Article(id=1329420712814686290, tenantId=1146029695717560320, journalId=1324759914146529478, issueId=1329420635937288321, articleNumber=null, orderNo=null, doi=10.7544/issn1000-1239.202660474, pmid=null, cstr=null, oa=null, hot=null, price=null, onlineType=0, articleFormat=0, articleType=null, articleTypeStr=null, receivedDate=1781712000000, receivedDateStr=2026-06-18, revisedDate=1789920000000, revisedDateStr=2026-09-21, acceptedDate=null, acceptedDateStr=null, onlineDate=1794888197376, onlineDateStr=2026-11-17, pubDate=1790784000000, pubDateStr=2026-10-01, doiRegisterDate=null, doiRegisterDateStr=null, onlineIssueDate=1794888197376, onlineIssueDateStr=2026-11-17, onlineJustAcceptDate=null, onlineJustAcceptDateStr=null, onlineFirstDate=null, onlineFirstDateStr=null, sourceXml=null, magXml=null, createTime=1794888197376, creator=13701087609, updateTime=1794888197376, updator=13701087609, issue=Issue{id=1329420635937288321, tenantId=1146029695717560320, journalId=1324759914146529478, year='2026', volume='63', issue='10', pageStart='2479', pageEnd='2766', issueExtLink='null', onlineDate='null', pubDate='1790784000000', pubDateStr='2026-10-01', beforeIssueId=null, nextIssueId=null, price=null, status=1, issueComplete=1, articleOrder=1, issueType=-1, specialIssue=null, createTime=1794888179048, creator='13701087609', updateTime=1794888179048, updator='13701087609', preIssue=null, nextIssue=null, articleTotal=null, ext=null, issueFiles=null, downloadFileDto=null}, startPage=2596, endPage=2610, ext={EN=ArticleExt(id=1329420713175396435, articleId=1329420712814686290, tenantId=1146029695717560320, journalId=1324759914146529478, language=EN, title=A Quantum-Native Federated Aggregation Algorithm for Heterogeneous Terminals in Dynamic Heterogeneous Networks, columnId=1329420636595794050, journalTitle=Journal of Computer Research and Development, columnName=Special Issue on Security Protection Technology of Massive Heterogeneous Terminals in Dynamic Heterogeneous Networks, runingTitle=null, highlight=null, articleAbstract=

In dynamic heterogeneous networks, massive heterogeneous terminals face a severe communication bottleneck when performing distributed collaborative learning: the frequent exchange of large-scale model parameters among terminals incurs heavy communication overhead. To address this issue, we propose QStar, a quantum-native federated aggregation algorithm for heterogeneous terminals. Unlike existing quantum federated learning schemes that convert quantum model updates into classical bit strings before transmission, QStar performs the transmission and aggregation of model updates within the quantum domain: clients encode local parameter updates into quantum superposition states so that a single quantum state transmission carries diverse candidate update strategies, and the server performs projective measurements guided by historical optimization trajectories to complete aggregation directly within the quantum domain. This mechanism makes the cost of carrying multiple candidate updates grow logarithmically rather than linearly with the number of candidates, and embeds candidate comparison and selection into the quantum measurement itself. Extensive experiments in typical heterogeneous terminal collaboration scenarios show that, compared with quantum federated learning baselines based on classical parameter transmission, QStar improves convergence stability and attains higher classification accuracy than parameter-quantization-based schemes on tasks with more classes and more complex features, providing a feasible technical pathway for collaborative learning of massive heterogeneous terminals in dynamic heterogeneous networks.

, authors=Zinuo Cai, Haiwei Zhu, Qirui Wang, Hongrui Yin, Ruhui Ma*, authorsList=Zinuo Cai, Haiwei Zhu, Qirui Wang, Hongrui Yin, Ruhui Ma, authorCompany=null, correspAuthors=Ruhui Ma, authorNote=

born in 1999. PhD candidate. His main research interests include quantum machine learning and cloud computing

born in 2005. Undergraduate. His main research interest includes federated learning

born in 2005. Undergraduate. His main research interests include quantum machine learning and federated learning

born in 2006. Undergraduate. His main research interests include quantum machine learning and federated learning

born in 1984. PhD, associate professor. His main research interests include distributed computing, cloud computing, and large language models

, correspAuthorsNote=null, copyrightStatement=Copyright © 2026 Journal of Computer Research and Development. All rights reserved., copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=null, magXml=null, pdfUrl=null, pdf=null, pdfFileSize=null, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=null, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=null, mapNumber=null, fund=null), CN=ArticleExt(id=1329420716975435903, articleId=1329420712814686290, tenantId=1146029695717560320, journalId=1324759914146529478, language=CN, title=面向动态异构网络中异质终端的量子原生联邦聚合算法, columnId=1329420636847452292, journalTitle=计算机研究与发展, columnName=动态异构网络中海量异质终端安全防护技术专题, runingTitle=null, highlight=null, articleAbstract=

在动态异构网络中,海量异质终端在进行分布式协同学习时面临严峻的通信瓶颈:终端之间需要频繁交换大规模模型参数,导致沉重的通信开销。为应对上述问题,提出QStar——一种面向异质终端的量子原生联邦聚合算法。与现有量子联邦学习方案需要将量子模型更新转换为经典比特串再进行传输不同,QStar在量子域内完成模型更新的传输与聚合:客户端将本地参数更新编码为量子叠加态,利用单次量子态传输即可承载多样化的候选更新策略;服务器端依据历史优化轨迹执行投影测量,直接在量子域内完成聚合。该机制使承载多个候选更新的通信代价由随候选数线性增长降为对数增长,并将候选比较与选优嵌入量子测量本身。在典型异质终端协同场景下开展的实验表明,相较于基于经典参数传输的量子联邦学习基线,QStar在收敛稳定性方面具有优势,并在类别数较多、特征较复杂的任务上取得了优于参数级量化方案的分类准确率,为动态异构网络中海量异质终端的协同学习提供了一条可行的技术路径。

, authors=蔡子诺, 朱海威, 汪启瑞, 银泓锐, 马汝辉*, authorsList=蔡子诺, 朱海威, 汪启瑞, 银泓锐, 马汝辉, authorCompany=null, correspAuthors=马汝辉, authorNote=

1999年生。博士研究生。主要研究方向为量子机器学习、云计算

2005年生。本科生。主要研究方向为联邦学习

2005年生。本科生。主要研究方向为量子机器学习、联邦学习

2006年生。本科生。主要研究方向为量子机器学习、联邦学习

1984年生。博士,副研究员。主要研究方向为分布式计算、云计算、大模型

, correspAuthorsNote=
马汝辉()
, copyrightStatement=版权所有 © 2026 计算机研究与发展编辑部, copyrightOwner=null, extLink=null, articleAbsUrl=null, sourceXml=fVeXQwMMICpKJw4ZbMRd7Q==, magXml=rNIEQPVMUFJQa5KTcNPvSw==, pdfUrl=null, pdf=6Xgc85+trI1oSKnPNJRqFg==, pdfFileSize=1600571, pdfExtLink=null, richHtmlUrl=null, mobilePdfUrl=null, reviewReport=null, pdfFirstPage=null, abstractGraph=1+6Lv0+RBqSmj28VL9SYRw==, abstractGraphContent=null, abstractVideo=null, citation=null, cebUrl=null, magXmlContent=OK1d72PExUR8xhpIcQqL1A==, mapNumber=null, fund=null)}, authors=[Author(id=1329420717428420740, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1329420717529084038, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420717428420740, language=EN, stringName=Zinuo Cai, firstName=Zinuo, middleName=null, lastName=Cai, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240, bio={"content":"

born in 1999. PhD candidate. His main research interests include quantum machine learning and cloud computing

"}, bioImg=null, bioContent=

born in 1999. PhD candidate. His main research interests include quantum machine learning and cloud computing

, aboutCorrespAuthor=null), CN=AuthorExt(id=1329420717591998599, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420717428420740, language=CN, stringName=蔡子诺, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=上海交通大学计算机学院 上海 200240, bio={"img":"mMqpK+kR2rKtNiAUViPsog==","content":"

1999年生。博士研究生。主要研究方向为量子机器学习、云计算

"}, bioImg=mMqpK+kR2rKtNiAUViPsog==, bioContent=

1999年生。博士研究生。主要研究方向为量子机器学习、云计算

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1329420717319368832, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, xref=null, ext=[AuthorCompanyExt(id=1329420717348728961, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240), AuthorCompanyExt(id=1329420717361311874, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=上海交通大学计算机学院 上海 200240)])]), Author(id=1329420717650718857, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1329420717722022027, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420717650718857, language=EN, stringName=Haiwei Zhu, firstName=Haiwei, middleName=null, lastName=Zhu, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240, bio={"content":"

born in 2005. Undergraduate. His main research interest includes federated learning

"}, bioImg=null, bioContent=

born in 2005. Undergraduate. His main research interest includes federated learning

, aboutCorrespAuthor=null), CN=AuthorExt(id=1329420717784936588, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420717650718857, language=CN, stringName=朱海威, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=上海交通大学计算机学院 上海 200240, bio={"img":"90z2rbVSo8BkhMs8v3qVjA==","content":"

2005年生。本科生。主要研究方向为联邦学习

"}, bioImg=90z2rbVSo8BkhMs8v3qVjA==, bioContent=

2005年生。本科生。主要研究方向为联邦学习

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1329420717319368832, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, xref=null, ext=[AuthorCompanyExt(id=1329420717348728961, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240), AuthorCompanyExt(id=1329420717361311874, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=上海交通大学计算机学院 上海 200240)])]), Author(id=1329420717843656846, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1329420717902377104, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420717843656846, language=EN, stringName=Qirui Wang, firstName=Qirui, middleName=null, lastName=Wang, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240, bio={"content":"

born in 2005. Undergraduate. His main research interests include quantum machine learning and federated learning

"}, bioImg=null, bioContent=

born in 2005. Undergraduate. His main research interests include quantum machine learning and federated learning

, aboutCorrespAuthor=null), CN=AuthorExt(id=1329420717982068881, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420717843656846, language=CN, stringName=汪启瑞, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=上海交通大学计算机学院 上海 200240, bio={"img":"BaRhKkBgn6iFaXVDef+DvQ==","content":"

2005年生。本科生。主要研究方向为量子机器学习、联邦学习

"}, bioImg=BaRhKkBgn6iFaXVDef+DvQ==, bioContent=

2005年生。本科生。主要研究方向为量子机器学习、联邦学习

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1329420717319368832, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, xref=null, ext=[AuthorCompanyExt(id=1329420717348728961, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240), AuthorCompanyExt(id=1329420717361311874, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=上海交通大学计算机学院 上海 200240)])]), Author(id=1329420718070149267, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1329420718133063829, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420718070149267, language=EN, stringName=Hongrui Yin, firstName=Hongrui, middleName=null, lastName=Yin, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240, bio={"content":"

born in 2006. Undergraduate. His main research interests include quantum machine learning and federated learning

"}, bioImg=null, bioContent=

born in 2006. Undergraduate. His main research interests include quantum machine learning and federated learning

, aboutCorrespAuthor=null), CN=AuthorExt(id=1329420718200172694, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420718070149267, language=CN, stringName=银泓锐, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=上海交通大学计算机学院 上海 200240, bio={"img":"n0Kl023bus8nwNNRor1ZLQ==","content":"

2006年生。本科生。主要研究方向为量子机器学习、联邦学习

"}, bioImg=n0Kl023bus8nwNNRor1ZLQ==, bioContent=

2006年生。本科生。主要研究方向为量子机器学习、联邦学习

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1329420717319368832, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, xref=null, ext=[AuthorCompanyExt(id=1329420717348728961, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240), AuthorCompanyExt(id=1329420717361311874, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=上海交通大学计算机学院 上海 200240)])]), Author(id=1329420718275670168, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, orderNo=4, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=ruhuima@sjtu.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1329420718334390426, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420718275670168, language=EN, stringName=Ruhui Ma, firstName=Ruhui, middleName=null, lastName=Ma, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=*, address=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240, bio={"content":"

born in 1984. PhD, associate professor. His main research interests include distributed computing, cloud computing, and large language models

"}, bioImg=null, bioContent=

born in 1984. PhD, associate professor. His main research interests include distributed computing, cloud computing, and large language models

, aboutCorrespAuthor=null), CN=AuthorExt(id=1329420718418276507, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420718275670168, language=CN, stringName=马汝辉, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=*, address=上海交通大学计算机学院 上海 200240, bio={"img":"vweECuXEgo59RzIlHOEKPA==","content":"

