{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,4,22]],"date-time":"2026-04-22T20:55:04Z","timestamp":1776891304612,"version":"3.51.2"},"reference-count":163,"publisher":"Institute of Electrical and Electronics Engineers (IEEE)","license":[{"start":{"date-parts":[[2025,1,1]],"date-time":"2025-01-01T00:00:00Z","timestamp":1735689600000},"content-version":"vor","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by-nc-nd\/4.0\/"}],"funder":[{"name":"European Regional Development Fund of European Union and Greek National Funds through the Operational Program Competitiveness, Entrepreneurship and Innovation, under the Call RESEARCH\u2014CREATE\u2014INNOVATE","award":["T2EDK-01231"],"award-info":[{"award-number":["T2EDK-01231"]}]}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":["IEEE Access"],"published-print":{"date-parts":[[2025]]},"DOI":"10.1109\/access.2025.3535085","type":"journal-article","created":{"date-parts":[[2025,1,27]],"date-time":"2025-01-27T18:55:41Z","timestamp":1738004141000},"page":"30993-31015","source":"Crossref","is-referenced-by-count":8,"title":["A Survey on Parameter Server Architecture: Approaches for Optimizing Distributed Centralized Learning"],"prefix":"10.1109","volume":"13","author":[{"ORCID":"https:\/\/orcid.org\/0009-0000-0931-1479","authenticated-orcid":false,"given":"Nikodimos","family":"Provatas","sequence":"first","affiliation":[{"name":"School of Electrical and Computer Engineering, National Technical University of Athens, Athens, Greece"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"ORCID":"https:\/\/orcid.org\/0000-0002-7142-8106","authenticated-orcid":false,"given":"Ioannis","family":"Konstantinou","sequence":"additional","affiliation":[{"name":"Department of Informatics and Telecommunications, University of Thessaly, Lamia, Greece"}],"role":[{"role":"author","vocabulary":"crossref"}]},{"given":"Nectarios","family":"Koziris","sequence":"additional","affiliation":[{"name":"School of Electrical and Computer Engineering, National Technical University of Athens, Athens, Greece"}],"role":[{"role":"author","vocabulary":"crossref"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1145\/3065386"},{"key":"ref2","article-title":"Very deep convolutional networks for large-scale image recognition","author":"Simonyan","year":"2014","journal-title":"arXiv:1409.1556"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1109\/ICASSP.2013.6638947"},{"key":"ref4","first-page":"173","article-title":"Deep speech 2: End-to-end speech recognition in English and Mandarin","volume-title":"Proc. 33rd Int. Conf. Int. Conf. Mach. Learn.","author":"Amodei"},{"key":"ref5","doi-asserted-by":"publisher","DOI":"10.1007\/978-981-10-5209-5"},{"key":"ref6","doi-asserted-by":"publisher","DOI":"10.1145\/1390156.1390177"},{"key":"ref7","volume-title":"Deep Learning","author":"Goodfellow","year":"2016"},{"key":"ref8","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2014.2325029"},{"key":"ref9","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV.2017.97"},{"key":"ref10","volume-title":"Deep Learning Pipeline: Building a Deep Learning Model With TensorFlow","author":"El-Amir","year":"2019"},{"key":"ref11","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2009.5206848"},{"key":"ref13","article-title":"Symbolic discovery of optimization algorithms","author":"Chen","year":"2023","journal-title":"arXiv:2302.06675"},{"key":"ref14","article-title":"CoCa: Contrastive captioners are image-text foundation models","author":"Yu","year":"2022","journal-title":"arXiv:2205.01917"},{"key":"ref15","first-page":"23965","article-title":"Model soups: Averaging weights of multiple fine-tuned models improves accuracy without increasing inference time","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Wortsman"},{"key":"ref16","article-title":"PaLI: A jointly-scaled