FedTriNet: A Pseudo Labeling Method with Three Players for Federated Semi-supervised Learning

被引:26
作者
Che, Liwei [1 ]
Long, Zewei [2 ]
Wang, Jiaqi [1 ]
Wang, Yaqing [3 ]
Xiao, Houping [4 ]
Ma, Fenglong [1 ]
机构
[1] Penn State Univ, Coll IST, State Coll, PA 16801 USA
[2] Univ Illinois, Dept Comp Sci, Champaign, IL USA
[3] Purdue Univ, Sch Elect & Comp Engn, W Lafayette, IN USA
[4] Georgia State Univ, Inst Insight, Atlanta, GA USA
来源
2021 IEEE INTERNATIONAL CONFERENCE ON BIG DATA (BIG DATA) | 2021年
关键词
federated learning; semi-supervised learning; pseudo labeling;
D O I
10.1109/BigData52589.2021.9671374
中图分类号
TP18 [人工智能理论];
学科分类号
081104 ; 0812 ; 0835 ; 1405 ;
摘要
Federated Learning has shown great potentials for the distributed data utilization and privacy protection. Most existing federated learning approaches focus on the supervised setting, which means all the data stored in each client has labels. However, in real-world applications, the client data are impossible to be fully labeled. Thus, how to exploit the unlabeled data should be a new challenge for federated learning. Although a few studies are attempting to overcome this challenge, they may suffer from information leakage or misleading information usage problems. To tackle these issues, in this paper, we propose a novel federated semi-supervised learning method named FedTriNet, which consists of two learning phases. In the first phase, we pre-train FedTriNet using labeled data with FedAvg. In the second phase, we aim to make most of the unlabeled data to help model learning. In particular, we propose to use three networks and a dynamic quality control mechanism to generate high-quality pseudo labels for unlabeled data, which are added to the training set. Finally, FedTriNet uses the new training set to retrain the model. Experimental results on three publicly available datasets show that the proposed FedTriNet outperforms state-of-the-art baselines under both IID and Non-IID settings.
引用
收藏
页码:715 / 724
页数:10
相关论文
共 27 条
[1]  
Albaseer A., 2020, 2020 INT WIRELESS CO
[2]  
[Anonymous], 2013, ICML WORKSHOP CHALLE
[3]  
Athiwaratkun B., INT C LEARN REPR ICL
[4]  
Berthelot D., 2019, INT C LEARNING REPRE
[5]  
Bhowmick A., 2018, Protection against reconstruction and its applications in private federated learning
[6]  
Bonawitz K., 2016, NIPS WORKSHOP PRIVAT
[7]  
Caldas S, 2018, ARXIV
[8]  
Cascante-Bonilla P., 2020, CURRICULUM LABELING
[9]  
Chapelle O., 2009, IEEE Transactions on Neural Networks, V20, P542, DOI [DOI 10.1109/TNN.2009.2015974, 10.1109/TNN.2009.2015974]
[10]  
Chen F., 2018, ARXIV180207876