1984年生。博士,副研究员。主要研究方向为分布式计算、云计算、大模型

"}, bioImg=vweECuXEgo59RzIlHOEKPA==, bioContent=

1984年生。博士,副研究员。主要研究方向为分布式计算、云计算、大模型

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1329420717319368832, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, xref=null, ext=[AuthorCompanyExt(id=1329420717348728961, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240), AuthorCompanyExt(id=1329420717361311874, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=上海交通大学计算机学院 上海 200240)])])], keywords=[Keyword(id=1329420718518939805, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, orderNo=1, keyword=dynamic heterogeneous networks), Keyword(id=1329420718586048670, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, orderNo=2, keyword=federated learning), Keyword(id=1329420718644768927, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, orderNo=3, keyword=quantum computing), Keyword(id=1329420718728655008, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, orderNo=4, keyword=quantum machine learning), Keyword(id=1329420718804152481, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, orderNo=5, keyword=distributed machine learning), Keyword(id=1329420718862872738, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, orderNo=1, keyword=动态异构网络), Keyword(id=1329420718925787299, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, orderNo=2, keyword=联邦学习), Keyword(id=1329420718976118948, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, orderNo=3, keyword=量子计算), Keyword(id=1329420719043227813, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, orderNo=4, keyword=量子机器学习), Keyword(id=1329420719106142374, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, orderNo=5, keyword=分布式机器学习)], refs=[Reference(id=1329420721631113417, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=1, rfOrder=0, authorNames=null, journalName=null, refType=null, unstructuredReference=Alsagheer D, Xu Lei, Shi Weidong. Decentralized machine learning governance: Overview, opportunities, and challenges[J]. IEEE Access, 2023, 11: 96718−96732, articleTitle=null, refAbstract=null), Reference(id=1329420721702416586, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=2, rfOrder=1, authorNames=null, journalName=null, refType=null, unstructuredReference=Albrecht J P. How the GDPR will change the world[J]. European Data Protection Law Review, 2016, 2: 287, articleTitle=null, refAbstract=null), Reference(id=1329420721765331147, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=3, rfOrder=2, authorNames=null, journalName=null, refType=null, unstructuredReference=Fallah A, Mokhtari A, Ozdaglar A. Personalized federated learning with theoretical guarantees: A model-agnostic meta-learning approach[J]. Advances in Neural Information Processing Systems, 2020, 33: 3557−3568, articleTitle=null, refAbstract=null), Reference(id=1329420721832440012, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=4, rfOrder=3, authorNames=null, journalName=null, refType=null, unstructuredReference=冯奕铭, 钱珍, 李光辉, 等. 异构边缘环境下自适应分层联邦学习协同优化方法[J]. 计算机研究与发展, 2025, 62(6): 1416−1433, articleTitle=null, refAbstract=null), Reference(id=1329420721903743181, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=4, rfOrder=4, authorNames=null, journalName=null, refType=null, unstructuredReference=Feng Yiming, Qian Zhen, Li Guanghui, et al. Synergistic optimization method for adaptive hierarchical federated learning in heterogeneous edge environments[J]. Journal of Computer Research and Development, 2025, 62(6): 1416−1433 (in Chinese), articleTitle=null, refAbstract=null), Reference(id=1329420721975046350, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=5, rfOrder=5, authorNames=null, journalName=null, refType=null, unstructuredReference=Nguyen D C, Ding M, Pathirana P N, et al. Federated learning for Internet of things: A comprehensive survey[J]. IEEE Communications Surveys & tutorials, 2021, 23(3): 1622−1658, articleTitle=null, refAbstract=null), Reference(id=1329420722037960911, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=6, rfOrder=6, authorNames=null, journalName=null, refType=null, unstructuredReference=Liu Dianqi, Bai Liang, Yu Tianyuan, et al. Towards method of horizontal federated learning: A survey[C]//Proc of 2022 8th Int Conf on Big Data and Information Analytics (BigDIA). Piscataway, NJ: IEEE, 2022: 259−266, articleTitle=null, refAbstract=null), Reference(id=1329420722113458384, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=7, rfOrder=7, authorNames=null, journalName=null, refType=null, unstructuredReference=Li Qinbin, Wen Zeyi, Wu Zhaomin, et al. A survey on federated learning systems: Vision, hype and reality for data privacy and protection[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35(4): 3347−3366, articleTitle=null, refAbstract=null), Reference(id=1329420722184761553, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=8, rfOrder=8, authorNames=null, journalName=null, refType=null, unstructuredReference=Yan Rui, Qu Liangqiong, Wei Qingyue, et al. Label-efficient self-supervised federated learning for tackling data heterogeneity in medical imaging[J]. IEEE Transactions on Medical Imaging, 2023, 42(7): 1932−1943, articleTitle=null, refAbstract=null), Reference(id=1329420722239287506, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=9, rfOrder=9, authorNames=null, journalName=null, refType=null, unstructuredReference=Duan Moming, Liu Duo, Chen Xianzhang, et al. Astraea: Self-balancing federated learning for improving classification accuracy of mobile deep learning applications[C]//Proc of the 37th Int Conf on Computer Design (ICCD). Piscataway, NJ: IEEE, 2019: 246−254, articleTitle=null, refAbstract=null), Reference(id=1329420722302202067, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=10, rfOrder=10, authorNames=null, journalName=null, refType=null, unstructuredReference=Jiang Wenxian, Chen Mengjuan, Tao Jun. Federated learning with blockchain for privacy-preserving data sharing in Internet of vehicles[J]. China Communications, 2023, 20(3): 69−85, articleTitle=null, refAbstract=null), Reference(id=1329420722356728020, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=11, rfOrder=11, authorNames=null, journalName=null, refType=null, unstructuredReference=Dang Trung Kien, Lan Xiang, Weng Jianshu, et al. Federated learning for electronic health records[J]. ACM Transactions on Intelligent Systems and Technology, 2022, 13(5): 1−17, articleTitle=null, refAbstract=null), Reference(id=1329420722449002709, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=12, rfOrder=12, authorNames=null, journalName=null, refType=null, unstructuredReference=李程, 袁勇, 郑志勇, 等. 基于区块链的联邦学习: 模型、方法与应用[J]. 自动化学报, 2024, 50(6): 1059−1085, articleTitle=null, refAbstract=null), Reference(id=1329420722511917270, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=12, rfOrder=13, authorNames=null, journalName=null, refType=null, unstructuredReference=Li Cheng, Yuan Yong, Zheng Zhiyong, et al. Blockchain-enabled federated learning: Models, methods and applications[J]. Acta Automatica Sinica, 2024, 50(6): 1059−1085 (in Chinese), articleTitle=null, refAbstract=null), Reference(id=1329420722566443223, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=13, rfOrder=14, authorNames=null, journalName=null, refType=null, unstructuredReference=Wu J, Dong F, Leung H, et al. Topology-aware federated learning in edge computing: A comprehensive survey[J]. ACM Computing Surveys, 2024, 56(10): 1−41, articleTitle=null, refAbstract=null), Reference(id=1329420722667106520, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=14, rfOrder=15, authorNames=null, journalName=null, refType=null, unstructuredReference=Pfeiffer K, Rapp M, Khalili R, et al. Federated learning for computationally constrained heterogeneous devices: A survey[J]. ACM Computing Surveys, 2023, 55(14s): 1−27, articleTitle=null, refAbstract=null), Reference(id=1329420722734215385, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=15, rfOrder=16, authorNames=null, journalName=null, refType=null, unstructuredReference=Gill S S, Kumar A, Singh H, et al. Quantum computing: A taxonomy, systematic review and future directions[J]. Software: Practice and Experience, 2022, 52(1): 66−114, articleTitle=null, refAbstract=null), Reference(id=1329420722792935642, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=16, rfOrder=17, authorNames=null, journalName=null, refType=null, unstructuredReference=Schuld M, Killoran N. Quantum machine learning in feature Hilbert spaces[J]. Physical Review Letters, 2019, 122(4): 040504, articleTitle=null, refAbstract=null), Reference(id=1329420722855850203, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=17, rfOrder=18, authorNames=null, journalName=null, refType=null, unstructuredReference=Biamonte J, Wittek P, Pancotti N, et al. Quantum machine learning[J]. Nature, 2017, 549(7671): 195−202, articleTitle=null, refAbstract=null), Reference(id=1329420722931347676, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=18, rfOrder=19, authorNames=null, journalName=null, refType=null, unstructuredReference=Uddin M R, Shaon S, Rahman R, et al. Quantum federated learning: A comprehensive survey[J]. IEEE Communications Surveys & Tutorials, 2025, 28: 3942−3975, articleTitle=null, refAbstract=null), Reference(id=1329420722998456541, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=19, rfOrder=20, authorNames=null, journalName=null, refType=null, unstructuredReference=Banchi L, Pereira J, Pirandola S. Generalization in quantum machine learning: A quantum information standpoint[J]. PRX Quantum, 2021, 2(4): 040321, articleTitle=null, refAbstract=null), Reference(id=1329420723082342622, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=20, rfOrder=21, authorNames=null, journalName=null, refType=null, unstructuredReference=Wei Wenqi, Liu Ling, Wu Yanzhao, et al. Gradient-leakage resilient federated learning[C]//Proc of the 41st Int Conf on Distributed Computing Systems (ICDCS). Piscataway, NJ: IEEE, 2021: 797−807, articleTitle=null, refAbstract=null), Reference(id=1329420723141062879, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=21, rfOrder=22, authorNames=null, journalName=null, refType=null, unstructuredReference=Alistarh D, Grubic D, Li J, et al. QSGD: Communication-efficient SGD via gradient quantization and encoding[J]. Advances in Neural Information Processing Systems, 2017, 30: 1709−1720, articleTitle=null, refAbstract=null), Reference(id=1329420723187200224, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=22, rfOrder=23, authorNames=null, journalName=null, refType=null, unstructuredReference=Lin Yujun, Han Song, Mao Huizi, et al. Deep gradient compression: Reducing the communication bandwidth for distributed training[C]//Proc of Int Conf on Learning Representations (ICLR). 2018. https://openreview.net/forum?id=SkhQHMW0W, articleTitle=null, refAbstract=null), Reference(id=1329420723250114785, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=23, rfOrder=24, authorNames=null, journalName=null, refType=null, unstructuredReference=Long Zhong, Chen Yuling, Dou Hui, et al. FedSQ: Sparse-quantized federated learning for communication efficiency[J]. IEEE Transactions on Consumer Electronics, 2024, 70(1): 4050−4061, articleTitle=null, refAbstract=null), Reference(id=1329420723321417954, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=24, rfOrder=25, authorNames=null, journalName=null, refType=null, unstructuredReference=Rothchild D, Panda A, Ullah E, et al. FetchSGD: Communication-efficient federated learning with sketching[C]//Proc of Int Conf on Machine Learning. New York: PMLR, 2020: 8253−8265, articleTitle=null, refAbstract=null), Reference(id=1329420723396915427, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=25, rfOrder=26, authorNames=null, journalName=null, refType=null, unstructuredReference=Guan Yixuan, Liu Xuefeng, Niu Jianwei, et al. FedTC: Enabling communication-efficient federated learning via Transform coding[C]//Proc of IEEE INFOCOM 2024-IEEE Conf on Computer Communications. Piscataway, NJ: IEEE, 2024: 821−830, articleTitle=null, refAbstract=null), Reference(id=1329420723464024292, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=26, rfOrder=27, authorNames=null, journalName=null, refType=null, unstructuredReference=王鑫澳, 陈珂, 寿黎但, 等. 基于联邦学习的BERT模型高效训练框架[J]. 软件学报, 2025, 36(9): 4110−4133, articleTitle=null, refAbstract=null), Reference(id=1329420723518550245, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=26, rfOrder=28, authorNames=null, journalName=null, refType=null, unstructuredReference=Wang Xinao, Chen Ke, Shou Lidan, et al. Efficient framework for BERT model training based on federated learning[J]. Journal of Software, 2025, 36(9): 4110−4133 (in Chinese), articleTitle=null, refAbstract=null), Reference(id=1329420723577270502, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=27, rfOrder=29, authorNames=null, journalName=null, refType=null, unstructuredReference=McMahan B, Moore E, Ramage D, et al. Communication-efficient learning of deep networks from decentralized data[C]//Proc of Artificial Intelligence and Statistics. New York: PMLR, 2017: 1273−1282, articleTitle=null, refAbstract=null), Reference(id=1329420723635990759, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=28, rfOrder=30, authorNames=null, journalName=null, refType=null, unstructuredReference=Li Tian, Sahu A K, Zaheer M, et al. Federated optimization in heterogeneous networks[J]. Proceedings of Machine Learning and Systems, 2020, 2: 429−450, articleTitle=null, refAbstract=null), Reference(id=1329420723690516712, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=29, rfOrder=31, authorNames=null, journalName=null, refType=null, unstructuredReference=Karimireddy S P, Kale S, Mohri M, et al. SCAFFOLD: Stochastic controlled averaging for federated learning[C]//Proc of Int Conf on Machine Learning. New York: PMLR, 2020: 5132−5143, articleTitle=null, refAbstract=null), Reference(id=1329420723757625578, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=30, rfOrder=32, authorNames=null, journalName=null, refType=null, unstructuredReference=Li Qinbin, He Bingsheng, Song D. Model-contrastive federated learning[C]//Proc of the IEEE/CVF Conf on Computer Vision and Pattern Recognition. Piscataway, NJ: IEEE, 2021: 10713−10722, articleTitle=null, refAbstract=null), Reference(id=1329420723820540139, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=31, rfOrder=33, authorNames=null, journalName=null, refType=null, unstructuredReference=杜甜, 陈星延, 寇纲, 等. 面向云边个性化模型解耦的聚类联邦学习方法[J]. 计算机学报, 2025, 48(2): 407−432, articleTitle=null, refAbstract=null), Reference(id=1329420723883454700, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=31, rfOrder=34, authorNames=null, journalName=null, refType=null, unstructuredReference=Du Tian, Chen Xingyan, Kou Gang, et al. Clustered federated learning with cloud-edge personalized model decoupling[J]. Chinese Journal of Computers, 2025, 48(2): 407−432 (in Chinese), articleTitle=null, refAbstract=null), Reference(id=1329420723954757869, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=32, rfOrder=35, authorNames=null, journalName=null, refType=null, unstructuredReference=Daskin A. Federated learning with distributed fixed design quantum chips and quantum channels[J]. arXiv preprint arXiv: 2401.13421, 2024, articleTitle=null, refAbstract=null), Reference(id=1329420724021866734, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=33, rfOrder=36, authorNames=null, journalName=null, refType=null, unstructuredReference=Park S, Jung S, Kim J. Dynamic quantum federated learning for satellite-ground integrated systems using slimmable quantum neural networks[J]. IEEE Access, 2024, 12: 58239−58247, articleTitle=null, refAbstract=null), Reference(id=1329420724088975599, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=34, rfOrder=37, authorNames=null, journalName=null, refType=null, unstructuredReference=Havlíček V, Córcoles A D, Temme K, et al. Supervised learning with quantum-enhanced feature spaces[J]. Nature, 2019, 567(7747): 209−212, articleTitle=null, refAbstract=null), Reference(id=1329420724156084464, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=35, rfOrder=38, authorNames=null, journalName=null, refType=null, unstructuredReference=Sandler M, Howard A, Zhu Menglong, et al. MobileNetV2: Inverted residuals and linear bottlenecks[C]//Proc of 2018 IEEE/CVF Conf on Computer Vision and Pattern Recognition. Piscataway, NJ: IEEE, 2018: 4510−4520, articleTitle=null, refAbstract=null), Reference(id=1329420724227387633, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=36, rfOrder=39, authorNames=null, journalName=null, refType=null, unstructuredReference=He Kaiming, Zhang Xiangyu, Ren Shaoqing, et al. Deep residual learning for image recognition[C]//Proc of the IEEE Conf on Computer Vision and Pattern Recognition. Piscataway, NJ: IEEE, 2016: 770−778, articleTitle=null, refAbstract=null), Reference(id=1329420724294496498, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=37, rfOrder=40, authorNames=null, journalName=null, refType=null, unstructuredReference=Li Xiaoxiao, Jiang Meirui, Zhang Xiaofei, et al. FedBN: Federated learning on Non-IID features via local batch normalization[C]//Proc of Int Conf on Learning Representations (ICLR). 2021. https://openreview.net/forum?id=6YEQUn0QICG, articleTitle=null, refAbstract=null), Reference(id=1329420724378382579, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=38, rfOrder=41, authorNames=null, journalName=null, refType=null, unstructuredReference=Arivazhagan M G, Aggarwal V, Singh A K, et al. Federated learning with personalization layers[J]. arXiv preprint arXiv: 1912.00818, 2019, articleTitle=null, refAbstract=null), Reference(id=1329420724449685748, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=39, rfOrder=42, authorNames=null, journalName=null, refType=null, unstructuredReference=Qu Z, Zhao X, Sun L, et al. DAQFL: Dynamic aggregation quantum federated learning algorithm for intelligent diagnosis in Internet of medical things[J]. IEEE Internet of Things Journal, 2025, 12(19): 39313−39325, articleTitle=null, refAbstract=null), Reference(id=1329420724520988917, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=40, rfOrder=43, authorNames=null, journalName=null, refType=null, unstructuredReference=Gurung D, Pokhrel S R. Performance analysis and design of a weighted personalized quantum federated learning[J]. IEEE Transactions on Artificial Intelligence, 2025, 6(8): 2302−2313, articleTitle=null, refAbstract=null), Reference(id=1329420724575514870, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, doi=null, pmid=null, pmcid=null, year=null, volume=null, issue=null, pageStart=null, pageEnd=null, url=null, language=null, rfNumber=41, rfOrder=44, authorNames=null, journalName=null, refType=null, unstructuredReference=Li Xiang, Huang Kaixuan, Yang Wenhao, et al. On the convergence of FedAvg on Non-IID data[C]//Proc of Int Conf on Learning Representations (ICLR). 2020. https://openreview.net/forum?id=HJxNAnVtDS, articleTitle=null, refAbstract=null)], funds=null, companyList=[AuthorCompany(id=1329420717319368832, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, xref=null, ext=[AuthorCompanyExt(id=1329420717348728961, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240), AuthorCompanyExt(id=1329420717361311874, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=上海交通大学计算机学院 上海 200240)])], figs=[ArticleFig(id=1329420719219388583, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 1, caption=Flow chart of federal learning, figureFileSmall=P8BV/qPa0RBVKD7n6qMEZw==, figureFileBig=1+6Lv0+RBqSmj28VL9SYRw==, tableContent=null), ArticleFig(id=1329420719278108840, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图1, caption=联邦学习流程图, figureFileSmall=P8BV/qPa0RBVKD7n6qMEZw==, figureFileBig=1+6Lv0+RBqSmj28VL9SYRw==, tableContent=null), ArticleFig(id=1329420719466852521, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 2, caption=Flow chart of quantum machine learning, figureFileSmall=yAClC1/cEzSzqaXWy/6RCg==, figureFileBig=9p1XSkppmVFCk7y+Xqc6WQ==, tableContent=null), ArticleFig(id=1329420719554932906, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图2, caption=量子机器学习流程图, figureFileSmall=yAClC1/cEzSzqaXWy/6RCg==, figureFileBig=9p1XSkppmVFCk7y+Xqc6WQ==, tableContent=null), ArticleFig(id=1329420719613653163, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 3, caption=Flow chart of QStar system, figureFileSmall=3tdpWwF7sMqN1nQ6gu899Q==, figureFileBig=JZeIFy2f3XyWHelYfC8q2Q==, tableContent=null), ArticleFig(id=1329420719672373420, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图3, caption=QStar系统流程图, figureFileSmall=3tdpWwF7sMqN1nQ6gu899Q==, figureFileBig=JZeIFy2f3XyWHelYfC8q2Q==, tableContent=null), ArticleFig(id=1329420719735287981, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 4, caption=The results of each strategy under MNIST dataset, mild Non-IID, and four classification tasks, figureFileSmall=QNG3jtx5q7ng+aomV3AOJw==, figureFileBig=vHH/FTasxOL4lLxCDcCXow==, tableContent=null), ArticleFig(id=1329420719802396846, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图4, caption=各策略在MNIST数据集、轻度Non-IID、四分类任务下的结果, figureFileSmall=QNG3jtx5q7ng+aomV3AOJw==, figureFileBig=vHH/FTasxOL4lLxCDcCXow==, tableContent=null), ArticleFig(id=1329420719861117103, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 5, caption=The results of each strategy under FMNIST dataset, mild Non-IID, and four classification tasks, figureFileSmall=LKGzppOTPcexE6tNfXRWlw==, figureFileBig=1oZeCDMYFDLqWNaFxz8omw==, tableContent=null), ArticleFig(id=1329420719945003184, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图5, caption=各策略在FMNIST数据集、轻度Non-IID、四分类任务下的结果, figureFileSmall=LKGzppOTPcexE6tNfXRWlw==, figureFileBig=1oZeCDMYFDLqWNaFxz8omw==, tableContent=null), ArticleFig(id=1329420720012112049, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 6, caption=The results of each strategy under CIFAR-10T dataset, mild Non-IID, and four classification tasks, figureFileSmall=QmdME5kAhfDug/3d9v8zgw==, figureFileBig=zJrKluZIXEJegJFGvIBZLw==, tableContent=null), ArticleFig(id=1329420720070832306, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图6, caption=各策略在CIFAR-10数据集、轻度Non-IID、四分类任务下的结果, figureFileSmall=QmdME5kAhfDug/3d9v8zgw==, figureFileBig=zJrKluZIXEJegJFGvIBZLw==, tableContent=null), ArticleFig(id=1329420720137941171, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 7, caption=The results of each strategy under MNIST dataset, severe Non-IID, and four classification tasks, figureFileSmall=KudJ2ZVknNr6d2LzJIAZ5g==, figureFileBig=Tq4fwhwvZhXBiEz9FaHzkQ==, tableContent=null), ArticleFig(id=1329420720200855732, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图7, caption=各策略在MNIST数据集、重度Non-IID、四分类任务下的结果, figureFileSmall=KudJ2ZVknNr6d2LzJIAZ5g==, figureFileBig=Tq4fwhwvZhXBiEz9FaHzkQ==, tableContent=null), ArticleFig(id=1329420720267964597, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 8, caption=The results of each strategy under FMNIST dataset, severe Non-IID, and four classification tasks, figureFileSmall=F/3stBs2NanmVkTGd52sjg==, figureFileBig=/Fo2WTNoYmmsGlrgowqjEA==, tableContent=null), ArticleFig(id=1329420720335073462, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图8, caption=各策略在FMNIST数据集、重度Non-IID、四分类任务下的结果, figureFileSmall=F/3stBs2NanmVkTGd52sjg==, figureFileBig=/Fo2WTNoYmmsGlrgowqjEA==, tableContent=null), ArticleFig(id=1329420720393793719, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 9, caption=The results of each strategy under CIFAR-10 dataset, severe Non-IID, and four classification tasks, figureFileSmall=2JbJcKQtT94z/Xx7p8XRmA==, figureFileBig=IQeu2KDlQnJ2kQKvWYxbSQ==, tableContent=null), ArticleFig(id=1329420720448319672, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图9, caption=各策略在CIFAR-10数据集、重度Non-IID、四分类任务下的结果, figureFileSmall=2JbJcKQtT94z/Xx7p8XRmA==, figureFileBig=IQeu2KDlQnJ2kQKvWYxbSQ==, tableContent=null), ArticleFig(id=1329420720511234233, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 10, caption=The results of each strategy under MNIST dataset, mild Non-IID, and two classification tasks, figureFileSmall=0lE9cEL3wwIN9jYoH/HLQw==, figureFileBig=EXzUiqho3ysm5RlKWCsYLA==, tableContent=null), ArticleFig(id=1329420720578343098, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图10, caption=各策略在MNIST数据集、轻度Non-IID、二分类任务下的结果, figureFileSmall=0lE9cEL3wwIN9jYoH/HLQw==, figureFileBig=EXzUiqho3ysm5RlKWCsYLA==, tableContent=null), ArticleFig(id=1329420720637063355, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 11, caption=The results of each strategy under FMNIST dataset, mild Non-IID, and two classification tasks, figureFileSmall=fojM0QJTUqSnBWsjB0d+Ag==, figureFileBig=QTtvx5eBeu+wtHMhCKb8/Q==, tableContent=null), ArticleFig(id=1329420720695783612, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图11, caption=各策略在FMNIST数据集、轻度Non-IID、二分类任务下的结果, figureFileSmall=fojM0QJTUqSnBWsjB0d+Ag==, figureFileBig=QTtvx5eBeu+wtHMhCKb8/Q==, tableContent=null), ArticleFig(id=1329420720758698173, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 12, caption=The results of each strategy under CIFAR-10 dataset, mild Non-IID, and two classification tasks, figureFileSmall=t1/lIF++SZheB6eC0jz0jA==, figureFileBig=Qkg4xDdxVtGdtY5VCQp18Q==, tableContent=null), ArticleFig(id=1329420720817418430, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图12, caption=各策略在CIFAR-10数据集、轻度Non-IID、二分类任务下的结果, figureFileSmall=t1/lIF++SZheB6eC0jz0jA==, figureFileBig=Qkg4xDdxVtGdtY5VCQp18Q==, tableContent=null), ArticleFig(id=1329420720901304511, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 13, caption=The results of each strategy under MNIST dataset, severe Non-IID, and two classification tasks, figureFileSmall=hmMqAINkWnss2aMlcusGTg==, figureFileBig=KrDAcUX8YLPsjbBIjtTL8Q==, tableContent=null), ArticleFig(id=1329420720980996288, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图13, caption=各策略在MNIST数据集、重度Non-IID、二分类任务下的结果, figureFileSmall=hmMqAINkWnss2aMlcusGTg==, figureFileBig=KrDAcUX8YLPsjbBIjtTL8Q==, tableContent=null), ArticleFig(id=1329420721043910849, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 14, caption=The results of each strategy under FMNIST dataset, severe Non-IID, and two classification tasks, figureFileSmall=dAa+v1seCEDYOWGJQK2zPw==, figureFileBig=sQbDrtLqbsoBekFqh7G62g==, tableContent=null), ArticleFig(id=1329420721102631106, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图14, caption=各策略在FMNIST数据集、重度Non-IID、二分类任务下的结果, figureFileSmall=dAa+v1seCEDYOWGJQK2zPw==, figureFileBig=sQbDrtLqbsoBekFqh7G62g==, tableContent=null), ArticleFig(id=1329420721157157059, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 15, caption=The results of each strategy under CIFAR-10 dataset, severe Non-IID, and two classification tasks, figureFileSmall=/7sR4HklN8mdJ+w+J5UkSA==, figureFileBig=g4N4sQHGcvX0rZL9KJsH0A==, tableContent=null), ArticleFig(id=1329420721215877316, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图15, caption=各策略在CIFAR-10数据集、重度Non-IID、二分类任务下的结果, figureFileSmall=/7sR4HklN8mdJ+w+J5UkSA==, figureFileBig=g4N4sQHGcvX0rZL9KJsH0A==, tableContent=null), ArticleFig(id=1329420721299763397, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Figure 16, caption=Ablation study of QStar, figureFileSmall=F21J6WCSap3Y0uzaZUwHpw==, figureFileBig=ziViqAGEiYAJx2Y9heHt5w==, tableContent=null), ArticleFig(id=1329420721362677958, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=图16, caption=QStar消融实验, figureFileSmall=F21J6WCSap3Y0uzaZUwHpw==, figureFileBig=ziViqAGEiYAJx2Y9heHt5w==, tableContent=null), ArticleFig(id=1329420721425592519, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=EN, label=Table 1, caption=