multilingual language-image model","author":"Chen","year":"2022","journal-title":"arXiv:2209.06794"},{"key":"ref17","first-page":"3965","article-title":"CoAtNet: Marrying convolution and attention for all data sizes","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Dai"},{"key":"ref18","first-page":"2266","article-title":"Modeling order in neural word embeddings at scale","volume-title":"Proc. 32nd Int. Conf. Int. Conf. Mach. Learn.","author":"Trask"},{"key":"ref19","first-page":"1877","article-title":"Language models are few-shot learners","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"33","author":"Brown"},{"key":"ref20","article-title":"The AI index 2024 annual report","author":"Maslej","year":"2024"},{"key":"ref21","doi-asserted-by":"publisher","DOI":"10.1016\/j.sysarc.2019.101635"},{"key":"ref22","doi-asserted-by":"publisher","DOI":"10.1109\/JSTARS.2019.2959707"},{"key":"ref23","doi-asserted-by":"publisher","DOI":"10.1145\/3320060"},{"key":"ref24","doi-asserted-by":"publisher","DOI":"10.1145\/3343211.3343218"},{"key":"ref25","first-page":"1223","article-title":"Large scale distributed deep networks","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"25","author":"Dean"},{"key":"ref26","first-page":"1","article-title":"Measuring the effects of data parallelism on neural network training","volume":"20","author":"Shallue","year":"2018","journal-title":"J. Mach. Learn. Res."},{"key":"ref27","article-title":"Measuring the limits of data parallel training for neural networks","author":"Shallue","year":"2019","journal-title":"arXiv:1811.03600"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.14778\/1920841.1920931"},{"key":"ref29","first-page":"1","article-title":"Parameter server for distributed machine learning","volume-title":"Proc. Big Learn. NIPS Workshop","volume":"6","author":"Li"},{"key":"ref30","doi-asserted-by":"publisher","DOI":"10.1145\/2640087.2644155"},{"key":"ref31","doi-asserted-by":"publisher","DOI":"10.1145\/2934664"},{"issue":"1","key":"ref32","first-page":"1235","article-title":"MLlib: Machine learning in apache spark","volume":"17","author":"Meng","year":"2016","journal-title":"J. Mach. Learn. Res."},{"key":"ref33","doi-asserted-by":"publisher","DOI":"10.1145\/3357223.3362707"},{"key":"ref34","first-page":"265","article-title":"TensorFlow: Large-scale machine learning on heterogeneous distributed systems","volume-title":"Proc. 12th USENIX Symp. Operating Syst. Design Implement. (OSDI)","author":"Abadi"},{"key":"ref35","first-page":"8026","article-title":"PyTorch: An imperative style, high-performance deep learning library","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Paszke"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.1109\/ETCT.2016.7882969"},{"key":"ref37","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2016.2591592"},{"issue":"2","key":"ref38","doi-asserted-by":"crossref","first-page":"1","DOI":"10.1145\/3377454","article-title":"A survey on distributed machine learning","volume":"53","author":"Verbraeken","year":"2020","journal-title":"ACM Comput. Surv."},{"key":"ref39","doi-asserted-by":"publisher","DOI":"10.1109\/MNET.011.2000530"},{"key":"ref40","doi-asserted-by":"publisher","DOI":"10.1016\/j.jpdc.2020.11.005"},{"key":"ref41","doi-asserted-by":"publisher","DOI":"10.1016\/j.infsof.2008.09.009"},{"key":"ref42","first-page":"571","article-title":"Project adam: Building an efficient and scalable deep learning training system","volume-title":"Proc. 11th USENIX Symp. Operating Syst. Design Implement. (OSDI)","author":"Chilimbi"},{"key":"ref43","doi-asserted-by":"publisher","DOI":"10.1109\/TBDATA.2015.2472014"},{"key":"ref44","volume-title":"TensorFlow: Large-scale machine learning on heterogeneous systems","author":"Abadi","year":"2015"},{"key":"ref45","volume-title":"Distributed Communication