{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2025,12,3]],"date-time":"2025-12-03T18:14:16Z","timestamp":1764785656545,"version":"3.44.0"},"reference-count":48,"publisher":"Elsevier BV","license":[{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/tdm\/userlicense\/1.0\/"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"tdm","delay-in-days":0,"URL":"https:\/\/www.elsevier.com\/legal\/tdmrep-license"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-017"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-012"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2026,1,1]],"date-time":"2026-01-01T00:00:00Z","timestamp":1767225600000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-004"}],"content-domain":{"domain":["elsevier.com","sciencedirect.com"],"crossmark-restriction":true},"short-container-title":["Future Generation Computer Systems"],"published-print":{"date-parts":[[2026,1]]},"DOI":"10.1016\/j.future.2025.107983","type":"journal-article","created":{"date-parts":[[2025,7,2]],"date-time":"2025-07-02T02:24:19Z","timestamp":1751423059000},"page":"107983","update-policy":"https:\/\/doi.org\/10.1016\/elsevier_cm_policy","source":"Crossref","is-referenced-by-count":1,"special_numbering":"C","title":["Approximate Gradient Synchronization With Adaptive Quantized Gradient Broadcast"],"prefix":"10.1016","volume":"174","author":[{"ORCID":"https:\/\/orcid.org\/0000-0002-4041-3681","authenticated-orcid":false,"given":"Shouxi","family":"Luo","sequence":"first","affiliation":[]},{"given":"Xue","family":"Liu","sequence":"additional","affiliation":[]},{"given":"Ke","family":"Li","sequence":"additional","affiliation":[]},{"ORCID":"https:\/\/orcid.org\/0000-0002-6345-7265","authenticated-orcid":false,"given":"Huanlai","family":"Xing","sequence":"additional","affiliation":[]},{"given":"Xu","family":"Zhang","sequence":"additional","affiliation":[]}],"member":"78","reference":[{"key":"10.1016\/j.future.2025.107983_b1","series-title":"Proceedings of the 6th APNet","first-page":"101","article-title":"Approximate gradient synchronization with AQGB","author":"Liu","year":"2022"},{"key":"10.1016\/j.future.2025.107983_b2","series-title":"Proceedings of the 11th OSDI","first-page":"583","article-title":"Scaling distributed machine learning with the parameter server","author":"Li","year":"2014"},{"key":"10.1016\/j.future.2025.107983_b3","unstructured":"L. Zheng, Z. Li, H. Zhang, Y. Zhuang, Z. Chen, Y. Huang, Y. Wang, Y. Xu, D. Zhuo, E.P. Xing, J.E. Gonzalez, I. Stoica, Alpa: Automating Inter- and Intra-Operator Parallelism for Distributed Deep Learning, in: Proceedings of the 16th OSDI, ISBN: 978-1-939133-28-1, 2022, pp. 559\u2013578."},{"key":"10.1016\/j.future.2025.107983_b4","series-title":"Proceedings of the 27th SOSP","first-page":"16","article-title":"A generic communication scheduler for distributed DNN training acceleration","author":"Peng","year":"2019"},{"key":"10.1016\/j.future.2025.107983_b5","doi-asserted-by":"crossref","unstructured":"H. Xu, C.-Y. Ho, A.M. Abdelmoniem, A. Dutta, E.H. Bergou, K. Karatsenidis, M. Canini, P. Kalnis, GRACE: A Compressed Communication Framework for Distributed Machine Learning, in: Proceedings of the 41st ICDCS, 2021, pp. 561\u2013572.","DOI":"10.1109\/ICDCS51616.2021.00060"},{"issue":"3","key":"10.1016\/j.future.2025.107983_b6","doi-asserted-by":"crossref","first-page":"230","DOI":"10.1109\/MNET.011.2000530","article-title":"A quantitative survey of communication optimizations in distributed deep learning","volume":"35","author":"Shi","year":"2021","journal-title":"IEEE Netw."},{"key":"10.1016\/j.future.2025.107983_b7","unstructured":"S.H. Hashemi, S. Abdu Jyothi, R. Campbell, TicTac: Accelerating Distributed Deep Learning with Communication Scheduling, in: Proceedings of the 2nd SysML Conference, Vol. 1, 2019, pp. 418\u2013430."},{"key":"10.1016\/j.future.2025.107983_b8","series-title":"Proceedings of the 21st NSDI","first-page":"1191","article-title":"THC: Accelerating distributed deep learning using tensor homomorphic compression","author":"Li","year":"2024"},{"issue":"5","key":"10.1016\/j.future.2025.107983_b9","doi-asserted-by":"crossref","first-page":"4488","DOI":"10.1109\/TNET.2024.3423380","article-title":"Releasing the power of in-network aggregation with aggregator-aware routing optimization","volume":"32","author":"Luo","year":"2024","journal-title":"IEEE\/ACM Trans. Netw."