{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,1,21]],"date-time":"2026-01-21T21:52:36Z","timestamp":1769032356663,"version":"3.49.0"},"reference-count":36,"publisher":"IEEE","license":[{"start":{"date-parts":[[2025,11,14]],"date-time":"2025-11-14T00:00:00Z","timestamp":1763078400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-029"},{"start":{"date-parts":[[2025,11,14]],"date-time":"2025-11-14T00:00:00Z","timestamp":1763078400000},"content-version":"stm-asf","delay-in-days":0,"URL":"https:\/\/doi.org\/10.15223\/policy-037"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"published-print":{"date-parts":[[2025,11,14]]},"DOI":"10.1109\/cloudcom67567.2025.11331331","type":"proceedings-article","created":{"date-parts":[[2026,1,20]],"date-time":"2026-01-20T20:37:16Z","timestamp":1768941436000},"page":"1-8","source":"Crossref","is-referenced-by-count":0,"title":["An All-Reduce Compatible Top-$K$ Compressor for Communication-Efficient Distributed Learning"],"prefix":"10.1109","author":[{"given":"Chuyan","family":"Chen","sequence":"first","affiliation":[{"name":"Peking University Beijing,China"}]},{"given":"Chenyang","family":"Ma","sequence":"additional","affiliation":[{"name":"Peking University Beijing,China"}]},{"given":"Zhangxin","family":"Li","sequence":"additional","affiliation":[{"name":"Peking University Beijing,China"}]},{"given":"Yutong","family":"He","sequence":"additional","affiliation":[{"name":"Peking University Beijing,China"}]},{"given":"Yanjie","family":"Dong","sequence":"additional","affiliation":[{"name":"Shenzhen MSU-BIT University,Shenzhen,P. R. China"}]},{"given":"Kun","family":"Yuan","sequence":"additional","affiliation":[{"name":"Peking University Beijing,China"}]}],"member":"263","reference":[{"key":"ref1","doi-asserted-by":"publisher","DOI":"10.1109\/JIOT.2016.2579198"},{"key":"ref2","doi-asserted-by":"publisher","DOI":"10.1109\/MC.2017.9"},{"key":"ref3","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2025.3542324"},{"key":"ref4","article-title":"Federated optimization: Distributed machine learning for on-device intelligence","author":"Kone\u010dny","year":"2016","journal-title":"arXiv preprint"},{"key":"ref5","first-page":"429","article-title":"Federated optimization in heterogeneous networks","volume-title":"Proceedings of Machine learning and systems","volume":"2","author":"Li"},{"key":"ref6","article-title":"Gradient sparsification for communication-efficient distributed optimization","volume":"31","author":"Wangni","year":"2018","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref7","article-title":"Sparsified sgd with memory","volume":"31","author":"Stich","year":"2018","journal-title":"Advances in neural information processing systems"},{"key":"ref8","first-page":"76444","article-title":"Momentum provably im-proves error feedback!","volume":"36","author":"Fatkhullin","year":"2023","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref9","first-page":"571","article-title":"Project adam: Building an efficient and scalable deep learning training system","volume-title":"11th USENIX symposium on operating systems design and implementation (OSDI 14)","author":"Chilimbi"},{"key":"ref10","doi-asserted-by":"publisher","DOI":"10.14778\/1920841.1920931"},{"key":"ref11","article-title":"Communication effi-cient distributed machine learning with the parameter server","volume":"27","author":"Li","year":"2014","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref12","doi-asserted-by":"publisher","DOI":"10.1109\/JPROC.2019.2918951"},{"key":"ref13","first-page":"1273","article-title":"Communication-efficient learning of deep networks from decentralized data","author":"McMahan","year":"2017","journal-title":"Artificial intelligence and statistics"},{"key":"ref14","doi-asserted-by":"publisher","DOI":"10.1561\/2200000083"},{"key":"ref15","first-page":"5132","article-title":"Scaffold: Stochastic controlled averaging for federated learning","volume-title":"International conference on machine learning","author":"Karimireddy"},{"key":"ref16","first-page":"7611","article-title":"Tackling the ob-jective inconsistency problem in heterogeneous federated optimization","volume":"33","author":"Wang","year":"2020","journal-title":"Advances in neural information processing systems"},{"key":"ref17","first-page":"1877","article-title":"Language mod-els are few-shot