Package-torch.Distributed\u2014PyTorch 1.6.0 Documentation","year":"2024"},{"key":"ref46","article-title":"MXNet: A flexible and efficient machine learning library for heterogeneous distributed systems","author":"Chen","year":"2015","journal-title":"arXiv:1512.01274"},{"key":"ref47","doi-asserted-by":"publisher","DOI":"10.1145\/3394486.3406703"},{"key":"ref48","first-page":"1","article-title":"ZeRO++: Extremely efficient collective communication for giant model training","volume-title":"Proc. ICLR","author":"Wang"},{"key":"ref49","doi-asserted-by":"publisher","DOI":"10.5555\/3291168.3291210"},{"key":"ref50","doi-asserted-by":"publisher","DOI":"10.1109\/SC41405.2020.00024"},{"key":"ref51","article-title":"Horovod: Fast and easy distributed deep learning in tensorflow","author":"Sergeev","year":"2018","journal-title":"arXiv:1802.05799"},{"key":"ref52","doi-asserted-by":"publisher","DOI":"10.1145\/3219819.3219837"},{"key":"ref53","doi-asserted-by":"publisher","DOI":"10.1109\/ICDM.2016.0028"},{"key":"ref54","doi-asserted-by":"publisher","DOI":"10.1145\/79173.79181"},{"key":"ref55","doi-asserted-by":"publisher","DOI":"10.1145\/1327452.1327492"},{"key":"ref56","first-page":"281","article-title":"Map-reduce for machine learning on multicore","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Sch\u00f6lkopf"},{"key":"ref57","doi-asserted-by":"publisher","DOI":"10.1145\/293347.293351"},{"key":"ref58","doi-asserted-by":"publisher","DOI":"10.1038\/323533a0"},{"key":"ref59","doi-asserted-by":"publisher","DOI":"10.1214\/aos\/1176345513"},{"key":"ref60","doi-asserted-by":"publisher","DOI":"10.2307\/2346830"},{"key":"ref61","first-page":"1231","article-title":"Efficient large-scale distributed training of conditional maximum entropy models","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"22","author":"McDonald"},{"key":"ref62","first-page":"2595","article-title":"Parallelized stochastic gradient descent","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"23","author":"Zinkevich"},{"key":"ref63","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2017.2696365"},{"key":"ref64","first-page":"693","article-title":"Hogwild: A lock-free approach to parallelizing stochastic gradient descent","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Recht"},{"key":"ref65","first-page":"1","article-title":"Asynchronous parallel stochastic gradient for nonconvex optimization","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"28","author":"Lian"},{"key":"ref66","first-page":"1223","article-title":"More effective distributed ML via a stale synchronous parallel parameter Server","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Ho"},{"key":"ref67","doi-asserted-by":"publisher","DOI":"10.1145\/2987550.2987554"},{"key":"ref68","first-page":"181","article-title":"Poseidon: An efficient communication architecture for distributed deep learning on $GPU$ clusters","volume-title":"Proc. USENIX Annu. Tech. Conf. (USENIX ATC)","author":"Zhang"},{"key":"ref69","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS.2019.00028"},{"key":"ref70","doi-asserted-by":"publisher","DOI":"10.1109\/TC.2020.2974461"},{"key":"ref71","first-page":"1","article-title":"Communication-efficient distributed blockwise momentum SGD with error-feedback","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Zheng"},{"key":"ref72","first-page":"6155","article-title":"DoubleSqueeze: Parallel stochastic gradient descent with double-pass error-compensated compression","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Tang"},{"key":"ref73","first-page":"418","article-title":"TicTac: Accelerating distributed deep learning with communication scheduling","volume-title":"Proc. Mach. Learn. Syst.","author":"Hashemi"},{"key":"ref74","first-page":"13551","article-title":"ScaleCom: Scalable sparsified gradient compression for