},{"key":"10.1016\/j.future.2025.107983_b10","series-title":"Proceedings of the 40th ICML","article-title":"Cocktailsgd: Fine-tuning foundation models over 500Mbps networks","author":"WANG","year":"2023"},{"issue":"9","key":"10.1016\/j.future.2025.107983_b11","doi-asserted-by":"crossref","first-page":"2144","DOI":"10.1109\/TPDS.2021.3062721","article-title":"Overlapping communication with computation in parameter server for scalable DL training","volume":"32","author":"Wang","year":"2021","journal-title":"IEEE Trans. Parallel Distrib. Syst."},{"key":"10.1016\/j.future.2025.107983_b12","series-title":"Proceedings of the 35th ICML","first-page":"5325","article-title":"Error compensated quantized SGD and its applications to large-scale distributed optimization","volume":"Vol. 80","author":"Wu","year":"2018"},{"key":"10.1016\/j.future.2025.107983_b13","series-title":"Proceedings of the 36th ICML","first-page":"7202","article-title":"Distributed learning over unreliable networks","volume":"Vol. 97","author":"Yu","year":"2019"},{"key":"10.1016\/j.future.2025.107983_b14","doi-asserted-by":"crossref","DOI":"10.1016\/j.jnca.2020.102590","article-title":"JPAS: Job-progress-aware flow scheduling for deep learning clusters","volume":"158","author":"Zhou","year":"2020","journal-title":"J. Netw. Comput. Appl."},{"key":"10.1016\/j.future.2025.107983_b15","series-title":"Proceedings of the 3rd APNet","first-page":"22","article-title":"Rethinking transport layer design for distributed machine learning","author":"Xia","year":"2019"},{"key":"10.1016\/j.future.2025.107983_b16","unstructured":"A. Koloskova*, T. Lin*, S.U. Stich, M. Jaggi, Decentralized Deep Learning with Arbitrary Communication Compression, in: Proceedings of ICLR, 2020."},{"year":"2020","series-title":"On communication compression for distributed optimization on heterogeneous data","author":"Stich","key":"10.1016\/j.future.2025.107983_b17"},{"key":"10.1016\/j.future.2025.107983_b18","series-title":"Proceedings of the 31st Annual Conference on Neural Information Processing Systems","first-page":"1710","article-title":"QSGD: Communication-efficient SGD via randomized quantization and encoding","author":"Alistarh","year":"2017"},{"key":"10.1016\/j.future.2025.107983_b19","doi-asserted-by":"crossref","unstructured":"A.M. Abdelmoniem, M. Canini, DC2: Delay-aware Compression Control for Distributed Machine Learning, in: Proceedings of INFOCOM, 2021, pp. 1\u201310.","DOI":"10.1109\/INFOCOM42981.2021.9488810"},{"issue":"2","key":"10.1016\/j.future.2025.107983_b20","doi-asserted-by":"crossref","DOI":"10.1145\/3377454","article-title":"A survey on distributed machine learning","volume":"53","author":"Verbraeken","year":"2020","journal-title":"ACM Comput. Surv."},{"issue":"3","key":"10.1016\/j.future.2025.107983_b21","doi-asserted-by":"crossref","first-page":"1465","DOI":"10.1109\/TAC.2016.2585302","article-title":"Distributed reinforcement learning via gossip","volume":"62","author":"Mathkar","year":"2017","journal-title":"IEEE Trans. Autom. Control"},{"issue":"1","key":"10.1016\/j.future.2025.107983_b22","doi-asserted-by":"crossref","first-page":"156","DOI":"10.1109\/TSC.2024.3506480","article-title":"Efficient parameter synchronization for peer-to-peer distributed learning with selective multicast","volume":"18","author":"Luo","year":"2025","journal-title":"IEEE Trans. Serv. Comput."},{"issue":"5","key":"10.1016\/j.future.2025.107983_b23","doi-asserted-by":"crossref","first-page":"4793","DOI":"10.1109\/TNSE.2024.3419030","article-title":"Efficient inter-datacenter AllReduce with multiple trees","volume":"11","author":"Luo","year":"2024","journal-title":"IEEE Trans. Netw. Sci. Eng."