Tech Papers"},{"key":"ref101","article-title":"CuDNN: Efficient primitives for deep learning","author":"chetlur","year":"2014","journal-title":"Arxiv 1410 0759"},{"key":"ref26","year":"0","journal-title":"TensorFlow Lite"},{"key":"ref100","year":"0","journal-title":"MLCommons"},{"key":"ref25","doi-asserted-by":"publisher","DOI":"10.1080\/23746149.2016.1259585"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1109\/TPAMI.2020.2975796"},{"key":"ref51","article-title":"EEG-TCNet: An accurate temporal convolutional network for embedded motor-imagery brain-machine interfaces","author":"mar ingolfsson","year":"2020","journal-title":"arXiv 2006 00622"},{"key":"ref59","first-page":"70","article-title":"Anti-bandit neural architecture search for model defense","author":"chen","year":"2020","journal-title":"Proc Eur Conf Comput Vis"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2020.acl-main.686"},{"key":"ref57","first-page":"216","article-title":"SkyNet: A hardware-efficient method for object detection and tracking on embedded systems","author":"zhang","year":"2020","journal-title":"Proc Mach Learn Syst"},{"key":"ref56","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01249-6_18"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.00293"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1109\/CVPRW.2018.00215"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR42600.2020.01204"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2019.01156"},{"key":"ref40","first-page":"1273","article-title":"Communication-efficient learning of deep networks from decentralized data","author":"mcmahan","year":"2017","journal-title":"Proc 20th Int Conf Artif Intell Statist"},{"key":"ref4","article-title":"Edge intelligence: Architectures, challenges, and applications","author":"xu","year":"2020","journal-title":"arXiv 2003 12172"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2020.2970550"},{"key":"ref6","year":"0","journal-title":"Raspberry Pi Series Processors"},{"key":"ref5","year":"0","journal-title":"Arm Cortex-M Series Processors"},{"key":"ref8","year":"0","journal-title":"Sipeed MAIX Series"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.1007\/978-3-030-01264-9_8"},{"key":"ref7","year":"0","journal-title":"NanoPi Series Processors"},{"key":"ref9","year":"0","journal-title":"Nvidia Jetson Embedded Systems"},{"key":"ref46","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00474"},{"key":"ref45","article-title":"MobileNets: Efficient convolutional neural networks for mobile vision applications","author":"howard","year":"2017","journal-title":"arXiv 1704 04861"},{"key":"ref48","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2018.00716"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2019.00140"},{"key":"ref42","doi-asserted-by":"publisher","DOI":"10.1145\/3298981"},{"key":"ref41","article-title":"Towards federated learning at scale: System design","author":"bonawitz","year":"2019","journal-title":"arXiv 1902 01046"},{"key":"ref44","article-title":"SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and < 0.5MB model size","author":"iandola","year":"2016","journal-title":"arXiv 1602 07360"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/ISCA.2018.00069"},{"key":"ref127","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v34i04.5954"},{"key":"ref126","first-page":"1","article-title":"Co-exploration of neural architectures and heterogeneous ASIC accelerator designs targeting multiple tasks","author":"yang","year":"2020","journal-title":"Proc 57th ACM\/IEEE Design Autom Conf (DAC)"},{"key":"ref125","doi-asserted-by":"publisher","DOI":"10.1109\/DAC18072.2020.9218749"},{"key":"ref124","doi-asserted-by":"publisher","DOI":"10.1109\/ICCAD45719.2019.8942055"},{"key":"ref73","article-title":"AdaScale: Towards real-time video object detection using adaptive scaling","author":"chin","year":"2019","journal-title":"arXiv 1902 02910"},{"key":"ref72","first-page":"1886","article-title":"Channel gating neural networks","author":"hua","year":"2019","journal-title":"Proc Adv Neural Inf Process Syst"},{"key":"ref129","article-title":"CoCoPIE: Making mobile AI sweet As PIE&#x2013;compression-compilation co-design goes a long way","author":"liu","year":"2020","journal-title":"Commun ACM"},{"key":"ref71","doi-asserted-by":"publisher","DOI":"10.1145\/3366636"},{"key":"ref128","doi-asserted-by":"publisher","DOI":"10.1145\/3373376.3378534"},{"key":"ref70","article-title":"PareCO: Pareto-aware channel optimization for slimmable neural networks","author":"chin","year":"2020","journal-title":"arXiv 2007 11752"},{"key":"ref76","doi-asserted-by":"publisher","DOI":"10.1145\/3352460.3358283"},{"key":"ref130","article-title":"MCUNet: Tiny deep learning on IoT devices","author":"lin","year":"2020","journal-title":"Proc Annu Conf Neural Inf Process Syst (NeurIPS)"},{"key":"ref77","doi-asserted-by":"publisher","DOI":"10.1109\/SMARTCOMP50058.2020.00065"},{"key":"ref74","article-title":"Precision gating: Improving neural network efficiency with dynamic dual-precision activations","author":"zhang","year":"2020","journal-title":"Proc Int Conf Learn Represent (ICLR)"},{"key":"ref75","first-page":"171","article-title":"FracBNN: Accurate and FPGA-efficient binary neural networks with fractional activations","author":"zhang","year":"2020","journal-title":"Proc Int Conf Learn Represent (ICLR)"},{"key":"ref133","doi-asserted-by":"publisher","DOI":"10.1109\/MM.2017.3711645"},{"key":"ref134","doi-asserted-by":"publisher","DOI":"10.1109\/MDAT.2017.2750907"},{"key":"ref131","year":"0","journal-title":"PULP&#x2014;An Open Parallel Ultra-Low-Power Processing-Platform"},{"key":"ref78","doi-asserted-by":"publisher","DOI":"10.1088\/1741-2552\/aace8c"},{"key":"ref132","doi-asserted-by":"publisher","DOI":"10.1016\/j.sse.2015.11.015"},{"key":"ref79","doi-asserted-by":"publisher","DOI":"10.3390\/jlpea10020019"},{"key":"ref136","doi-asserted-by":"publisher","DOI":"10.1109\/ASAP.2018.8445101"},{"key":"ref135","doi-asserted-by":"publisher","DOI":"10.1109\/ESSCIRC.2018.8494247"},{"key":"ref138","doi-asserted-by":"publisher","DOI":"10.1098\/rsta.2019.0155"},{"key":"ref137","doi-asserted-by":"publisher","DOI":"10.23919\/DATE48585.2020.9116529"},{"key":"ref60","article-title":"NASS: Optimizing secure inference via neural architecture search","author":"bian","year":"2020","journal-title":"arXiv 2001 11854"},{"key":"ref139","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2020.2976702"},{"key":"ref62","doi-asserted-by":"publisher","DOI":"10.1109\/TC.2020.2995593"},{"key":"ref61","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/P19-1355"},{"key":"ref63","first-page":"7543","article-title":"Improving neural network quantization without retraining using outlier channel splitting","author":"zhao","year":"2019","journal-title":"Proc ICML"},{"key":"ref64","article-title":"Memory-driven mixed low precision quantization for enabling deep network inference on microcontrollers","author":"rusci","year":"2019","journal-title":"arXiv 1905 13082"},{"key":"ref140","doi-asserted-by":"publisher","DOI":"10.1109\/MDT.2009.69"},{"key":"ref65","doi-asserted-by":"publisher","DOI":"10.1109\/TCSII.2020.2983648"},{"key":"ref141","doi-asserted-by":"publisher","DOI":"10.1073\/pnas.1915768117"},{"key":"ref66","doi-asserted-by":"publisher","DOI":"10.3850\/9783981537079_0819"},{"key":"ref142","year":"0","journal-title":"AI