communication-efficient distributed training","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Chen"},{"key":"ref75","doi-asserted-by":"publisher","DOI":"10.1145\/3190508.3190517"},{"key":"ref76","doi-asserted-by":"publisher","DOI":"10.1145\/3187009.3177734"},{"key":"ref77","doi-asserted-by":"publisher","DOI":"10.1145\/3267809.3267840"},{"key":"ref78","doi-asserted-by":"publisher","DOI":"10.14778\/3407790.3407796"},{"key":"ref79","doi-asserted-by":"publisher","DOI":"10.14778\/3476311.3476325"},{"key":"ref80","doi-asserted-by":"publisher","DOI":"10.1145\/3514221.3517860"},{"key":"ref81","first-page":"19","article-title":"Communication efficient distributed machine learning with the parameter server","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"27","author":"Li"},{"key":"ref82","doi-asserted-by":"publisher","DOI":"10.1145\/2806777.2806778"},{"key":"ref83","doi-asserted-by":"publisher","DOI":"10.1145\/3097983.3098029"},{"key":"ref84","doi-asserted-by":"publisher","DOI":"10.1145\/3041021.3054225"},{"key":"ref85","doi-asserted-by":"publisher","DOI":"10.1109\/BigData.2018.8622358"},{"key":"ref86","article-title":"Staleness-aware async-SGD for distributed deep learning","author":"Zhang","year":"2015","journal-title":"arXiv:1511.05950"},{"key":"ref87","first-page":"803","article-title":"Slow and stale gradients can win the race: Error-runtime trade-offs in distributed SGD","volume-title":"Proc. Int. Conf. Artif. Intell. Statist.","author":"Dutta"},{"key":"ref88","first-page":"631","article-title":"Litz: Elastic framework for high-performance distributed machine learning","volume-title":"Proc. USENIX Annu. Tech. Conf. (USENIX ATC)","author":"Qiao"},{"key":"ref89","doi-asserted-by":"publisher","DOI":"10.1145\/3337821.3337828"},{"key":"ref90","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS.2019.00150"},{"key":"ref91","doi-asserted-by":"publisher","DOI":"10.1145\/3419111.3421296"},{"key":"ref92","first-page":"4447","article-title":"Sparsified SGD with memory","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"31","author":"Stich"},{"key":"ref93","first-page":"1508","article-title":"TernGrad: Ternary gradients to reduce communication in distributed deep learning","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"30","author":"Wen"},{"key":"ref94","first-page":"1","article-title":"QSGD: Communication-efficient SGD via gradient quantization and encoding","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Alistarh"},{"key":"ref95","first-page":"1","article-title":"ATOMO: Communication-efficient learning via atomic sparsification","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","volume":"31","author":"Wang"},{"key":"ref96","doi-asserted-by":"publisher","DOI":"10.1145\/3234944.3234978"},{"key":"ref97","first-page":"196","article-title":"Communication-efficient distributed optimization with quantized preconditioners","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Alimisis"},{"key":"ref98","first-page":"1","article-title":"Communication-efficient distributed SGD with sketching","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Ivkin"},{"key":"ref99","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2021.3084806"},{"key":"ref100","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS51616.2021.00060"},{"key":"ref101","first-page":"132","article-title":"Priority-based parameter propagation for distributed DNN training","volume-title":"Proc. Mach. Learn. Syst.","author":"Jayarajan"},{"key":"ref102","doi-asserted-by":"publisher","DOI":"10.1109\/TPDS.2021.3062721"},{"key":"ref103","doi-asserted-by":"publisher","DOI":"10.1109\/SC41405.2020.00048"},{"key":"ref104","doi-asserted-by":"publisher","DOI":"10.14778\/3137628.3137649"},{"key":"ref105","doi-asserted-by":"publisher","DOI":"10.1145\/3419394.3423637"},{"key":"ref106","doi-asserted-by":"publisher","DOI":"10.1145\/3419111.3421307"},{"key":"ref107","first-page":"185","article-title":"Effective