Performance of QStar and Other Baseline Methods on MNIST, FMNIST and CIFAR-10 Datasets %

, figureFileSmall=null, figureFileBig=null, tableContent=
数据集FedAvg[27]FedBN[37]FedPer[38]FedProx[28]DAQFL[39]WPQFL[40]QStar
四分类任务
MNIST97.0334.3134.6298.0895.0599.0597.25
FMNIST89.9094.9033.6691.9267.3394.2988.24
CIFAR-1063.8646.5956.6381.9363.1079.2773.56
数据集FedAvg[27]FedBN[37]FedPer[38]FedProx[28]DAQFL[39]WPQFL[40]QStar
二分类任务
MNIST100.0046.4383.64100.0073.68100.00100.00
FMNIST98.0420.4145.10100.00100.0098.08100.00
CIFAR-1081.824.8845.2492.6878.5778.0597.50
), ArticleFig(id=1329420721530450120, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, language=CN, label=表1, caption=

QStar和其他基线方法在MNIST、FMNIST和CIFAR-10数据集上的表现

, figureFileSmall=null, figureFileBig=null, tableContent=
数据集FedAvg[27]FedBN[37]FedPer[38]FedProx[28]DAQFL[39]WPQFL[40]QStar
四分类任务
MNIST97.0334.3134.6298.0895.0599.0597.25
FMNIST89.9094.9033.6691.9267.3394.2988.24
CIFAR-1063.8646.5956.6381.9363.1079.2773.56
数据集FedAvg[27]FedBN[37]FedPer[38]FedProx[28]DAQFL[39]WPQFL[40]QStar
二分类任务
MNIST100.0046.4383.64100.0073.68100.00100.00
FMNIST98.0420.4145.10100.00100.0098.08100.00
CIFAR-1081.824.8845.2492.6878.5778.0597.50
)], attaches=null, journal=Journal(id=1324759606947315908, delFlag=0, nameCn=计算机研究与发展, nameEn=Journal of Computer Research and Development, nameHistory1=null, nameHistory2=null, issn=1000-1239, eissn=null, cn=11-1777/TP, coden=null, periodic=0, language=CN, oaType=null, ccby=null, superviseOffice=null, ownerOffice=null, pubOffice=null, editorOffice=null, officeType=null, aims=null, clcCode=null, officeProv=null, officeCity=null, officeAddr=null, officeZip=null, officeEmail=null, officePhone=null, editDirector=null, officeDirector=null, officeDirectorPhone=null, officeStaffNum=null, officeEmpNum=null, coverPicUrl=k2l8zgqwstuxxevFX2qxVA==, journalPrice=null, startedYear=null, abbrevIsoEn=Journal of Computer Research and Development, journalRemark=null, publicationField=null, createdTime=1793776903141, updatedTime=1793850698612, createdBy=18614031015, updatedBy=13701087609, firstLetterCn=J, firstLetterEn=J, subjectCode=Electronic Information, subjectName=null, subjectCodeEn=Electronic Information, subjectNameEn=null, picCn=k2l8zgqwstuxxevFX2qxVA==, picEn=r6k0Ry+RD+c/VXj5QarmDw==, jcr=null, cjcr=null, exts=[JournalExt(id=1325069127710286688, language=CN, name=计算机研究与发展, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1793850698640, updatedTime=1793850698640, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=https://crad.ict.ac.cn/journalx/authorLogOn.action, submissionEditorUrl=https://crad.ict.ac.cn/journalx/editorLogOn.action, submissionReviewUrl=https://crad.ict.ac.cn/journalx/expertLogOn.action, submissionCeEditorUrl=https://crad.ict.ac.cn/journalx/editorInChiefLogOn.action, submissionAeEditorUrl=, option={"copyright":""}), JournalExt(id=1325069127777395553, language=EN, name=Journal of Computer Research and Development, nameHistory1=null, nameHistory2=null, managedBy=, sponsoredBy=, publishedBy=, editorOffice=, officeProv=null, officeCity=null, officeAddr=, officeZip=, editDirector=, officeDirector=null, officePhone=null, coverPicUrl=null, journalRemark=, submitArticleUrl=null, websiteUrl=, createdTime=1793850698656, updatedTime=1793850698656, createdBy=13701087609, updatedBy=13701087609, submissionGuidelinesUrl=, submissionAuthorUrl=https://crad.ict.ac.cn/journalx/authorLogOn.action, submissionEditorUrl=https://crad.ict.ac.cn/journalx/editorLogOn.action, submissionReviewUrl=https://crad.ict.ac.cn/journalx/expertLogOn.action, submissionCeEditorUrl=https://crad.ict.ac.cn/journalx/editorInChiefLogOn.action, submissionAeEditorUrl=, option={"copyright":""})], databaseList=null, tenantJournalId=1324759914146529478, websiteList=[Website(id=1325070300139892771, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1324759914146529478, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN, language=CN, createTime=1793850978169, createBy=18614031015, updateTime=1793851040259, updateBy=18614031015, name=计算机研究与发展-中文, tplId=1146099689490845704, title=计算机研究与发展, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1325070893751349360, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=articleTextType, value=kx, createTime=1793851119697, updateTime=1793851119697, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893730377837, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=banner, value=null, createTime=1793851119692, updateTime=1793851119692, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893772320883, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=grayFlag, value=0, createTime=1793851119702, updateTime=1793851119702, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893721989228, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=logo, value=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/file/pic?fileId=U27tDsy717YF7XSPA+Vwdw==, createTime=1793851119690, updateTime=1793851119690, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893784903797, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=minRunFlag, value=0, createTime=1793851119705, updateTime=1793851119705, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893742960751, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/file/pic, createTime=1793851119695, updateTime=1793851119695, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893780709492, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=silenceFlag, value=0, createTime=1793851119704, updateTime=1793851119704, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893738766446, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_cn_619/, createTime=1793851119694, updateTime=1793851119694, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893755543665, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=themeColor, value=null, createTime=1793851119698, updateTime=1793851119698, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070893763932274, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300139892771, code=themeStyle, value=null, createTime=1793851119700, updateTime=1793851119700, creator=18614031015, updator=18614031015)]), Website(id=1325070300580294698, webName=null, webTitle=null, webDomain=null, webCopyrigh=null, webIpcNo=null, seoTitle=null, seoKeywords=null, seoDescription=null, tenantJournalId=null, journalId=1324759914146529478, journalNameCn=null, journalNameEn=null, grayFlag=null, tenantId=1146029695717560320, platformId=null, journalGroupId=null, journalGroupNameCn=null, journalGroupNameEn=null, type=1, domain=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN, language=EN, createTime=1793850978274, createBy=18614031015, updateTime=1793851036638, updateBy=18614031015, name=计算机研究与发展-英文, tplId=1146101810881728533, title=Journal of Computer Research and Development, delFlag=0, indexPage=/home, props=[WebsiteProps(id=1325070925288321148, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=articleTextType, value=kx, createTime=1793851127216, updateTime=1793851127216, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925263155321, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=banner, value=null, createTime=1793851127210, updateTime=1793851127210, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925309292671, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=grayFlag, value=0, createTime=1793851127221, updateTime=1793851127221, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925254766712, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=logo, value=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/file/pic?fileId=U27tDsy717YF7XSPA+Vwdw==, createTime=1793851127208, updateTime=1793851127208, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925326069889, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=minRunFlag, value=0, createTime=1793851127225, updateTime=1793851127225, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925279932539, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=picServerUrl, value=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/file/pic, createTime=1793851127214, updateTime=1793851127214, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925317681280, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=silenceFlag, value=0, createTime=1793851127223, updateTime=1793851127223, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925271543930, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=staticResourcePath, value=https://castjournals.cast.org.cn/joweb/cast_kjdb_en_623/, createTime=1793851127212, updateTime=1793851127212, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925292515453, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=themeColor, value=null, createTime=1793851127217, updateTime=1793851127217, creator=18614031015, updator=18614031015), WebsiteProps(id=1325070925300904062, tenantId=1146029695717560320, journalId=null, journalGroupId=null, siteId=1325070300580294698, code=themeStyle, value=null, createTime=1793851127219, updateTime=1793851127219, creator=18614031015, updator=18614031015)])], journalTitle=计算机研究与发展, weixinUrl=null, journalUrl=https://crad.ict.ac.cn/, iacademicId=null, status=1, seqNo=null, journalTitleEn=Journal of Computer Research and Development, journalPhotoCn=k2l8zgqwstuxxevFX2qxVA==, journalPhotoEn=r6k0Ry+RD+c/VXj5QarmDw==, journalFirstLetter=J, journalRecommend=null, journalNew=null, journalCollection=null, jcrJf=null, cjcrJf=null, jcrJfStr=null, cjcrJfStr=null, submissionFirstDecision=null, sciSubjectClassification=null, casSubjectClassification=null, citeScore=null, totalCitationFrequency=null, icpCode=null, psCode=null, advertisingLicenseCode=null, copyrightInformation=null, country=null, option=, provinceCode=null, provinceName=null, collectFlag=false, interPubPlatform=, interPubPlatformUrl=null), detailUrlCn=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/10.7544/issn1000-1239.202660474, detailUrlEn=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/10.7544/issn1000-1239.202660474, pdfUrlCn=https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/PDF/10.7544/issn1000-1239.202660474, pdfUrlEn=https://castjournals.cast.org.cn/joweb/jsjyjyfz/EN/PDF/10.7544/issn1000-1239.202660474, aliStartDate=0, aliEndDate=0, collectionFlag=false, citedCount=null, citedUrl=null, previewStatus=0, delFlag=0, hasFullText=1, orderTime=1790784000000, fullTextJson=null, articleText=null, reference=null)
收藏切换
面向动态异构网络中异质终端的量子原生联邦聚合算法
蔡子诺 , 朱海威 , 汪启瑞 , 银泓锐 , 马汝辉 *
计算机研究与发展 | 动态异构网络中海量异质终端安全防护技术专题 2026,63(10): 2596-2610
收起
收藏切换
计算机研究与发展 |动态异构网络中海量异质终端安全防护技术专题 2026 , 63 (10) : 2596 -2610
面向动态异构网络中异质终端的量子原生联邦聚合算法
全屏
[Author(id=1329420717428420740, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, orderNo=0, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1329420717529084038, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420717428420740, language=EN, stringName=Zinuo Cai, firstName=Zinuo, middleName=null, lastName=Cai, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240, bio={"content":"

born in 1999. PhD candidate. His main research interests include quantum machine learning and cloud computing

"}, bioImg=null, bioContent=

born in 1999. PhD candidate. His main research interests include quantum machine learning and cloud computing

, aboutCorrespAuthor=null), CN=AuthorExt(id=1329420717591998599, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420717428420740, language=CN, stringName=蔡子诺, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=上海交通大学计算机学院 上海 200240, bio={"img":"mMqpK+kR2rKtNiAUViPsog==","content":"

1999年生。博士研究生。主要研究方向为量子机器学习、云计算

"}, bioImg=mMqpK+kR2rKtNiAUViPsog==, bioContent=

1999年生。博士研究生。主要研究方向为量子机器学习、云计算

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1329420717319368832, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, xref=null, ext=[AuthorCompanyExt(id=1329420717348728961, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240), AuthorCompanyExt(id=1329420717361311874, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=上海交通大学计算机学院 上海 200240)])]), Author(id=1329420717650718857, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, orderNo=1, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1329420717722022027, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420717650718857, language=EN, stringName=Haiwei Zhu, firstName=Haiwei, middleName=null, lastName=Zhu, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240, bio={"content":"

born in 2005. Undergraduate. His main research interest includes federated learning

"}, bioImg=null, bioContent=

born in 2005. Undergraduate. His main research interest includes federated learning

, aboutCorrespAuthor=null), CN=AuthorExt(id=1329420717784936588, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420717650718857, language=CN, stringName=朱海威, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=上海交通大学计算机学院 上海 200240, bio={"img":"90z2rbVSo8BkhMs8v3qVjA==","content":"

2005年生。本科生。主要研究方向为联邦学习

"}, bioImg=90z2rbVSo8BkhMs8v3qVjA==, bioContent=

2005年生。本科生。主要研究方向为联邦学习

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1329420717319368832, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, xref=null, ext=[AuthorCompanyExt(id=1329420717348728961, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240), AuthorCompanyExt(id=1329420717361311874, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=上海交通大学计算机学院 上海 200240)])]), Author(id=1329420717843656846, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, orderNo=2, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1329420717902377104, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420717843656846, language=EN, stringName=Qirui Wang, firstName=Qirui, middleName=null, lastName=Wang, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240, bio={"content":"

born in 2005. Undergraduate. His main research interests include quantum machine learning and federated learning

"}, bioImg=null, bioContent=

born in 2005. Undergraduate. His main research interests include quantum machine learning and federated learning

, aboutCorrespAuthor=null), CN=AuthorExt(id=1329420717982068881, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420717843656846, language=CN, stringName=汪启瑞, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=上海交通大学计算机学院 上海 200240, bio={"img":"BaRhKkBgn6iFaXVDef+DvQ==","content":"

2005年生。本科生。主要研究方向为量子机器学习、联邦学习

"}, bioImg=BaRhKkBgn6iFaXVDef+DvQ==, bioContent=

2005年生。本科生。主要研究方向为量子机器学习、联邦学习

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1329420717319368832, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, xref=null, ext=[AuthorCompanyExt(id=1329420717348728961, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240), AuthorCompanyExt(id=1329420717361311874, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=上海交通大学计算机学院 上海 200240)])]), Author(id=1329420718070149267, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, orderNo=3, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=null, emailSecond=null, emailThird=null, correspondingAuthor=0, authorType=1, ext={EN=AuthorExt(id=1329420718133063829, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420718070149267, language=EN, stringName=Hongrui Yin, firstName=Hongrui, middleName=null, lastName=Yin, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240, bio={"content":"

born in 2006. Undergraduate. His main research interests include quantum machine learning and federated learning

"}, bioImg=null, bioContent=

born in 2006. Undergraduate. His main research interests include quantum machine learning and federated learning

, aboutCorrespAuthor=null), CN=AuthorExt(id=1329420718200172694, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420718070149267, language=CN, stringName=银泓锐, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=null, address=上海交通大学计算机学院 上海 200240, bio={"img":"n0Kl023bus8nwNNRor1ZLQ==","content":"

2006年生。本科生。主要研究方向为量子机器学习、联邦学习

"}, bioImg=n0Kl023bus8nwNNRor1ZLQ==, bioContent=

2006年生。本科生。主要研究方向为量子机器学习、联邦学习

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1329420717319368832, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, xref=null, ext=[AuthorCompanyExt(id=1329420717348728961, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240), AuthorCompanyExt(id=1329420717361311874, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=上海交通大学计算机学院 上海 200240)])]), Author(id=1329420718275670168, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, orderNo=4, firstName=null, middleName=null, lastName=null, nameCn=null, orcid=null, stid=null, country=null, authorPic=null, dead=0, email=ruhuima@sjtu.edu.cn, emailSecond=null, emailThird=null, correspondingAuthor=1, authorType=1, ext={EN=AuthorExt(id=1329420718334390426, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420718275670168, language=EN, stringName=Ruhui Ma, firstName=Ruhui, middleName=null, lastName=Ma, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=*, address=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240, bio={"content":"

born in 1984. PhD, associate professor. His main research interests include distributed computing, cloud computing, and large language models

"}, bioImg=null, bioContent=

born in 1984. PhD, associate professor. His main research interests include distributed computing, cloud computing, and large language models

, aboutCorrespAuthor=null), CN=AuthorExt(id=1329420718418276507, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, authorId=1329420718275670168, language=CN, stringName=马汝辉, firstName=null, middleName=null, lastName=null, prefix=null, suffix=null, authorComment=null, nameInitials=null, affiliation=null, department=null, xref=*, address=上海交通大学计算机学院 上海 200240, bio={"img":"vweECuXEgo59RzIlHOEKPA==","content":"

1984年生。博士,副研究员。主要研究方向为分布式计算、云计算、大模型

"}, bioImg=vweECuXEgo59RzIlHOEKPA==, bioContent=

1984年生。博士,副研究员。主要研究方向为分布式计算、云计算、大模型

, aboutCorrespAuthor=null)}, companyList=[AuthorCompany(id=1329420717319368832, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, xref=null, ext=[AuthorCompanyExt(id=1329420717348728961, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=EN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240), AuthorCompanyExt(id=1329420717361311874, tenantId=1146029695717560320, journalId=1324759914146529478, articleId=1329420712814686290, companyId=1329420717319368832, language=CN, country=null, province=null, city=null, postcode=null, companyName=null, departmentName=null, remark=上海交通大学计算机学院 上海 200240)])])]
蔡子诺, 朱海威, 汪启瑞, 银泓锐, 马汝辉*
作者信息
  • 上海交通大学计算机学院 上海 200240
通讯作者:
马汝辉()
作者简介:

1999年生。博士研究生。主要研究方向为量子机器学习、云计算

2005年生。本科生。主要研究方向为联邦学习

2005年生。本科生。主要研究方向为量子机器学习、联邦学习

2006年生。本科生。主要研究方向为量子机器学习、联邦学习

1984年生。博士,副研究员。主要研究方向为分布式计算、云计算、大模型

A Quantum-Native Federated Aggregation Algorithm for Heterogeneous Terminals in Dynamic Heterogeneous Networks
Zinuo Cai, Haiwei Zhu, Qirui Wang, Hongrui Yin, Ruhui Ma*
Affiliations
  • School of Computer Science, Shanghai Jiao Tong University, Shanghai 200240
  • born in 1999. PhD candidate. His main research interests include quantum machine learning and cloud computing

    born in 2005. Undergraduate. His main research interest includes federated learning

    born in 2005. Undergraduate. His main research interests include quantum machine learning and federated learning

    born in 2006. Undergraduate. His main research interests include quantum machine learning and federated learning

    born in 1984. PhD, associate professor. His main research interests include distributed computing, cloud computing, and large language models

出版时间: 2026-10-01 doi: 10.7544/issn1000-1239.202660474
文章导航
收藏切换

在动态异构网络中,海量异质终端在进行分布式协同学习时面临严峻的通信瓶颈:终端之间需要频繁交换大规模模型参数,导致沉重的通信开销。为应对上述问题,提出QStar——一种面向异质终端的量子原生联邦聚合算法。与现有量子联邦学习方案需要将量子模型更新转换为经典比特串再进行传输不同,QStar在量子域内完成模型更新的传输与聚合:客户端将本地参数更新编码为量子叠加态,利用单次量子态传输即可承载多样化的候选更新策略;服务器端依据历史优化轨迹执行投影测量,直接在量子域内完成聚合。该机制使承载多个候选更新的通信代价由随候选数线性增长降为对数增长,并将候选比较与选优嵌入量子测量本身。在典型异质终端协同场景下开展的实验表明,相较于基于经典参数传输的量子联邦学习基线,QStar在收敛稳定性方面具有优势,并在类别数较多、特征较复杂的任务上取得了优于参数级量化方案的分类准确率,为动态异构网络中海量异质终端的协同学习提供了一条可行的技术路径。

