TY - GEN
T1 - Rethinking Personalized Federated Learning with Clustering-Based Dynamic Graph Propagation
AU - Wang, Jiaqi
AU - Chen, Yuzhong
AU - Wu, Yuhang
AU - Das, Mahashweta
AU - Yang, Hao
AU - Ma, Fenglong
N1 - Publisher Copyright:
© The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2024.
PY - 2024
Y1 - 2024
N2 - Most existing personalized federated learning approaches are based on intricate designs, which often require complex implementation and tuning. In order to address this limitation, we propose a simple yet effective personalized federated learning framework. Specifically, during each communication round, we group clients into multiple clusters based on their model training status and data distribution on the server side. We then consider each cluster center as a node equipped with model parameters and construct a graph that connects these nodes using weighted edges. Additionally, we update the model parameters at each node by propagating information across the entire graph. Subsequently, we design a precise personalized model distribution strategy to allow clients to obtain the most suitable model from the server side. We conduct experiments on three image benchmark datasets and create synthetic structured datasets with three types of typologies. Experimental results demonstrate the effectiveness of the proposed FedCedar.
AB - Most existing personalized federated learning approaches are based on intricate designs, which often require complex implementation and tuning. In order to address this limitation, we propose a simple yet effective personalized federated learning framework. Specifically, during each communication round, we group clients into multiple clusters based on their model training status and data distribution on the server side. We then consider each cluster center as a node equipped with model parameters and construct a graph that connects these nodes using weighted edges. Additionally, we update the model parameters at each node by propagating information across the entire graph. Subsequently, we design a precise personalized model distribution strategy to allow clients to obtain the most suitable model from the server side. We conduct experiments on three image benchmark datasets and create synthetic structured datasets with three types of typologies. Experimental results demonstrate the effectiveness of the proposed FedCedar.
UR - http://www.scopus.com/inward/record.url?scp=85192792601&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85192792601&partnerID=8YFLogxK
U2 - 10.1007/978-981-97-2259-4_12
DO - 10.1007/978-981-97-2259-4_12
M3 - Conference contribution
AN - SCOPUS:85192792601
SN - 9789819722617
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 155
EP - 167
BT - Advances in Knowledge Discovery and Data Mining - 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024, Proceedings
A2 - Yang, De-Nian
A2 - Xie, Xing
A2 - Tseng, Vincent S.
A2 - Pei, Jian
A2 - Huang, Jen-Wei
A2 - Lin, Jerry Chun-Wei
PB - Springer Science and Business Media Deutschland GmbH
T2 - 28th Pacific-Asia Conference on Knowledge Discovery and Data Mining, PAKDD 2024
Y2 - 7 May 2024 through 10 May 2024
ER -