straggler mitigation: Attack of the clones","volume-title":"Proc. 10th USENIX Symp. Networked Syst. Design Implement. (NSDI)","author":"Ananthanarayanan"},{"key":"ref108","doi-asserted-by":"publisher","DOI":"10.1145\/3419111.3421299"},{"key":"ref109","first-page":"2331","article-title":"Slow learners are fast","volume-title":"Proc. 22nd Int. Conf. Neural Inf. Process. Syst.","author":"Langford"},{"key":"ref110","doi-asserted-by":"publisher","DOI":"10.1145\/3035918.3035933"},{"key":"ref111","doi-asserted-by":"publisher","DOI":"10.1109\/HPCC\/SmartCity\/DSS.2019.00042"},{"key":"ref112","doi-asserted-by":"publisher","DOI":"10.14778\/2536222.2536229"},{"key":"ref113","first-page":"1","article-title":"Key management in distributed systems","volume":"2010","author":"Acar","year":"2010","journal-title":"Microsoft Res."},{"key":"ref114","first-page":"453","article-title":"Practical hardening of crash-tolerant systems","volume-title":"Proc. USENIX Annu. Tech. Conf. (USENIX ATC)","author":"Correia"},{"key":"ref115","doi-asserted-by":"publisher","DOI":"10.1109\/PADSW.2018.8644613"},{"key":"ref116","first-page":"1273","article-title":"Communication-efficient learning of deep networks from decentralized data","volume-title":"Proc. Artif. Intell. Statist.","author":"McMahan"},{"key":"ref117","article-title":"Federated learning: Strategies for improving communication efficiency","author":"Kone\u010d n\u00fd","year":"2016","journal-title":"arXiv:1610.05492"},{"key":"ref118","doi-asserted-by":"publisher","DOI":"10.1016\/j.neucom.2020.08.034"},{"key":"ref119","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52729.2023.01566"},{"key":"ref120","first-page":"374","article-title":"Towards federated learning at scale: System design","volume-title":"Proc. Mach. Learn. Syst.","author":"Bonawitz"},{"key":"ref121","doi-asserted-by":"publisher","DOI":"10.1609\/aaai.v37i7.26023"},{"key":"ref122","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2023.3271859"},{"key":"ref123","article-title":"Personalized federated learning: A meta-learning approach","author":"Fallah","year":"2020","journal-title":"arXiv:2002.07948"},{"key":"ref124","first-page":"39801","article-title":"Personalized federated learning with inferred collaboration graphs","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Ye"},{"key":"ref125","doi-asserted-by":"publisher","DOI":"10.1109\/ICCV51070.2023.01755"},{"key":"ref126","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1706.06083"},{"key":"ref127","doi-asserted-by":"publisher","DOI":"10.1109\/TCE.2023.3315415"},{"key":"ref128","doi-asserted-by":"publisher","DOI":"10.1109\/TNET.2023.3302016"},{"key":"ref129","doi-asserted-by":"publisher","DOI":"10.1016\/j.knosys.2021.106775"},{"key":"ref130","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2020.2986024"},{"key":"ref131","doi-asserted-by":"publisher","DOI":"10.1109\/COMST.2021.3075439"},{"key":"ref132","doi-asserted-by":"publisher","DOI":"10.1109\/TKDE.2021.3124599"},{"key":"ref133","doi-asserted-by":"publisher","DOI":"10.1007\/s10115-022-01664-x"},{"key":"ref134","doi-asserted-by":"publisher","DOI":"10.1145\/3460427"},{"key":"ref135","doi-asserted-by":"publisher","DOI":"10.1016\/j.future.2020.10.007"},{"key":"ref136","doi-asserted-by":"publisher","DOI":"10.1145\/3501296"},{"key":"ref137","doi-asserted-by":"publisher","DOI":"10.1109\/TNNLS.2022.3160699"},{"key":"ref138","doi-asserted-by":"publisher","DOI":"10.1145\/3524104"},{"key":"ref139","doi-asserted-by":"publisher","DOI":"10.1007\/s13042-022-01647-y"},{"key":"ref140","doi-asserted-by":"publisher","DOI":"10.1145\/3625558"},{"key":"ref141","doi-asserted-by":"publisher","DOI":"10.1145\/3575810"},{"key":"ref142","first-page":"1","article-title":"Efficient large-scale