},{"year":"2023","series-title":"Global-QSGD: Practical floatless quantization for distributed learning with theoretical guarantees","author":"Xin","key":"10.1016\/j.future.2025.107983_b24"},{"year":"2023","series-title":"Challenges of quantization in machine learning (ML)","key":"10.1016\/j.future.2025.107983_b25"},{"issue":"4","key":"10.1016\/j.future.2025.107983_b26","doi-asserted-by":"crossref","DOI":"10.1145\/3510587","article-title":"Communication-efficient federated learning with adaptive quantization","volume":"13","author":"Mao","year":"2022","journal-title":"ACM Trans. Intell. Syst. Technol."},{"issue":"9","key":"10.1016\/j.future.2025.107983_b27","doi-asserted-by":"crossref","first-page":"2678","DOI":"10.1109\/JSAC.2022.3192050","article-title":"AC-SGD: Adaptively compressed SGD for communication-efficient distributed learning","volume":"40","author":"Yan","year":"2022","journal-title":"IEEE J. Sel. Areas Commun."},{"key":"10.1016\/j.future.2025.107983_b28","series-title":"Proceedings of the 17th EuroSys","first-page":"435","article-title":"Out-of-order backprop: An effective scheduling technique for deep learning","author":"Oh","year":"2022"},{"key":"10.1016\/j.future.2025.107983_b29","doi-asserted-by":"crossref","unstructured":"S. Wang, D. Li, J. Geng, Geryon: Accelerating Distributed CNN Training by Network-Level Flow Scheduling, in: Proceedings of INFOCOM, 2020, pp. 1678\u20131687.","DOI":"10.1109\/INFOCOM41043.2020.9155282"},{"key":"10.1016\/j.future.2025.107983_b30","series-title":"Proceedings of the ACM Web Conference","first-page":"1764","article-title":"Modeling and optimizing the scaling performance in distributed deep learning training","author":"Liu","year":"2022"},{"key":"10.1016\/j.future.2025.107983_b31","series-title":"Proceedings of the 50th ICPP","article-title":"Prophet: Speeding up distributed DNN training with predictable communication scheduling","author":"Zhang","year":"2021"},{"key":"10.1016\/j.future.2025.107983_b32","series-title":"Proceedings of the 25th PPoPP","first-page":"45","article-title":"Taming unbalanced training workloads in deep learning with partial collective operations","author":"Li","year":"2020"},{"key":"10.1016\/j.future.2025.107983_b33","unstructured":"A. Eisenman, K.K. Matam, S. Ingram, D. Mudigere, R. Krishnamoorthi, K. Nair, M. Smelyanskiy, M. Annavaram, Check-N-Run: a Checkpointing System for Training Deep Learning Recommendation Models, in: Proceedings of the 19th NSDI, Renton, WA, ISBN: 978-1-939133-27-4, 2022, pp. 929\u2013943."},{"key":"10.1016\/j.future.2025.107983_b34","series-title":"Proceedings of the 14th OSDI","first-page":"937","article-title":"Kungfu: Making training in distributed machine learning adaptive","author":"Mai","year":"2020"},{"issue":"11","key":"10.1016\/j.future.2025.107983_b35","doi-asserted-by":"crossref","first-page":"2224","DOI":"10.1109\/TPDS.2024.3460185","article-title":"Efficient cross-cloud partial reduce with CREW","volume":"35","author":"Luo","year":"2024","journal-title":"IEEE Trans. Parallel Distrib. Syst."},{"key":"10.1016\/j.future.2025.107983_b36","series-title":"IEEE Std 754-2019 (Revision of IEEE 754-2008)","first-page":"1","article-title":"IEEE standard for floating-point arithmetic","year":"2019"},{"key":"10.1016\/j.future.2025.107983_b37","unstructured":"X. Sun, N. Wang, C.-y. Chen, J.-m. Ni, A. Agrawal, X. Cui, S. Venkataramani, K. El Maghraoui, V. Srinivasan, K. Gopalakrishnan, Ultra-Low Precision 4-Bit Training of Deep Neural Networks, in: Proceedings of the 34th NIPS, ISBN: 9781713829546, 2020."},{"issue":"1","key":"10.1016\/j.future.2025.107983_b38","doi-asserted-by":"crossref","first-page":"178","DOI":"10.1109\/TNET.2022.3187821","article-title":"Meeting coflow deadlines in data center networks with policy-based selective completion","volume":"31","author":"Luo","year":"2023","journal-title":"IEEE\/ACM Trans. Netw."