动态异构网络  /  联邦学习  /  量子计算  /  量子机器学习  /  分布式机器学习

In dynamic heterogeneous networks, massive heterogeneous terminals face a severe communication bottleneck when performing distributed collaborative learning: the frequent exchange of large-scale model parameters among terminals incurs heavy communication overhead. To address this issue, we propose QStar, a quantum-native federated aggregation algorithm for heterogeneous terminals. Unlike existing quantum federated learning schemes that convert quantum model updates into classical bit strings before transmission, QStar performs the transmission and aggregation of model updates within the quantum domain: clients encode local parameter updates into quantum superposition states so that a single quantum state transmission carries diverse candidate update strategies, and the server performs projective measurements guided by historical optimization trajectories to complete aggregation directly within the quantum domain. This mechanism makes the cost of carrying multiple candidate updates grow logarithmically rather than linearly with the number of candidates, and embeds candidate comparison and selection into the quantum measurement itself. Extensive experiments in typical heterogeneous terminal collaboration scenarios show that, compared with quantum federated learning baselines based on classical parameter transmission, QStar improves convergence stability and attains higher classification accuracy than parameter-quantization-based schemes on tasks with more classes and more complex features, providing a feasible technical pathway for collaborative learning of massive heterogeneous terminals in dynamic heterogeneous networks.