and Compute"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.1109\/TCAD.2018.2857338"},{"key":"ref143","year":"0","journal-title":"IBM Quamtum Computing"},{"key":"ref68","doi-asserted-by":"publisher","DOI":"10.1145\/3240765.3240796"},{"key":"ref144","year":"0","journal-title":"IBM Research Quantum Computing for Machine Learning"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2019.2918951"},{"key":"ref69","article-title":"Slimmable neural networks","author":"yu","year":"2018","journal-title":"Proc Int Conf Learn Represent"},{"key":"ref145","first-page":"1","article-title":"A co-design framework of neural networks and quantum circuits towards quantum advantage","volume":"12","author":"jiang","year":"2021","journal-title":"Nature Commun"},{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2016.2579198"},{"key":"ref109","doi-asserted-by":"publisher","DOI":"10.1145\/3061639.3062195"},{"key":"ref95","doi-asserted-by":"publisher","DOI":"10.1109\/HPCA47549.2020.00062"},{"key":"ref108","doi-asserted-by":"publisher","DOI":"10.1109\/ICCAD.2015.7372592"},{"key":"ref94","article-title":"ChewBaccaNN: A flexible 223 TOPS\/W BNN accelerator","author":"andri","year":"2020","journal-title":"arXiv 2005 07137"},{"key":"ref107","doi-asserted-by":"publisher","DOI":"10.1145\/3373087.3375306"},{"key":"ref93","doi-asserted-by":"publisher","DOI":"10.1109\/JSSC.2016.2616357"},{"key":"ref106","doi-asserted-by":"publisher","DOI":"10.1109\/FCCM.2019.00020"},{"key":"ref92","doi-asserted-by":"publisher","DOI":"10.1145\/2996864"},{"key":"ref105","article-title":"Automated design space exploration for optimised deployment of DNN on arm Cortex-A CPUs","author":"de prado","year":"2020","journal-title":"arXiv 2006 05181"},{"key":"ref91","doi-asserted-by":"crossref","DOI":"10.1109\/JETCAS.2020.3014503","article-title":"Optimizing temporal convolutional network inference on FPGA-based accelerators","author":"carreras","year":"2020"},{"key":"ref104","year":"2018","journal-title":"CUDA Profiling Tools Interface"},{"key":"ref90","doi-asserted-by":"publisher","DOI":"10.1109\/DAC18072.2020.9218684"},{"key":"ref103","first-page":"125","author":"chen","year":"2013","journal-title":"Reservoir simulation on NVIDIA Tesla GPUs"},{"key":"ref102","year":"2020","journal-title":"Intel VTune"},{"key":"ref111","author":"kara","year":"2021","journal-title":"DAC-SDC&#x2019;18 2nd Place Winner in FPGA Track"},{"key":"ref112","author":"hao","year":"2021","journal-title":"DAC-SDC&#x2019;18 3rd Place Winner in FPGA Track"},{"key":"ref110","year":"0","journal-title":"2020 Design Automation Conference System Design Contest (DAC-SDC)"},{"key":"ref98","first-page":"414","article-title":"The design and implementation of a scalable deep learning benchmarking platform","author":"li","year":"2020","journal-title":"Proc IEEE 13th Int Conf Cloud Comput (CLOUD)"},{"key":"ref99","doi-asserted-by":"publisher","DOI":"10.1109\/IPDPS47924.2020.00042"},{"key":"ref96","article-title":"Non-structured DNN weight pruning&#x2014;Is it beneficial in any platform?","author":"ma","year":"2019","journal-title":"arXiv 1907 02124"},{"key":"ref97","article-title":"MLModelScope: Evaluate and measure ML models within AI pipelines","author":"dakkak","year":"2018","journal-title":"arXiv 1811 09737"},{"key":"ref10","year":"0","journal-title":"AMD Ryzen Embedded Family"},{"key":"ref11","year":"0","journal-title":"Arm Mali GPUs"},{"key":"ref12","first-page":"1","article-title":"Pushing the limits of narrow precision inferencing at cloud scale with Microsoft floating