language model training on GPU clusters using megatron-LM","volume-title":"Proc. Int. Conf. High Perform. Comput., Netw., Storage Anal. (SC21)","author":"Narayanan"},{"key":"ref143","first-page":"745","article-title":"MegaScale: Scaling large language model training to more than 10,000 GPUs","volume-title":"Proc. 21st USENIX Symp. Networked Syst. Design Implement. (NSDI)","author":"Jiang"},{"key":"ref144","doi-asserted-by":"publisher","DOI":"10.1145\/3065386"},{"key":"ref145","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/D16-1264"},{"key":"ref146","first-page":"1","article-title":"SuperGLUE: A stickier benchmark for general-purpose language understanding systems","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Wang"},{"key":"ref147","doi-asserted-by":"publisher","DOI":"10.1111\/nyas.15007"},{"key":"ref148","article-title":"HEMM: Holistic evaluation of multimodal foundation models","author":"Pu Liang","year":"2024","journal-title":"arXiv:2407.03418"},{"key":"ref149","first-page":"1","article-title":"Measuring massive multitask language understanding","volume-title":"Proc. Int. Conf. Learn. Represent.","author":"Hendrycks"},{"key":"ref150","first-page":"1","article-title":"Chatbot arena: An open platform for evaluating LLMs by human preference","volume-title":"Proc. 41st Int. Conf. Mach. Learn.","author":"Chiang"},{"key":"ref151","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2022.acl-long.229"},{"key":"ref152","doi-asserted-by":"publisher","DOI":"10.18653\/v1\/2023.emnlp-main.397"},{"key":"ref153","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR52733.2024.00913"},{"key":"ref154","first-page":"1","article-title":"Understanding social reasoning in language models with language models","volume-title":"Proc. Adv. Neural Inf. Process. Syst.","author":"Gandhi"},{"key":"ref155","first-page":"1","article-title":"AgentBench: Evaluating LLMs as agents","volume-title":"Proc. 12th Int. Conf. Learn. Represent.","author":"Liu"},{"key":"ref156","doi-asserted-by":"publisher","DOI":"10.1145\/3694715.3695970"},{"key":"ref157","doi-asserted-by":"publisher","DOI":"10.1109\/ACCESS.2024.3409745"},{"key":"ref158","first-page":"119","article-title":"Zeus: Understanding and optimizing GPU energy consumption of DNN training","volume-title":"Proc. 20th USENIX Symp. Networked Syst. Design Implement. (NSDI)","author":"You"},{"key":"ref159","first-page":"851","article-title":"EnvPipe: Performance-preserving DNN atraining framework for saving energy","volume-title":"Proc. USENIX Annu. Tech. Conf. (USENIX ATC 23)","author":"Choi"},{"key":"ref160","article-title":"An overview of gradient descent optimization algorithms","author":"Ruder","year":"2016","journal-title":"arXiv:1609.04747"},{"key":"ref161","article-title":"Adam: A method for stochastic optimization","author":"Kingma","year":"2014","journal-title":"arXiv:1412.6980"},{"key":"ref162","article-title":"Rmsprop and equilibrated adaptive learning rates for nonconvex optimization","author":"Bengio","year":"2015","journal-title":"arXiv:1502.04390"},{"key":"ref163","first-page":"1139","article-title":"On the importance of initialization and momentum in deep learning","volume-title":"Proc. Int. Conf. Mach. Learn.","author":"Sutskever"}],"container-title":["IEEE Access"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/6287639\/10820123\/10855445.pdf?arnumber=10855445","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2025,2,20]],"date-time":"2025-02-20T20:48:23Z","timestamp":1740084503000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/10855445\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025]]},"references-count":163,"URL":"https:\/\/doi.org\/10.1109\/access.2025.3535085","relation":{},"ISSN":["2169-3536"],"issn-type":[{"value":"2169-3536","type":"electronic"}],"subject":[],"published":{"date-parts":[[2025]]}}}