dynamic heterogeneous networks  /  federated learning  /  quantum computing  /  quantum machine learning  /  distributed machine learning
蔡子诺, 朱海威, 汪启瑞, 银泓锐, 马汝辉. 面向动态异构网络中异质终端的量子原生联邦聚合算法. 计算机研究与发展, 2026 , 63 (10) : 2596 -2610 . DOI: 10.7544/issn1000-1239.202660474
Zinuo Cai, Haiwei Zhu, Qirui Wang, Hongrui Yin, Ruhui Ma. A Quantum-Native Federated Aggregation Algorithm for Heterogeneous Terminals in Dynamic Heterogeneous Networks[J]. Journal of Computer Research and Development, 2026 , 63 (10) : 2596 -2610 . DOI: 10.7544/issn1000-1239.202660474
在动态异构网络中,海量异质终端持续产生着规模庞大、分布广泛且高度敏感的本地数据。这些终端在硬件算力、网络带宽和数据分布等维度上存在显著差异。传统的集中式数据汇集与模型训练范式面临根本性困难:一方面,将分散于各终端的敏感数据集中到云端服务器会引发数据合规风险与通信成本[1-2];另一方面,异质终端之间算力与带宽的严重不对等使得统一训练策略难以适配所有参与方[3-4]。因此,如何在尊重终端异质性的前提下高效地协同利用多方数据价值,成为动态异构网络环境下的核心研究挑战。
联邦学习[5]是一种分布式机器学习范式,它使得多个参与方能够在不暴露各自本地原始数据的前提下,通过仅交换模型参数或梯度更新,共同训练一个全局共享模型。根据参与方数据特征空间与样本空间的重叠关系,联邦学习进一步细分为横向联邦学习、纵向联邦学习和联邦迁移学习等形态[6-7],分别在用户特征重叠较多、样本重叠较多或两者均稀疏的场景下发挥作用。正是由于这种“数据不动模型动”的独特优势,联邦学习被迅速应用于金融风控、医疗影像分析和智能交通等隐私敏感任务中,成为破解数据孤岛与隐私合规困境的关键技术[8-12]。
然而,联邦学习在带来隐私保护优势的同时,也不可避免地陷入了计算与通信效率的双重瓶颈。在通信层面[13],模型训练的每一轮迭代都需要在中心服务器与海量客户端之间传输完整的模型参数或梯度更新,当模型参数量达到百万乃至数十亿级别时,反复上传与下发所产生的通信开销急剧膨胀,成为制约训练速度和实时性的首要短板。在计算层面[14],瓶颈体现在两端:一方面,大量客户端本身即为算力有限的智能手机、物联网终端等设备,本地训练需消耗其本就紧张的电量与计算资源;另一方面,为在不泄露梯度信息的前提下完成安全聚合,服务器端常需引入同态加密、安全多方计算等复杂密码学操作,这些运算会带来数倍乃至数十倍的额外计算负担。通信与计算瓶颈相互交织,终端异构性、网络带宽波动与数据分布不均等因素进一步放大了这一矛盾,如何在维持模型精度和隐私保护强度的同时有效降低端到端成本,成为联邦学习迈向大规模实际部署所必须攻克的核心难题。
量子计算[15]是一种遵循量子力学规律来调控量子信息单元进行计算的全新计算范式。它以量子比特为基本单位,利用量子叠加、纠缠和干涉等特性,使得量子计算机在特定问题上具备远超经典计算机的潜在运算能力。在机器学习领域,量子机器学习[16]旨在将这些量子特性引入模型设计与训练流程,当前最具实用前景的是变分量子电路(variational quantum circuit,VQC)[17],这类混合量子-经典模型以参数化量子门构成的量子层替代经典网络中难以高效计算的部分,通过梯度下降进行端到端训练,在同等参数量下对高维数据的表达与利用效率显著优于经典模型。随着量子硬件与算法的持续进步,量子机器学习为解决联邦学习中的计算瓶颈与模型效率问题提供了极具前景的新技术路径。
量子联邦学习(quantum federated learning,QFL)[18]将量子机器学习与联邦学习深度融合,延续了经典联邦学习的分布式框架,但将本地训练主体替换为量子神经网络或混合量子-经典模型。得益于量子模型参数的高效性——同等表达能力下量子电路参数数量远少于经典深度网络——量子联邦学习可大幅压缩单轮通信所需传输的数据量。
然而,当前的含噪声中等规模量子(NISQ)硬件没有实现完整的量子纠错,量子门保真度和相干时间都有限,尚不足以支持高保真、高速率传输量子态的网络,因此现有量子联邦学习方法仍将量子模型更新转换为经典浮点数后,经由经典信道进行传输与聚合,并未利用量子态的叠加与纠缠特性实现真正的原生态信息交互。这种“量子训练-经典传输”的范式割裂导致了2个关键局限:其一,信息传输效率受限[19],量子模型训练中产生的丰富信息在转换为经典表示时被大量丢弃,且单次传输仅能携带单一的确定性更新;其二,附加处理开销偏高[20],经典信道中的参数传输往往还需叠加额外的编解码与密码学处理,进一步加重了系统整体的计算负担。
针对上述瓶颈,本文提出QStar——一种基于量子原生通信框架的联邦聚合算法。QStar不再将量子模型更新转换为经典参数进行传输,而是在量子域内直接完成模型更新的通信、选择与聚合,构建了从客户端量子态编码、量子信道传输到服务端量子测量的完整闭环。这不仅利用量子信道压缩了通信量,也降低了直接对旋转角做算术平均带来的风险:量子电路参数定义在以$ 2\text{π} $为周期的圆周上,当各客户端的角度分布跨越±180°边界时,算术平均会给出偏离真实中心的结果。考虑到量子硬件噪声、退相干、隐形传态保真度波动的影响,转换为经典浮点数聚合会放大客户端漂移和不稳定性。而该架构使量子联邦学习摆脱了对经典中间转换的依赖,真正发挥量子态叠加性与相干性在信息传输与聚合中的原生优势,获得性能上的提升。
具体来说,在客户端侧,本文设计了一种多候选量子叠加编码方法。客户端将本地数据集划分为多个互不重叠的子集,在每个子集上独立执行梯度下降以生成多样化的候选更新策略,随后采用振幅编码将所有候选更新及其性能评分编码为单个量子叠加态。通过单次量子态传输即可并行承载全部候选策略,避免了传统方法逐一传输所造成的通信开销累积。
在服务器侧,本文设计了一种方向一致性投影测量聚合方法。服务器首先记录过去若干轮的全局模型更新并计算历史方向参考向量,以此捕捉全局优化轨迹的主导趋势。对各客户端上传量子叠加态中的每个候选更新,服务器计算其与历史方向向量之间的余弦相似度,并将方向相似度与客户端提供的本地性能评分综合为测量权重,据此调制叠加态中对应候选的振幅幅度。随后,服务器执行计算基投影测量,量子态将以较高概率坍缩至与全局优化方向最为一致且本地性能优异的候选更新。这一过程将候选比较与选优嵌入量子测量本身,避免了逐一穷举比较的经典计算冗余。
本文的主要贡献如下:
1)提出QStar框架,以量子态通信取代经典参数传输,构建了从客户端编码、量子信道传输到服务端量子测量的端到端量子原生聚合闭环;
2)设计客户端多候选量子叠加编码方法,在单次量子态传输中并行承载多种候选更新策略;
3)设计服务器端方向一致性投影测量聚合方法,将历史全局优化轨迹融入量子测量过程,以方向感知的方式实现智能选优与高效聚合;
4)基于PennyLane与Flower构建实验环境,在典型分布式协同场景下验证了QStar在通信开销、收敛速度和模型精度方面相较经典参数传输基线的优势。
在经典联邦学习领域,通信开销一直是制约系统可扩展性的首要瓶颈。为缓解这一问题,研究者从模型压缩、通信调度和聚合策略3个维度提出了大量优化方法。
模型压缩方向的核心思路是减少单次通信传输的数据量,主要包括梯度量化和梯度稀疏化两大类技术。QSGD[21]将32位浮点梯度随机量化为低位宽表示,在不显著损失模型精度的前提下实现40%~70%的带宽节省;深度梯度压缩(deep gradient compression,DGC)[22]通过Top-k稀疏化仅传输梯度幅值最大的少量参数;FedSQ[23]进一步将稀疏化与量化进行联合设计,在保持收敛精度的同时最大化通信压缩效果。此外,FetchSGD[24]利用Count Sketch等草图算法将高维梯度投影至低维空间进行传输,FedTC[25]则引入变换编码将梯度转换至能量集中的紧凑域再进行稀疏化,这些方法均在不同应用场景下展现出显著的通信压缩效果[26]。
通信调度方向着眼于减少通信轮次和调整通信频率。FedAvg[27]本身即可视为一种基础的通信调度策略,其通过在服务器聚合前让客户端执行多轮本地更新来降低通信频率。后续工作在此基础上引入自适应通信机制,根据模型收敛状态或网络带宽动态调整本地更新轮数及参与客户端数量,例如在训练初期使用较少本地迭代以快速探索以及在收敛阶段使用较多本地迭代以降低通信成本。
高级聚合策略旨在改善非独立同分布(Non-IID)数据分布下的收敛性能,间接减少达到目标精度所需的通信轮次。FedProx[28]在客户端本地训练目标中引入近端项以抑制客户端漂移;SCAFFOLD[29]利用控制变量修正客户端更新方向,从根本上减少Non-IID数据引起的梯度方差;MOON[30]通过对比学习拉近客户端本地表征与全局表征的距离。这些方法虽然主要面向统计异质性挑战,但通过加速收敛和减少通信轮次,也在客观上产生了显著的通信优化效果[31]。
然而,在量子联邦学习场景下,模型主体变为量子神经网络,其参数空间具有与经典模型截然不同的结构特性,使得经典压缩范式的直接迁移面临根本性挑战。而利用量子编码技术的叠加特性可在同一量子态中并行承载多份信息,使单次传输能够同时传递远超经典编码的信息量,为通信效率优化提供了新思路。
量子联邦学习的通信优化是最近2年才兴起的研究方向,现有工作可大致分为参数级优化、电路级优化和量子原生优化3个层次。
参数级优化延续经典联邦学习中梯度压缩的基本思路,将量化、稀疏化等操作直接施加于量子电路参数。例如,部分工作将QSGD范式迁移至量子联邦学习(QFL)场景,对上传的变分量子电路(VQC)旋转角度进行随机量化以降低每轮通信比特数。然而,量子电路参数的微小摄动可能通过多层的酉变换传播累积,导致模型输出发生显著扰动,因此参数级量化的精度损失对QFL模型性能的影响通常比对经典模型更为敏感。
电路级优化从量子电路结构本身入手,通过设计更紧凑的参数化量子电路或引入参数共享机制来减少需要传输的参数总量。变分量子电路的结构搜索已被引入联邦学习场景,以自动发现通信成本与模型表达能力之间的最优电路拓扑。此外,在量子比特资源受限的场景下,量子电路切割技术将大规模量子电路分解为多个可独立训练的小型子电路并分别部署于不同客户端,以分布式计算替代集中式的大规模量子计算。
量子原生优化是量子联邦学习通信优化的最新发展方向,其区别于参数级和电路级优化的核心特征是直接利用量子力学的基本特性,如叠加态、纠缠、干涉来完成信息的编码、传输与聚合,而非对经典优化方法进行简单适配。Daskin[32]的量子态梯度传输方案和Park等人[33]的叠加编码方案均属于这一方向的初步探索,证明了量子原生机制在理论上可以实现远超经典压缩极限的通信效率。然而,现有量子原生方案仍存在2个关键局限:其一,量子态编码未充分利用叠加性承载信息多样性,每个客户端每轮仅向服务器传输单一的确定性模型更新,服务器缺乏多候选比较与择优的空间;其二,服务端聚合缺乏历史方向感知的选择机制,未将全局优化的历史轨迹纳入聚合决策,在Non-IID数据分布下容易使高质量更新被大量低质量更新稀释,拖慢全局收敛。更进一步地,变分量子电路的可训练参数为旋转角,定义在以2π为周期的圆周上;当不同客户端的角度分布跨越±180°边界时,算术平均会给出偏离真实中心的结果。这意味着沿用经典的参数平均,在量子模型的参数空间上并非一个良定义的聚合算子。
相较参数级与电路级优化沿袭经典压缩思路,量子原生优化直接利用叠加、纠缠等量子特性完成信息的编码、传输与聚合,存在本质差异。针对上述第1个局限,QStar通过量子编码策略将多组候选更新叠加编码并上传,提供更多更新信息;对于第2个局限,QStar采用与历史轨迹投影测量一样的聚合方法予以解决。
联邦学习的基本架构如图1所示,由中心服务器和多个参与客户端构成,训练过程遵循“下发—本地训练—上传—聚合”的循环迭代模式。
设共有$ K $个参与客户端,每个客户端拥有本地数据集$ {\mathcal{D}}_{k}=\left\{\left({\boldsymbol{x}}_{i},{y}_{i}\right)\right\}_{i=1}^{{n}_{k}} $,其中$ {n}_{k}=\left| {\mathcal{D}}_{k}\right| $为客户端$ k $的样本数,总样本数为$ n=\displaystyle\sum_{k=1}^{K}{n}_{k} $,$ {\boldsymbol{x}}_{i} $是第$ i $ 个样本的输入特征,$ {y}_{i} $为其对应的标签。其训练目标为求解分布式优化问题$ \underset{\boldsymbol{w}}{\min } \,\mathcal{L}\left(\boldsymbol{w}\right)=\displaystyle\sum_{k=1}^{K}\dfrac{{n}_{k}}{n}{\mathcal{L}}_{k}\left(\boldsymbol{w}\right) $。其中$ \boldsymbol{w} $为全局模型参数,$ {\mathcal{L}}_{k}\left(\boldsymbol{w}\right)=\dfrac{1}{{n}_{k}}\displaystyle\sum_{\left({\boldsymbol{x}}_{i},{y}_{i}\right)\in {\mathcal{D}}_{k}}\ell\left(f\left({\boldsymbol{x}}_{i};\boldsymbol{w}\right),{y}_{i}\right) $为客户端$ k $的本地经验损失,$ \ell\left(\cdot ,\cdot \right) $为样本级损失函数。
最经典的联邦学习算法是联邦平均,其每一轮训练迭代$ t $包括4个步骤:
①全局模型下发。服务器将当前全局模型参数$ \boldsymbol{w}_{\rm g}^{(t)} $下发至本轮选中的客户端。
②客户端本地训练。客户端k以$ \boldsymbol{w}_{\rm g}^{(t)} $为初始模型,在$ {\mathcal{D}}_{k} $上执行E轮随机梯度下降:
$ \boldsymbol{w}_{k}^{\left(t,0\right)}=\boldsymbol{w}_{\rm g}^{(t)}\;,$
$ \boldsymbol{w}_{k}^{\left(t,e+1\right)}=\boldsymbol{w}_{k}^{\left(t,e\right)}-\eta \nabla {\mathcal{L}}_{k}\left(\boldsymbol{w}_{k}^{\left(t,e\right)}\right),\;\; e=0{,}1,… ,E-1\;,$
其中$ \eta $为学习率。E轮迭代后,客户端k得到本地模型$ \boldsymbol{w}_{k}^{\left(t+1\right)}=\boldsymbol{w}_{k}^{\left(t,E\right)} $,并计算更新量$ \mathit{\Delta }\boldsymbol{w}_{k}^{(t)}=\boldsymbol{w}_{k}^{\left(t+1\right)}-\boldsymbol{w}_{\rm g}^{(t)} $。
③模型更新上传。客户端将本地更新$ \mathit{\Delta }\boldsymbol{w}_{k}^{(t)} $(或完整的本地模型$ \boldsymbol{w}_{k}^{\left(t+1\right)} $)上传至中心服务器。
④服务器聚合更新。服务器按样本量对各客户端模型加权平均,得到新一轮全局模型:
$ \boldsymbol{w}_{\rm g}^{\left(t+1\right)}=\displaystyle\sum_{k=1}^{K}\dfrac{{n}_{k}}{n}\boldsymbol{w}_{k}^{\left(t+1\right)}\; 。 $
量子计算是基于量子力学原理进行信息处理的计算范式,其基本信息单元为量子比特。与经典比特只能处于0或1确定状态不同,量子比特可以同时处于$ |0\rangle $和$ |1\rangle $的叠加态,这赋予了量子计算独特的并行计算潜力。
1)量子态与叠加原理
单个量子比特的纯态可以用二维复希尔伯特空间$ {\mathbb{C}}^{2} $中的单位向量表示:
$ \left| \psi \rangle =\alpha \right| 0\rangle +\beta \left| 1\rangle ,\quad \right| \alpha {|}^{2}+|\beta {|}^{2}=1\;,$
其中$ \alpha ,\beta \in \mathbb{C} $为概率幅,$ |0\rangle $和$ |1\rangle $为计算基矢。测量该量子比特时,将以概率$ {\left| \alpha \right|^{2} } $坍缩到状态$ |0\rangle $,以概率$ {\left| \beta \right|^{2} } $坍缩到$ |1\rangle $。
2)量子门与量子电路
量子计算通过酉变换操纵量子态,对应的算子为量子门。常用的单量子比特门包括Pauli门(X、Y、Z)、Hadamard门与参数化旋转门(RX、RY、RZ);常用的双量子比特门包括CNOT门与CZ门,用于在量子比特之间产生纠缠。其中参数化旋转门是变分量子电路中的可训练部分。
3)量子测量
量子测量是从量子态中提取经典信息的唯一途径。对于可观测量O(自伴算子$ O={O}^{\dagger } $),其谱分解为$ O=\displaystyle\sum_{a}{\lambda }_{a}{P}_{a} $,其中$ {\lambda }_{a} $为特征值,$ {P}_{a}=\left| a\rangle \langle a\right| $为到对应特征子空间的投影算子。根据玻恩规则,在量子态 $ |\psi \rangle $上测量O,得到结果$ {\lambda }_{a} $的概率为:
$ P\left(a\right)=\left\langle \psi \left| {P}_{a}\right| \psi \right\rangle =\left| \langle a\right| \psi \rangle {|}^{2}\; 。 $
值得注意的是,量子测量是不可逆的,即测量之后量子态将以概率$ P\left(a\right) $坍缩到对应的本征态$ |a\rangle $,原有的叠加信息随之消失。
量子机器学习是将量子计算与经典机器学习融合的交叉领域,旨在利用量子叠加、纠缠和干涉等量子力学特性增强或加速机器学习模型的表达能力与训练效率。当前最具实用前景的是混合量子-经典计算模型,其核心思想是利用量子电路承担非经典可高效计算的前向传播任务,同时保留经典优化器在参数空间中进行梯度下降搜索的灵活性。如图2所示,量子机器学习的基本流程涵盖4个核心阶段:数据编码、量子模型、量子测量和参数优化。
①数据编码。将经典数据加载到量子计算机中是量子机器学习的第1步,称为量子态编码。常见的编码方式包括振幅编码和角度编码。振幅编码将$ N $维经典数据向量$ \boldsymbol{x}=\left({x}_{1},{x}_{2},… ,{x}_{N}\right) $编码为$ \left\lceil {\rm lb} N\right\rceil $个量子比特的振幅[34];角度编码将数据的每个特征值编码为旋转门的旋转角度。
②量子模型。变分量子电路是一种由可训练参数$ \boldsymbol{\theta }=\left({\theta }_{1},{\theta }_{2},… ,{\theta }_{M}\right) $控制的参数化量子电路,它在量子机器学习中扮演着量子神经网络的角色。VQC通常由旋转层和门纠缠层交替堆叠构成。其中旋转层由$ {R}_{X} ({\theta }_{j})、{R}_{Y} ({\theta }_{j})、{R}_{Z} ({\theta }_{j}) $等参数化单量子比特构成,纠缠层由CNOT、CZ等双量子比特门构成,用于产生量子纠缠。
③量子测量。量子模型处理完毕后,需通过量子测量提取经典输出信息。通常选定一个可观测量,在输出态$ |{\psi }_{\text{out}}\rangle $上重复多次测量并计算期望值:
$ f\left(\boldsymbol{x};\boldsymbol{\theta }\right)=\left\langle {\psi }_{\text{out}}\left(\boldsymbol{x};\boldsymbol{\theta }\right)\left| O\right| {\psi }_{\text{out}}\left(\boldsymbol{x};\boldsymbol{\theta }\right)\right\rangle \; 。 $
④参数优化。VQC的参数优化通过经典优化器在经典计算机上完成。利用参数偏移法则,可有效计算量子电路关于参数$ {\theta }_{j} $的梯度:
$ \dfrac{\partial f\left(\boldsymbol{x};\boldsymbol{\theta }\right)}{\partial {\theta }_{j}}=\dfrac{1}{2}\left[f\left(\boldsymbol{x};\boldsymbol{\theta }+\dfrac{\pi }{2}{\boldsymbol{e}}_{j}\right)-f\left(\boldsymbol{x};\boldsymbol{\theta }-\dfrac{\pi }{2}{\boldsymbol{e}}_{j}\right)\right]\;,$
其中$ {\boldsymbol{e}}_{j} $为第j个方向上的单位向量。式(7)的关键优势在于梯度计算仅需2次额外的电路前向评估,无需对量子电路本身进行反向传播。获得梯度后,经典优化器按照标准方式更新参数:
$ {\boldsymbol{\theta }}^{\left(t+1\right)}={\boldsymbol{\theta }}^{(t)}-\eta {\nabla }_{\boldsymbol{\theta }}\mathcal{L} ({\boldsymbol{\theta }}^{(t)})\; 。 $
QStar是一种基于量子通信的量子联邦学习聚合算法,旨在解决现有QFL系统依赖经典信道逐一传输参数更新的瓶颈问题。与传统方法将量子模型参数提取为经典浮点数后通过经典信道传输的范式不同,QStar直接在量子域内完成模型更新的通信与聚合,充分利用量子态的叠加性与相干性实现多候选更新的并行传输和智能化选择。
QStar采用经典的客户端-服务器联邦架构,并在此基础上引入量子通信层。系统流程图如图3所示。
一轮完整训练流程如下:
①全局模型分发。服务器将当前轮次的全局模型参数$ \boldsymbol{\boldsymbol{\mathit{\boldsymbol{\theta}}}}\mathrm{_{\rm{g}lobal}} $下发至所有参与的客户端,确保各客户端在相同的全局模型基础上开展本地训练。
②本地训练与候选生成。每个客户端$ c $在接收到全局模型后,利用本地数据集$ {D}_{c} $执行本地训练,生成多个候选更新并编码为量子叠加态$ |{\mathit{\Psi }}_{c}\rangle $。
③量子态传输。客户端通过量子信道将编码后的叠加态$ |{\mathit{\Psi }}_{c}\rangle $发送至服务器,实现多候选更新的一次性并行传输。
④方向一致性投影测量聚合。服务器接收来自所有客户端的量子叠加态后,结合历史全局更新方向对各候选进行加权测量,从每个客户端的叠加态中坍缩出最优候选更新$ \widetilde{\delta \boldsymbol \theta }_{{i}^{*}}^{(c)} $。
⑤全局模型更新。服务器对所有客户端选出的最优候选进行聚合,更新全局模型:
$ {\boldsymbol \theta }_{\rm global}\leftarrow {\boldsymbol \theta }_{\rm global}+\dfrac{1}{N}\displaystyle\sum_{c=1}^{N}\widetilde{\delta \boldsymbol \theta }_{{i}^{*}}^{(c)}\;,$
其中N为参与本轮训练的客户端总数。
相较于传统QFL中逐一传输单一经典参数更新的方式,QStar的量子原生通信范式在参数传递,全局参数更新这2方面都带来了显著优化。在参数传递上,QStar利用量子态的叠加性,使得客户端能够在单次量子传输中并行编码$ k $个候选更新,通信开销从$O(k\cdot D) $降低至$O(D) $(D为参数维度),显著减少了通信轮次和带宽消耗。在服务端聚合策略上,QStar通过带有历史全局更新方向测量偏好的投影测量对各候选更新在量子域进行选择,直接选出最合适的候选更新,避免了经典方法中比较候选结果的计算冗余,实现了量子投影测量聚合。不仅如此,通过这种带有历史全局更新方向的测量方法,系统能够在动态训练过程中保持优化轨迹的稳定性,有效抑制Non-IID数据环境下的客户端漂移现象。
多候选量子叠加编码目标是在单次量子传输中并行携带多个候选更新策略,充分挖掘量子叠加态的信息承载能力,从而降低通信开销并为服务端的智能选优提供丰富的候选空间。
1)候选生成
每个客户端$ c $首先将其本地数据集$ {D}_{c} $划分为$ k $个互不重叠的子集$ \{D_{c}^{(1)},D_{c}^{(2)},… ,D_{c}^{(k)}\} $,满足:
$ {D}_{c}=\bigcup \limits_{i=1}^{k}D_{c}^{(i)},\quad D_{c}^{(i)}\cap D_{c}^{(j)}=\varnothing \quad (i\neq j) \; 。 $
在每个子集上,客户端以当前全局模型参数$ \boldsymbol{\theta}_{\rm{g}lobal} $为起点,执行本地梯度下降,计算相应的参数更新量:
$ \delta \boldsymbol \theta _{i}^{(c)}=-\eta \nabla L({\boldsymbol \theta }_{\rm global},D_{c}^{(i)}) \;,$
其中$ \eta $为学习率,$ L(\cdot ) $为损失函数。由此,客户端生成了$ k $个具有多样性的候选更新$ \{\delta \boldsymbol\theta _{1}^{(c)},\delta \boldsymbol\theta _{2}^{(c)},… ,\delta \boldsymbol\theta _{k}^{(c)}\} $,这些候选反映了局部数据不同子分布下的优化方向,为后续的模型更新提供了多样化的候选空间。
为便于后续量子态编码,对每个候选更新进行$ {L}_{2} $归一化:
$ \widetilde{\delta \boldsymbol \theta }_{i}^{(c)}=\dfrac{\delta \boldsymbol \theta _{i}^{(c)}}{\| \delta \boldsymbol \theta _{i}^{(c)}{\| }_{2}}\; 。 $
归一化操作确保了各候选更新在量子态振幅分配中具有可比性,避免了因更新幅度差异悬殊而导致的振幅权重失衡。
2)评分计算
为了评估每个候选更新的本地性能,客户端使用预留的本地验证集$ D_{c}^{\rm val} $对各候选进行评分。具体地,将候选更新施加于全局模型后,在验证集上计算模型的损失值,以其负值作为性能评分:
$ {score}_{i}=-L\left({\boldsymbol \theta }_{\rm global}+\widetilde{\delta \boldsymbol \theta }_{i}^{(c)},D_{c}^{\rm val}\right)\; 。 $
评分越高,说明该候选更新在本地验证集上使模型性能提升越显著,即该候选更新的质量越优。这一评分机制使得客户端能够在不暴露原始数据的前提下,以隐私保护的方式提供候选质量信息。
3)振幅权重分配
基于候选更新的性能评分,客户端采用带温度参数$ \beta $的softmax函数为各候选分配量子态振幅权重:
$ {\alpha }_{i}=\dfrac{\exp(\beta \cdot {score}_{i})}{\displaystyle\sum_{j=1}^{k}\exp\left(\beta \cdot {score}_{j}\right)}\;,$
其中温度参数$ \beta \gt 0 $控制权重分配的“软硬”程度:当$ \beta \rightarrow 0 $时,各候选获得均等权重,退化为均匀叠加;当$ \beta \rightarrow \infty $时,权重集中于评分最高的候选,趋近于确定性选择。通过调节$ \beta $,系统可以在探索多样性与利用最优解之间取得平衡。权重$ {\alpha }_{i} $满足归一化条件$ \displaystyle\sum_{i=1}^{k}{\alpha }_{i}=1 $,可直接用作量子态的振幅系数。
4)量子态编码与传输
客户端将$ k $个归一化候选更新及其对应的振幅权重编码为一个联合量子叠加态,其构造形式为:
$ |{\mathit{\Psi }}_{c}\rangle =\displaystyle\sum_{i=1}^{k}{\alpha }_{i}|i{\rangle }_{\rm index}\otimes |\widetilde{\delta \boldsymbol \theta }_{i}^{(c)}{\rangle }_{\rm param}\; 。 $
该量子态由2个寄存器张量积构成。其中,索引寄存器使用$ \lceil {\rm lb} k\rceil $个量子比特编码候选序号$ i $,参数寄存器使用$\lceil {\rm lb} D \rceil $个量子比特以振幅编码方式存放对应的归一化更新向量$ \widetilde{\delta \boldsymbol \theta }_{i}^{(c)} $。整个联合态$ |{\mathit{\Psi }}_{c}\rangle $是$ k $个候选更新的相干叠加,各候选的振幅由其性能评分决定,从而将性能偏好信息内嵌于量子态的幅度结构之中。
最终,客户端将编码完成的量子叠加态$ |{\mathit{\Psi }}_{c}\rangle $通过量子信道一次性发送至服务器,实现了多候选更新的并行传输,避免了逐一传输所带来的通信开销累积。以量子比特数计,承载$ k $个D维向量候选更新仅需索引寄存器$\lceil {\rm lb} k\rceil $个、参数寄存器$\lceil{\rm lb}D\rceil $个量子比特,相比经典方案逐一传输$ k $个D维向量的线性代价,其随候选数的增长由线性降为对数;需要说明的是,振幅编码的态制备与读出开销仍与D维向量相关,本文在仿真层面对该过程进行建模。
服务端目标是在接收到各客户端传输的量子叠加态后,通过引入历史全局更新方向作为测量偏好,以方向一致性投影测量的方式从叠加态中智能地选出最优候选更新,从而在提升聚合稳定性的同时加速全局模型的收敛。
1)历史方向计算
为了获取全局优化的历史轨迹,服务器记录过去$ T $轮的全局模型更新,并对其进行时间平均以获得历史方向参考向量:
$ {\boldsymbol d}_{\rm hist}=\dfrac{1}{T}\displaystyle\sum_{t'=t-T}^{t-1}{\varDelta }{\boldsymbol \theta }^{(t')}\;,$