learners","volume":"33","author":"Brown","year":"2020","journal-title":"Advances in neural information processing systems"},{"issue":"140","key":"ref18","first-page":"1","article-title":"Exploring the limits of transfer learning with a unified text-to-text transformer","volume":"21","author":"Raffel","year":"2020","journal-title":"Journal of machine learning research"},{"key":"ref19","article-title":"Training compute-optimal large language models","author":"Hoffmann","year":"2022","journal-title":"arXiv preprint"},{"key":"ref20","article-title":"Llama: Open and efficient foundation language models","author":"Touvron","year":"2023","journal-title":"arXiv preprint"},{"key":"ref21","article-title":"Powersgd: Practical low-rank gradient compression for distributed optimization","volume":"32","author":"Vogels","year":"2019","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref22","article-title":"Galore: Memory-efficient llm training by gradient low-rank projection","author":"Zhao","year":"2024","journal-title":"arXiv preprint"},{"key":"ref23","article-title":"Greedy low-rank gradient compression for distributed learning with convergence guarantees","author":"Chen","year":"2025","journal-title":"arXiv preprint"},{"key":"ref24","article-title":"Subspace optimization for large language models with convergence guarantees","author":"He","year":"2024","journal-title":"arXiv preprint"},{"key":"ref25","article-title":"Qsgd: Communication-efficient sgd via gradient quantization and encoding","volume":"30","author":"Alistarh","year":"2017","journal-title":"Advances in neural information processing systems"},{"key":"ref26","first-page":"129","article-title":"Natural compression for distributed deep learning","author":"Horv\u00f3th","year":"2022","journal-title":"Mathematical and Scientific Machine Learning"},{"key":"ref27","article-title":"Deep gradient compression: Reducing the communication bandwidth for distributed training","author":"Lin","year":"2017","journal-title":"arXiv preprint"},{"key":"ref28","doi-asserted-by":"publisher","DOI":"10.1109\/ICDCS57875.2023.00031"},{"key":"ref29","doi-asserted-by":"publisher","DOI":"10.21437\/Interspeech.2014-274"},{"key":"ref30","first-page":"3252","article-title":"Error feedback fixes signsgd and other gradient compression schemes","volume-title":"International Conference on Machine Learning","author":"Karimireddy"},{"key":"ref31","first-page":"4384","article-title":"Ef21: A new, simpler, theoretically better, and practically faster error feedback","volume":"34","author":"Richtarik","year":"2021","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref32","first-page":"18955","article-title":"Lower bounds and nearly optimal algorithms in distributed learning with communication com-pression","volume":"35","author":"Huang","year":"2022","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref33","first-page":"47991","article-title":"Unbiased compression saves commu-nication in distributed optimization: When and how much?","volume":"36","author":"He","year":"2023","journal-title":"Advances in Neural Information Processing Systems"},{"key":"ref34","article-title":"Atomo: Communication-efficient learning via atomic spar-sification","volume":"31","author":"Wang","year":"2018","journal-title":"Advances in neural information processing systems"},{"key":"ref35","doi-asserted-by":"publisher","DOI":"10.1109\/CVPR.2016.90"},{"key":"ref36","doi-asserted-by":"publisher","DOI":"10.48550\/ARXIV.1907.11692"}],"event":{"name":"2025 lEEE International Conference on Cloud Computing Technology and Science (CloudCom)","location":"Shenzhen, China","start":{"date-parts":[[2025,11,14]]},"end":{"date-parts":[[2025,11,16]]}},"container-title":["2025 lEEE International Conference on Cloud Computing Technology and Science (CloudCom)"],"original-title":[],"link":[{"URL":"http:\/\/xplorestaging.ieee.org\/ielx8\/11330195\/11331311\/11331331.pdf?arnumber=11331331","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,1,21]],"date-time":"2026-01-21T07:08:13Z","timestamp":1768979293000},"score":1,"resource":{"primary":{"URL":"https:\/\/ieeexplore.ieee.org\/document\/11331331\/"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2025,11,14]]},"references-count":36,"URL":"https:\/\/doi.org\/10.1109\/cloudcom67567.2025.11331331","relation":{},"subject":[],"published":{"date-parts":[[2025,11,14]]}}}