典型文献
A Federated Domain Adaptation Algorithm Based on Knowledge Distillation and Contrastive Learning
文献摘要：
Smart manufacturing suffers from the heterogeneity of local data distribution across parties,mutual information silos and lack of privacy protection in the process of industry chain collabo-ration.To address these problems,we propose a federated domain adaptation algorithm based on knowledge distillation and contras-tive learning.Knowledge distillation is used to extract transferable integration knowledge from the different source domains and the quality of the extracted integration knowledge is used to assign reasonable weights to each source domain.A more rational weighted average aggregation is used in the aggregation phase of the center server to optimize the global model,while the local model of the source domain is trained with the help of contrastive learning to constrain the local model optimum towards the global model optimum,mitigating the inherent heterogeneity between lo-cal data.Our experiments are conducted on the largest domain ad-aptation dataset,and the results show that compared with other tra-ditional federated domain adaptation algorithms,the algorithm we proposed trains a more accurate model,requires fewer communi-cation rounds,makes more effective use of imbalanced data in the industrial area,and protects data privacy.
文献关键词：
中图分类号：
作者姓名：
HUANG Fang;FANG Zhijun;SHI Zhicai;ZHUANG Lehui;Li Xingchen;HUANG Bo
作者机构：
School of Electrical and Electronic Engineering,Shanghai University of Engineering Science,Shanghai 201600,China
文献出处：
引用格式：
[1]HUANG Fang;FANG Zhijun;SHI Zhicai;ZHUANG Lehui;Li Xingchen;HUANG Bo-.A Federated Domain Adaptation Algorithm Based on Knowledge Distillation and Contrastive Learning)[J].武汉大学自然科学学报（英文版）,2022(06):499-507
A类：
contras,aptation
B类：
Federated,Domain,Adaptation,Algorithm,Based,Knowledge,Distillation,Contrastive,Learning,Smart,manufacturing,suffers,from,heterogeneity,local,distribution,across,parties,mutual,information,silos,lack,privacy,protection,process,industry,chain,collabo,To,address,these,problems,federated,adaptation,knowledge,distillation,learning,used,transferable,integration,different,source,domains,quality,extracted,assign,reasonable,weights,each,more,rational,weighted,average,aggregation,phase,center,server,optimize,global,model,while,trained,help,contrastive,constrain,optimum,towards,mitigating,inherent,between,Our,experiments,conducted,largest,dataset,results,show,that,compared,other,ditional,algorithms,proposed,trains,accurate,requires,fewer,communi,cation,rounds,makes,effective,imbalanced,industrial,area,protects
AB值：
0.584376
相似文献
机标中图分类号，由域田数据科技根据网络公开资料自动分析生成，仅供学习研究参考。