其中$ \mathit{\Delta }{\theta }^{(t')} $表示第$ t' $轮的全局更新量。历史方向向量$ {\boldsymbol d}_{\rm hist} $反映了近期全局优化的主导趋势,代表了模型参数在参数空间中的历史运动方向。通过引入这一历史信息,服务器能够识别并优先选取与当前优化轨迹方向一致的客户端候选更新,从而避免因单一轮次的噪声或数据异质性而导致优化方向的大幅偏转。
2)方向相似度计算
服务器对每个客户端$ c $传输的叠加态中的各候选更新$ \widetilde{\delta \boldsymbol \theta }_{i}^{(c)} $,计算其与历史方向向量$ {\boldsymbol d}_{\rm hist} $之间的余弦相似度,以量化该候选更新与历史全局优化方向的一致程度:
$ {sim}_{i}=\dfrac{\widetilde{\delta \boldsymbol \theta }_{i}^{(c)}\cdot {\boldsymbol d}_{\rm hist}}{\| \widetilde{\delta \boldsymbol \theta }_{i}^{(c)}\| \cdot \| {\boldsymbol d}_{\rm hist}\| }\; 。 $
余弦相似度$ {sim}_{i}\in [-1{,}1] $:当$ {sim}_{i} $接近$ 1 $时,表明该候选更新与历史方向高度一致,有助于延续当前的优化轨迹;当$ {sim}_{i} $接近$ -1 $时,则表明该候选与历史方向相悖,可能导致优化轨迹的振荡或偏转。通过这一度量,服务器能够在量子测量阶段对方向一致性高的候选赋予更大的选中概率。
3)综合测量权重计算
服务器将候选更新的本地性能评分与方向相似度综合考量,计算最终用于调整量子态振幅的综合测量权重:
$ {p}_{i}=\dfrac{\exp(\beta \cdot {score}_{i}+\gamma \cdot {sim}_{i})}{\displaystyle\sum_{j=1}^{k}\exp\left(\beta \cdot {score}_{j}+\gamma \cdot {sim}_{j}\right)}\;,$
其中$ \beta $为性能偏好系数,控制本地验证集性能在选择决策中的权重;$ \gamma $为方向一致性系数,控制历史更新方向偏好的影响强度。
4)量子测量与最优候选坍缩
服务器将量子叠加态索引寄存器中各候选的振幅修改为综合测量权重$ \sqrt{{p}_{i}} $(以确保概率归一化),并定义索引寄存器的投影算子:
$ {P}_{i}=|i\rangle \langle i{|}_{\rm index}\otimes {I}_{\rm param}\; 。 $
随后对整个联合量子态执行计算基投影测量。根据量子测量的概率性原理,测量结果以概率$ {p}_{i} $坍缩到候选$ i $所对应的参数态:
$ |{\mathit{\Psi }}_{c}\rangle \xrightarrow{\rm measure}|{i}^{*}{\rangle }_{\rm index}\otimes |\widetilde{\delta \boldsymbol \theta }_{{i}^{*}}^{(c)}{\rangle }_{\rm param}\;,$
其中$ {i}^{*}=\underset{i}{\arg\,\max } {p}_{i} $为测量获得的候选序号。测量完成后,服务器从参数寄存器中提取对应的最优候选更新$ \widetilde{\delta \boldsymbol \theta }_{{i}^{*}}^{(c)} $来作为客户端$ c $本轮的有效更新贡献。这一投影测量过程在量子域内直接完成了候选选优过程,无需对所有候选逐一进行经典比较,从而在量子域内完成了测量聚合。
5)全局模型聚合更新
在完成对所有$ N $个客户端量子态的投影测量后,服务器收集各客户端的最优候选更新,采用联邦平均的方式聚合得到本轮的全局更新:
$ {\boldsymbol\theta }_{\rm global}\leftarrow {\boldsymbol\theta }_{\rm global}+\dfrac{1}{N}\displaystyle\sum_{c=1}^{N}\widetilde{\delta \boldsymbol \theta }_{{i}^{*}}^{(c)}\; 。 $
更新后的全局模型$ {\boldsymbol\theta }_{\rm global} $将在下一轮训练开始时重新分发至各客户端,开启新一轮的量子联邦学习迭代。
1)实验环境与仿真框架。所有实验均在配备NVIDIA RTX 4090 GPU的服务器集群上完成。在软件层面,本文基于PennyLane量子计算框架构建变分量子电路并模拟客户端量子态的制备、传输与服务端的投影测量过程,基于Flower联邦学习框架实现客户端-服务器的分布式协同训练与聚合调度。
2)数据集。实验在3个标准图像分类数据集上进行评估,涵盖灰度与彩色、简单与复杂2类难度梯度,以全面考察方法在不同数据复杂度下的适应性。其中MNIST与FMNIST均为28×28的单通道灰度图像,各包含6万张训练样本与1万张测试样本,单样本为784特征维度,分别对应手写数字与服饰商品的十分类任务,难度相对较低;CIFAR-10为32×32的3通道彩色RGB图像,包含5万张训练样本与1万张测试样本,单样本为3072特征维度,对应普适物体的十分类任务,视觉特征更为复杂。
3)经典-量子混合模型。为适配不同数据集的复杂度并验证QStar在不同模型架构上的普适性,实验采用了2种规模的经典-量子混合模型,其结构均由经典特征提取器与变分量子层级联构成。经典部分负责高维图像特征的降维与提取,量子部分则以变分量子电路替代经典分类头完成最终决策。在难度较低的MNIST与FMNIST数据集上,采用轻量级MobileNetV2[35]作为特征提取器,并级联全连接强纠缠量子层,包含220万余个可冻结经典层参数、1万余个经典-量子转换参数和56个量子参数;在难度较高的CIFAR-10数据集上,采用表达能力更强的ResNet-18[36]作为特征提取器,并级联基于$ {U}_{\text{SU}\left(4\right)} $门的分层卷积-池化量子层,包含1120万余个可冻结经典层参数、1.2万余个经典-量子转换参数和297个量子参数,以匹配彩色图像更丰富的特征结构。
4)对比基线方法。实验将QStar与6种代表性基线方法进行对比,涵盖经典联邦学习与量子联邦学习两大类。经典联邦学习方法选取了4种聚合策略各异的代表:FedAvg[27]采用标准的按样本量加权平均,是最基础的联邦聚合范式;FedBN[37]让客户端在本地保留批量归一化层而仅聚合其余参数,以缓解客户端间的特征分布偏移;FedPer[38]让客户端本地保留个性化经典层而仅聚合共享层,侧重个性化建模;FedProx[28]则在本地训练目标中引入近端正则项,从优化角度抑制客户端漂移。量子联邦学习方法选取了2种基于参数级压缩的代表:DAQFL[39]基于动态振幅量化降低上传比特数,WPQFL[40]基于加权原型量化进行通信压缩;二者均沿用“量子训练-经典传输”的范式,可与QStar的量子原生通信形成直接对照。为保证对比公平,除FedBN与FedPer因方法设计需保持经典层可训练外,其余所有方法均冻结经典特征提取器及仅训练量子层参数,从而将性能差异聚焦于量子参数的通信与聚合环节。
5)评估指标。围绕QStar在精度与通信2个维度的设计目标,实验采用以下指标进行量化评估。在模型精度方面,以测试集准确率(Accuracy)衡量,该值越高表明分类性能越好;在通信开销方面,以单轮平均传输比特数衡量,即每轮通信中终端与服务器之间交换的平均比特数,该值越低表明通信开销越小。
6)超参数设置。为保证实验的可复现性与对比的公平性,所有方法均采用统一超参数配置。系统默认包含50个客户端,每轮随机选取50%的客户端参与训练与聚合,本地训练轮数为3,全局通信轮数为20,优化器学习率设为0.01,批次大小为32。
1)准确性分析
表1汇总了QStar与6种基线方法在3个数据集、二分类与四分类2种任务设置下的最终测试准确率。在更具挑战性的四分类任务中,QStar在MNIST、FMNIST与CIFAR-10上均取得较优结果,分别达到97.25%、88.24%与73.56%。稳定优于FedPer和DAQFL。综合3个数据集,QStar的准确率虽略低于FedProx和WPQFL,但仍处于第1梯队。这是因为实验在Non-IID分布下进行[41],FedProx在Non-IID下近端项惩罚带来稳定收敛能力,而WPQFL保留一定本地参数保留个性化参数,使得这2类策略有更优的表现。QStar与FedAvg、FedBN之间的差异则主要来自FedBN自身的不稳定性与Non-IID分布带来的训练波动。
QStar相对同类量子联邦学习方法具有以下优势:在全部3个四分类任务上,QStar均稳定超越基于参数级压缩的DAQFL,即在MNIST上QStar达到97.25%,较DAQFL的95.05%高出2.20个百分点;在FMNIST上达到88.24%,较DAQFL的67.33%高出20.91个百分点;在CIFAR-10上达到73.56%,较DAQFL的63.10%高出10.46个百分点。尤其是DAQFL在四分类任务上普遍出现严重退化,尤其在FMNIST与CIFAR-10等更复杂的数据集上准确率大幅下滑,表明其动态振幅量化在类别数增多、决策边界更复杂时引入了不可忽视的信息损失;而QStar以量子原生的候选叠加传输,从根本上规避了“量子训练-经典传输”范式中逐参数量化与截断所带来的精度损耗。此外,FedPer因仅聚合共享层、本地保留个性化经典层,在四分类任务上各数据集均出现显著崩塌,在MNIST、FMNIST与CIFAR-10上分别仅为34.62%、33.66%与56.63%,说明其个性化设计虽利于本地适配,却难以形成强泛化的全局模型。FedBN在MNIST和CIFAR-10数据集上也出现崩塌,准确率分别为34.31%与46.59%,说明其本地保留批量归一化层的设计,在经典层与量子层混合的架构下难以稳定发挥作用。
在相对简单的二分类任务中,多数方法均逼近性能上限,方法间差异主要体现在彩色复杂数据上。QStar在MNIST与FMNIST上均达到100.00%,并在CIFAR-10上以97.50%的准确率取得最优,反映出在低类别数、强可分场景下,候选叠加所带来的探索增益相对饱和,方法间的准确率排序更多取决于具体优化策略而非聚合范式。综合来看,QStar在难度更高、更能体现方法区分度的四分类与复杂彩色任务上展现出稳定的领先性,而其准确性优势主要来源于“客户端多候选生成—服务端方向一致性投影选优”的协同机制:客户端在多个互不重叠的数据子集上生成具有多样性的候选更新,使每一轮上传都蕴含对局部数据不同子分布的多种优化方向;服务端则结合本地验证评分与历史优化方向的余弦相似度对候选进行综合加权与投影测量选优,从而在数据异质、特征复杂的任务中筛选出与全局优化轨迹一致的高质量更新,有效抑制了单一更新受局部数据偏置干扰而导致的精度下滑。
2)训练过程
图4~15给出了各策略在3个数据集、2种任务分类设置和2种Non-IID程度下随通信轮次变化的准确率曲线,可从收敛速度与收敛稳定性2个角度刻画QStar的训练动态。
在收敛速度方面,相比其他聚合策略,QStar在训练初期即表现出更快的准确率爬升,往往能在更少的通信轮次内逼近其最终准确率。这是因为服务端在每一轮聚合中引入历史全局更新方向作为测量偏好,对各客户端候选更新计算其与历史方向的余弦相似度,并优先选取与既有优化轨迹一致的候选,使全局模型沿主导优化方向持续、稳定地前进,避免了因单轮噪声或客户端数据异质而产生的方向反复与回退。换言之,方向一致性投影测量相当于在量子测量阶段对“是否延续当前优化趋势”做了一次隐式甄别,使得每一轮的有效更新都更贴合全局最优方向,从而在通信轮次受限的条件下更快地积累有效进展。
在收敛稳定性方面,QStar在收敛后的准确率波动较小,曲线更为平滑,而部分基线方法(如DAQFL、FedPer)在训练后期常出现明显的准确率抖动甚至回落。这说明方向一致性投影测量在抑制Non-IID引起的更新冲突方面发挥了平滑作用:当个别客户端的本地更新因数据偏置而偏离全局趋势时,其候选与历史方向的余弦相似度下降,相应地,被投影测量选中的概率降低,从而减小了异常更新对全局模型的扰动。同时,客户端侧基于本地验证评分的softmax振幅分配进一步保证了被纳入叠加态的候选本身具有较优的本地性能,二者共同构成了“质量筛选+方向筛选”的双重稳健机制,使QStar在训练后期能够维持平稳收敛,而非在多个次优解之间震荡。这一稳定性优势在更易受异质性影响的复杂数据集上尤为明显。
3)通信效率
在MNIST数据集上统计各方法的单轮平均传输比特数。服务器端每轮聚合遍历收到参数中的每个张量并累加其字节长度,训练结束时除以聚合轮数得到单轮平均值。冻结经典特征提取器及仅训练量子层的方法单轮传输量在1.03~1.05 Mb量级,而FedBN与FedPer因需同步规模达数百万参数的经典骨干网络,单轮传输量分别达227.8 Mb与226.8 Mb,高出2个数量级。QStar与其他冻结经典层的方法处于同一量级,其设计优势在于承载多候选的代价随候选数对数增长,而非单轮绝对比特数的降低;DAQFL通过参数级量化可将单轮传输量进一步压缩至0.61 Mb,但伴随精度下降与收敛不稳定。
4)消融实验
图16给出了QStar与2个对照组在MNIST和CIFAR-10数据集上的最终测试准确率。在w/o多候选,客户端不再划分互不重叠的数据子集,也不生成候选集,而是在完整本地数据上执行标准的联邦平均式本地训练,每轮仅产生单一确定性更新;服务端保留历史方向机制,按各客户端更新与历史方向参考向量的余弦相似度对其加权聚合。在w/o方向感知中,客户端仍在多个数据子集上生成$ k $个候选更新并编码为量子叠加态;服务端不再计算历史方向参考向量,综合测量权重退化为仅由客户端上传的本地验证评分决定,即式(18)中取$ \gamma =0 $。
2个对照组在2个数据集上均低于完整的QStar,且方向感知的影响更为显著。在CIFAR-10上,QStar的准确率为89.41%,替换多候选生成后降至81.18%,降幅8.23个百分点;替换方向感知后降至66.67%,降幅达22.74个百分点。在MNIST上三者分别为97.14%、95.33%与91.35%,对应降幅为1.81与5.79个百分点。这表明服务端的方向感知选优是QStar在高异质性、复杂特征数据上保持精度的主导因素:其通过引入历史全局优化方向对各候选进行余弦相似度加权与投影选优,能够抑制Non-IID分布下由局部数据偏置引起的更新偏差;客户端多候选生成则为该选优机制提供候选空间,当其被替换为单一确定性更新后,选优空间收缩,精度随之下滑。
本文的验证工作存在以下局限。一方面,量子态的制备、传输与投影测量均在PennyLane中建模,未在真实量子硬件与量子信道上验证,实际部署还需考虑门保真度、退相干与传输损耗的影响。另一方面,多候选生成使客户端本地训练开销随候选数线性增长,本文以终端算力换取聚合质量,该权衡在算力受限的终端上仍需进一步评估。
本文面向动态异构网络中异质终端的协同学习需求,提出量子原生联邦聚合算法QStar,构建了以量子态为载体、贯通客户端编码、量子态传输与服务端聚合的端到端框架。在客户端,通过多候选量子叠加编码,将不同局部数据所产生的候选更新联合传输;在服务端,通过方向一致性投影测量聚合,将候选性能与历史全局优化方向相结合,实现量子域内的候选选择与模型更新。基于3个标准图像分类数据集的实验结果表明,QStar在收敛稳定性和复杂分类任务准确率方面具有竞争力,验证了所提方法在异质数据场景下的有效性。
蔡子诺负责方案整体设计、代码实现并撰写论文;朱海威、汪启瑞、银泓锐负责部分算法思路和代码实现;马汝辉提出指导意见并修改论文。