},{"key":"10.1016\/j.future.2025.107983_b39","doi-asserted-by":"crossref","unstructured":"M. Khani, M. Ghobadi, M. Alizadeh, Z. Zhu, M. Glick, K. Bergman, A. Vahdat, B. Klenk, E. Ebrahimi, SiP-ML: high-bandwidth optical network interconnects for machine learning training, in: Proceedings of the ACM SIGCOMM Conference, 2021, pp. 657\u2013675.","DOI":"10.1145\/3452296.3472900"},{"year":"2019","series-title":"SCALE-sim: Systolic CNN accelerator simulator","author":"Samajdar","key":"10.1016\/j.future.2025.107983_b40"},{"key":"10.1016\/j.future.2025.107983_b41","doi-asserted-by":"crossref","unstructured":"K. He, X. Zhang, S. Ren, J. Sun, Deep Residual Learning for Image Recognition, in: Proceedings of CVPR, 2016, pp. 770\u2013778.","DOI":"10.1109\/CVPR.2016.90"},{"issue":"12","key":"10.1016\/j.future.2025.107983_b42","doi-asserted-by":"crossref","first-page":"4326","DOI":"10.1109\/TPDS.2022.3187815","article-title":"Acctfm: An effective intra-layer model parallelization strategy for training large-scale transformer-based models","volume":"33","author":"Zeng","year":"2022","journal-title":"IEEE Trans. Parallel Distrib. Syst."},{"issue":"04","key":"10.1016\/j.future.2025.107983_b43","doi-asserted-by":"crossref","first-page":"3817","DOI":"10.1609\/aaai.v34i04.5793","article-title":"On the discrepancy between the theoretical analysis and practical implementations of compressed communication for distributed deep learning","volume":"34","author":"Dutta","year":"2020","journal-title":"Proc. AAAI"},{"issue":"1","key":"10.1016\/j.future.2025.107983_b44","doi-asserted-by":"crossref","first-page":"217","DOI":"10.1109\/JSAIT.2020.2985917","article-title":"Qsparse-local-SGD: Distributed SGD with quantization, sparsification, and local computations","volume":"1","author":"Basu","year":"2020","journal-title":"IEEE J. Sel. Areas Inf. Theory"},{"issue":"1","key":"10.1016\/j.future.2025.107983_b45","doi-asserted-by":"crossref","first-page":"166","DOI":"10.1007\/s11390-023-2894-6","article-title":"xCCL: A survey of industry-led collective communication libraries for deep learning","volume":"38","author":"Weingram","year":"2023","journal-title":"J. Comput. Sci. Tech."},{"issue":"6","key":"10.1016\/j.future.2025.107983_b46","doi-asserted-by":"crossref","first-page":"1161","DOI":"10.1109\/JSAC.2020.2986616","article-title":"Efficient file dissemination in data center networks with priority-based adaptive multicast","volume":"38","author":"Luo","year":"2020","journal-title":"IEEE J. Sel. Areas Commun."},{"key":"10.1016\/j.future.2025.107983_b47","unstructured":"R. H\u00f6nig, Y. Zhao, R. Mullins, DAdaQuant: Doubly-adaptive quantization for communication-efficient Federated Learning, in: Proceedings of the 39th ICML, Vol. 162, 2022, pp. 8852\u20138866."},{"key":"10.1016\/j.future.2025.107983_b48","series-title":"Proceedings of the 18th MASS","first-page":"136","article-title":"DQ-SGD: Dynamic quantization in SGD for communication-efficient distributed learning","author":"Yan","year":"2021"}],"container-title":["Future Generation Computer Systems"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S0167739X2500278X?httpAccept=text\/xml","content-type":"text\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/api.elsevier.com\/content\/article\/PII:S0167739X2500278X?httpAccept=text\/plain","content-type":"text\/plain","content-version":"vor","intended-application":"text-mining"}],"deposited":{"date-parts":[[2025,9,27]],"date-time":"2025-09-27T15:41:48Z","timestamp":1758987708000},"score":1,"resource":{"primary":{"URL":"https:\/\/linkinghub.elsevier.com\/retrieve\/pii\/S0167739X2500278X"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,1]]},"references-count":48,"alternative-id":["S0167739X2500278X"],"URL":"https:\/\/doi.org\/10.1016\/j.future.2025.107983","relation":{},"ISSN":["0167-739X"],"issn-type":[{"type":"print","value":"0167-739X"}],"subject":[],"published":{"date-parts":[[2026,1]]},"assertion":[{"value":"Elsevier","name":"publisher","label":"This article is maintained by"},{"value":"Approximate Gradient Synchronization With Adaptive Quantized Gradient Broadcast","name":"articletitle","label":"Article Title"},{"value":"Future Generation Computer Systems","name":"journaltitle","label":"Journal Title"},{"value":"https:\/\/doi.org\/10.1016\/j.future.2025.107983","name":"articlelink","label":"CrossRef DOI link to publisher maintained version"},{"value":"article","name":"content_type","label":"Content Type"},{"value":"\u00a9 2025 Elsevier B.V. All rights are reserved, including those for text and data mining, AI training, and similar technologies.","name":"copyright","label":"Copyright"}],"article-number":"107983"}}