Generalizable Heterogeneous Federated Cross-Correlation and Instance Similarity Learning
IEEE Transactions on Pattern Analysis and Machine Intelligence · 2024-02 · 46 (2) · 712-728
DOI: 10.1109/tpami.2023.3327373 ↗Kurz gefasst
Fachartikel in der Zeitschrift IEEE Transactions on Pattern Analysis and Machine Intelligence, 2024-02. Verfasst von Wenke Huang und 3 weiteren Autor:innen. Die Arbeit wurde bisher 78-mal zitiert (Crossref) und nennt 118 Literaturangaben. Gefördert von: National Natural Science Foundation of China, Key Research and Development Program of Hubei Province, Zhejiang lab.
Abstract (Originalsprache)
Federated learning is an important privacy-preserving multi-party learning paradigm, involving collaborative learning with others and local updating on private data. Model heterogeneity and catastrophic forgetting are two crucial challenges, which greatly limit the applicability and generalizability. This paper presents a novel FCCL+, federated correlation and similarity learning with non-target distillation, facilitating the both intra-domain discriminability and inter-domain generalization. For heterogeneity issue, we leverage irrelevant unlabeled public data for communication between the heterogeneous participants. We construct cross-correlation matrix and align instance similarity distribution on both logits and feature levels, which effectively overcomes the communication barrier and improves the generalizable ability. For catastrophic forgetting in local updating stage, FCCL+ introduces Federated Non Target Distillation, which retains inter-domain knowledge while avoiding the optimization conflict issue, fulling distilling privileged inter-domain information through depicting posterior classes relation. Considering that there is no standard benchmark for evaluating existing heterogeneous federated learning under the same setting, we present a comprehensive benchmark with extensive representative methods under four domain shift scenarios, supporting both heterogeneous and homogeneous federated settings. Empirical results demonstrate the superiority of our method and the efficiency of modules on various scenarios. The benchmark code for reproducing our results is available at https://github.com/WenkeHuang/FCCL.
Förderung
- National Natural Science Foundation of China
- Key Research and Development Program of Hubei Province
- Zhejiang lab
- CCF-Huawei Populus Grove Fund
- Special Fund of Hubei Luojia Laboratory
Zitieren
APA
Huang, W., Ye, M., Shi, Z., & Du, B. (2024). Generalizable Heterogeneous Federated Cross-Correlation and Instance Similarity Learning. IEEE Transactions on Pattern Analysis and Machine Intelligence, 46(2), 712-728. https://doi.org/10.1109/tpami.2023.3327373
Vancouver
Huang W, Ye M, Shi Z, Du B. Generalizable Heterogeneous Federated Cross-Correlation and Instance Similarity Learning. IEEE Transactions on Pattern Analysis and Machine Intelligence. 2024;46(2):712-728. doi:10.1109/tpami.2023.3327373
BibTeX
@article{Huang2024,
author = {Huang, Wenke and Ye, Mang and Shi, Zekun and Du, Bo},
title = {Generalizable Heterogeneous Federated Cross-Correlation and Instance Similarity Learning},
journal = {IEEE Transactions on Pattern Analysis and Machine Intelligence},
year = {2024},
volume = {46},
number = {2},
pages = {712--728},
doi = {10.1109/tpami.2023.3327373}
}
EUVIMED