参考文献 引证文献
排序方式:
1
Alsagheer D, Xu Lei, Shi Weidong. Decentralized machine learning governance: Overview, opportunities, and challenges[J]. IEEE Access, 2023, 11: 96718−96732
2
Albrecht J P. How the GDPR will change the world[J]. European Data Protection Law Review, 2016, 2: 287
3
Fallah A, Mokhtari A, Ozdaglar A. Personalized federated learning with theoretical guarantees: A model-agnostic meta-learning approach[J]. Advances in Neural Information Processing Systems, 2020, 33: 3557−3568
4
冯奕铭, 钱珍, 李光辉, 等. 异构边缘环境下自适应分层联邦学习协同优化方法[J]. 计算机研究与发展, 2025, 62(6): 1416−1433
Feng Yiming, Qian Zhen, Li Guanghui, et al. Synergistic optimization method for adaptive hierarchical federated learning in heterogeneous edge environments[J]. Journal of Computer Research and Development, 2025, 62(6): 1416−1433 (in Chinese)
5
Nguyen D C, Ding M, Pathirana P N, et al. Federated learning for Internet of things: A comprehensive survey[J]. IEEE Communications Surveys & tutorials, 2021, 23(3): 1622−1658
6
Liu Dianqi, Bai Liang, Yu Tianyuan, et al. Towards method of horizontal federated learning: A survey[C]//Proc of 2022 8th Int Conf on Big Data and Information Analytics (BigDIA). Piscataway, NJ: IEEE, 2022: 259−266
7
Li Qinbin, Wen Zeyi, Wu Zhaomin, et al. A survey on federated learning systems: Vision, hype and reality for data privacy and protection[J]. IEEE Transactions on Knowledge and Data Engineering, 2023, 35(4): 3347−3366
8
Yan Rui, Qu Liangqiong, Wei Qingyue, et al. Label-efficient self-supervised federated learning for tackling data heterogeneity in medical imaging[J]. IEEE Transactions on Medical Imaging, 2023, 42(7): 1932−1943
9
Duan Moming, Liu Duo, Chen Xianzhang, et al. Astraea: Self-balancing federated learning for improving classification accuracy of mobile deep learning applications[C]//Proc of the 37th Int Conf on Computer Design (ICCD). Piscataway, NJ: IEEE, 2019: 246−254
10
Jiang Wenxian, Chen Mengjuan, Tao Jun. Federated learning with blockchain for privacy-preserving data sharing in Internet of vehicles[J]. China Communications, 2023, 20(3): 69−85
11
Dang Trung Kien, Lan Xiang, Weng Jianshu, et al. Federated learning for electronic health records[J]. ACM Transactions on Intelligent Systems and Technology, 2022, 13(5): 1−17
12
李程, 袁勇, 郑志勇, 等. 基于区块链的联邦学习: 模型、方法与应用[J]. 自动化学报, 2024, 50(6): 1059−1085
Li Cheng, Yuan Yong, Zheng Zhiyong, et al. Blockchain-enabled federated learning: Models, methods and applications[J]. Acta Automatica Sinica, 2024, 50(6): 1059−1085 (in Chinese)
13
Wu J, Dong F, Leung H, et al. Topology-aware federated learning in edge computing: A comprehensive survey[J]. ACM Computing Surveys, 2024, 56(10): 1−41
14
Pfeiffer K, Rapp M, Khalili R, et al. Federated learning for computationally constrained heterogeneous devices: A survey[J]. ACM Computing Surveys, 2023, 55(14s): 1−27
15
Gill S S, Kumar A, Singh H, et al. Quantum computing: A taxonomy, systematic review and future directions[J]. Software: Practice and Experience, 2022, 52(1): 66−114
16
Schuld M, Killoran N. Quantum machine learning in feature Hilbert spaces[J]. Physical Review Letters, 2019, 122(4): 040504
17
Biamonte J, Wittek P, Pancotti N, et al. Quantum machine learning[J]. Nature, 2017, 549(7671): 195−202
18
Uddin M R, Shaon S, Rahman R, et al. Quantum federated learning: A comprehensive survey[J]. IEEE Communications Surveys & Tutorials, 2025, 28: 3942−3975
19
Banchi L, Pereira J, Pirandola S. Generalization in quantum machine learning: A quantum information standpoint[J]. PRX Quantum, 2021, 2(4): 040321
20
Wei Wenqi, Liu Ling, Wu Yanzhao, et al. Gradient-leakage resilient federated learning[C]//Proc of the 41st Int Conf on Distributed Computing Systems (ICDCS). Piscataway, NJ: IEEE, 2021: 797−807
21
Alistarh D, Grubic D, Li J, et al. QSGD: Communication-efficient SGD via gradient quantization and encoding[J]. Advances in Neural Information Processing Systems, 2017, 30: 1709−1720
22
Lin Yujun, Han Song, Mao Huizi, et al. Deep gradient compression: Reducing the communication bandwidth for distributed training[C]//Proc of Int Conf on Learning Representations (ICLR). 2018. https://openreview.net/forum?id=SkhQHMW0W
23
Long Zhong, Chen Yuling, Dou Hui, et al. FedSQ: Sparse-quantized federated learning for communication efficiency[J]. IEEE Transactions on Consumer Electronics, 2024, 70(1): 4050−4061
24
Rothchild D, Panda A, Ullah E, et al. FetchSGD: Communication-efficient federated learning with sketching[C]//Proc of Int Conf on Machine Learning. New York: PMLR, 2020: 8253−8265
25
Guan Yixuan, Liu Xuefeng, Niu Jianwei, et al. FedTC: Enabling communication-efficient federated learning via Transform coding[C]//Proc of IEEE INFOCOM 2024-IEEE Conf on Computer Communications. Piscataway, NJ: IEEE, 2024: 821−830
26
王鑫澳, 陈珂, 寿黎但, 等. 基于联邦学习的BERT模型高效训练框架[J]. 软件学报, 2025, 36(9): 4110−4133
Wang Xinao, Chen Ke, Shou Lidan, et al. Efficient framework for BERT model training based on federated learning[J]. Journal of Software, 2025, 36(9): 4110−4133 (in Chinese)
27
McMahan B, Moore E, Ramage D, et al. Communication-efficient learning of deep networks from decentralized data[C]//Proc of Artificial Intelligence and Statistics. New York: PMLR, 2017: 1273−1282
28
Li Tian, Sahu A K, Zaheer M, et al. Federated optimization in heterogeneous networks[J]. Proceedings of Machine Learning and Systems, 2020, 2: 429−450
29
Karimireddy S P, Kale S, Mohri M, et al. SCAFFOLD: Stochastic controlled averaging for federated learning[C]//Proc of Int Conf on Machine Learning. New York: PMLR, 2020: 5132−5143
30
Li Qinbin, He Bingsheng, Song D. Model-contrastive federated learning[C]//Proc of the IEEE/CVF Conf on Computer Vision and Pattern Recognition. Piscataway, NJ: IEEE, 2021: 10713−10722
31
杜甜, 陈星延, 寇纲, 等. 面向云边个性化模型解耦的聚类联邦学习方法[J]. 计算机学报, 2025, 48(2): 407−432
Du Tian, Chen Xingyan, Kou Gang, et al. Clustered federated learning with cloud-edge personalized model decoupling[J]. Chinese Journal of Computers, 2025, 48(2): 407−432 (in Chinese)
32
Daskin A. Federated learning with distributed fixed design quantum chips and quantum channels[J]. arXiv preprint arXiv: 2401.13421, 2024
33
Park S, Jung S, Kim J. Dynamic quantum federated learning for satellite-ground integrated systems using slimmable quantum neural networks[J]. IEEE Access, 2024, 12: 58239−58247
34
Havlíček V, Córcoles A D, Temme K, et al. Supervised learning with quantum-enhanced feature spaces[J]. Nature, 2019, 567(7747): 209−212
35
Sandler M, Howard A, Zhu Menglong, et al. MobileNetV2: Inverted residuals and linear bottlenecks[C]//Proc of 2018 IEEE/CVF Conf on Computer Vision and Pattern Recognition. Piscataway, NJ: IEEE, 2018: 4510−4520
36
He Kaiming, Zhang Xiangyu, Ren Shaoqing, et al. Deep residual learning for image recognition[C]//Proc of the IEEE Conf on Computer Vision and Pattern Recognition. Piscataway, NJ: IEEE, 2016: 770−778
37
Li Xiaoxiao, Jiang Meirui, Zhang Xiaofei, et al. FedBN: Federated learning on Non-IID features via local batch normalization[C]//Proc of Int Conf on Learning Representations (ICLR). 2021. https://openreview.net/forum?id=6YEQUn0QICG
38
Arivazhagan M G, Aggarwal V, Singh A K, et al. Federated learning with personalization layers[J]. arXiv preprint arXiv: 1912.00818, 2019
39
Qu Z, Zhao X, Sun L, et al. DAQFL: Dynamic aggregation quantum federated learning algorithm for intelligent diagnosis in Internet of medical things[J]. IEEE Internet of Things Journal, 2025, 12(19): 39313−39325
40
Gurung D, Pokhrel S R. Performance analysis and design of a weighted personalized quantum federated learning[J]. IEEE Transactions on Artificial Intelligence, 2025, 6(8): 2302−2313
41
Li Xiang, Huang Kaixuan, Yang Wenhao, et al. On the convergence of FedAvg on Non-IID data[C]//Proc of Int Conf on Learning Representations (ICLR). 2020. https://openreview.net/forum?id=HJxNAnVtDS
2026年第63卷第10期
86
40
0
引用本文
BibTeX
文章信息
doi: 10.7544/issn1000-1239.202660474
  • 接收时间:2026-06-18
  • 首发时间:2026-11-17
  • 出版时间:2026-10-01
补充材料
相关文章
文章信息
作者
出版历史
  • 收稿日期:2026-06-18
  • 修回日期:2026-09-21
基金
作者信息
    上海交通大学计算机学院 上海 200240

通讯作者:

马汝辉()
参考文献
分享链接
https://castjournals.cast.org.cn/joweb/jsjyjyfz/CN/10.7544/issn1000-1239.202660474
分享至
全文二维码

扫描看全文

引用本文
BibTeX
本文的引用情况
2种不同金属材料的力学参数
科
Family
属数
Number of
genus
种数
Number of
species
占总种数比例
Percentage of
total species (%)
属
Genus
种数
Number of
species
占总种数比例
Percentage of total
species (%)
鹅膏菌科Amanitaceae 2 11 5.26 鹅膏菌属 Amanita 10 4.78
小菇科 Mycenaceae 2 12 5.74 丝盖伞属 Inocybe 5 2.39
多孔菌科 Polyporaceae 8 14 6.70 蜡蘑属 Laccaria 5 2.39
红菇科 Russulaceae 3 23 11.00 小皮伞属 Marasmius 6 2.87
小菇属 Mycena 11 5.26
光柄菇属 Pluteus 5 2.39
红菇属 Russula 17 8.13
栓菌属 Trametes 5 2.39
关闭全屏