point","author":"rouhani","year":"2020","journal-title":"Proc NeurIPS"},{"key":"ref13","article-title":"Are FPGAs suitable for edge computing?","author":"biookaghazadeh","year":"2018","journal-title":"Proc USENIX Workshop Hot Topics Edge Comput (HotEdge)"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1145\/3020078.3021745"},{"key":"ref15","doi-asserted-by":"publisher","DOI":"10.1145\/3240765.3240801"},{"key":"ref118","author":"weixiong","year":"2021","journal-title":"DAC-SDC&#x2019;18 2nd Place Winner in FPGA Track"},{"key":"ref16","doi-asserted-by":"publisher","DOI":"10.1145\/3020078.3021740"},{"key":"ref82","doi-asserted-by":"publisher","DOI":"10.1109\/FCCM.2011.29"},{"key":"ref117","author":"liu","year":"2021","journal-title":"DAC-SDC&#x2019;20 3rd Place Winner in FPGA Track"},{"key":"ref17","year":"0","journal-title":"Xilinx zynq-7000 fpgas"},{"key":"ref81","doi-asserted-by":"publisher","DOI":"10.1145\/3194554.3194597"},{"key":"ref18","year":"0","journal-title":"Xilinx Spartan-7 FPGAs"},{"key":"ref84","doi-asserted-by":"publisher","DOI":"10.1109\/SASP.2009.5226333"},{"key":"ref119","author":"zhang","year":"2021","journal-title":"DAC-SDC&#x2019;19 1st Place Winner in GPU Track"},{"key":"ref19","year":"0","journal-title":"Intel Arria 10 FPGAs"},{"key":"ref83","doi-asserted-by":"publisher","DOI":"10.1109\/ASICON.2011.6157401"},{"key":"ref114","author":"kara","year":"2021","journal-title":"DAC-SDC&#x2019;19 3rd Place Winner in FPGA Track"},{"key":"ref113","author":"zeng","year":"2021","journal-title":"DAC-SDC&#x2019;18 1st Place Winner in FPGA Track"},{"key":"ref116","author":"hao","year":"2021","journal-title":"DAC-SDC&#x2019;18 1st Place Winner in FPGA Track"},{"key":"ref80","first-page":"894","article-title":"Machine learning on FPGAs to face the IoT revolution","author":"zhang","year":"2017","journal-title":"Proc IEEE\/ACM Int Conf Comput -Aided Design (ICCAD)"},{"key":"ref115","author":"zhao","year":"2021","journal-title":"DAC-SDC&#x2019;18 2nd Place Winner in FPGA Track"},{"key":"ref120","author":"zhan","year":"2021","journal-title":"DAC-SDC&#x2019;18 1st Place Winner in FPGA Track"},{"key":"ref89","doi-asserted-by":"publisher","DOI":"10.1109\/HPEC.2019.8916419"},{"key":"ref121","doi-asserted-by":"publisher","DOI":"10.1109\/ICSICT.2018.8564956"},{"key":"ref122","doi-asserted-by":"publisher","DOI":"10.1145\/3316781.3317829"},{"key":"ref123","doi-asserted-by":"publisher","DOI":"10.1145\/3316781.3317757"},{"key":"ref85","doi-asserted-by":"publisher","DOI":"10.1109\/TVLSI.2009.2013353"},{"key":"ref86","doi-asserted-by":"publisher","DOI":"10.1145\/3020078.3021741"},{"key":"ref87","doi-asserted-by":"publisher","DOI":"10.1109\/ISVLSI.2019.00012"},{"key":"ref88","doi-asserted-by":"publisher","DOI":"10.1109\/FPL.2018.00035"}],"container-title":["IEEE Design &amp; Test"],"original-title":[],"link":[{"URL":"https:\/\/ieeexplore.ieee.org\/ielam\/6221038\/9499233\/9391712-aam.pdf","content-type":"application\/pdf","content-version":"am","intended-application":"syndication"},{"URL":"http:\/\/xplorestaging.ieee.org\/ielx7\/6221038\/9499233\/09391712.pdf?arnumber=9391712","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2022,5,10]],"date-time":"2022-05-10T14:53:42Z","timestamp":1652194422000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/9391712\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2021,8]]},"references-count":145,"journal-issue":{"issue":"4"},"URL":"https:\/\/doi.org\/10.1109\/mdat.2021.3069952","relation":{},"ISSN":["2168-2356","2168-2364"],"issn-type":[{"value":"2168-2356","type":"print"},{"value":"2168-2364","type":"electronic"}],"subject":[],"published":{"date-parts":[[2021,8]]}}}