May, 2024
FedProK: 基于样本原型特征知识迁移的可信联邦增量学习
FedProK: Trustworthy Federated Class-Incremental Learning via Prototypical Feature Knowledge Transfer
Xin Gao, Xin Yang, Hao Yu, Yan Kang, Tianrui Li
TL;DRFederated Class-Incremental Learning (FCIL) aims to improve trustworthiness by transferring knowledge using FedProK, a method that leverages prototypical feature knowledge transfer, demonstrating superior performance in spatial-temporal knowledge transfer.