Federated Class-Incremental Learning with New-Class Augmented Self-Distillation

doi:10.1007/s11390-025-5186-5

CSpace

	Federated Class-Incremental Learning with New-Class Augmented Self-Distillation
	Wu, Zhi-Yuan 1,2; Sun, Sheng 1,2; He, Tian-Liu 1; Wang, Yu-Wei 1; Liu, Min 1,3; Gao, Bo 4; Jiang, Xue-Feng 1,2
	2025-09-01
发表期刊	JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY
ISSN	1000-9000
卷号	40 期号:5 页码:1427-1437
摘要	Federated learning (FL) enables collaborative model training among participants while guaranteeing the privacy of raw data. Mainstream FL methodologies overlook the dynamic nature of real-world data, particularly its tendency to grow in volume and diversify in classes over time. This oversight results in FL methods suffering from catastrophic forgetting, where the trained models inadvertently discard previously learned information upon assimilating new data. In response to this challenge, we propose a novel federated class-incremental learning (FCIL) method, named Federated Class-incremental Learning with New-Class Augmented Self-Distillation (FedCLASS). The core of FedCLASS is to enrich the class scores of historical models with new class scores predicted by current models and utilize the combined knowledge for self-distillation, enabling a more sufficient and precise knowledge transfer from historical models to current models. Theoretical analyses demonstrate that FedCLASS stands on reliable foundations, considering the scores of old classes predicted by historical models as conditional probabilities in the absence of new classes, and the scores of new classes predicted by current models as the conditional probabilities of class scores derived from historical models. Empirical experiments demonstrate the superiority of FedCLASS over four baseline algorithms in reducing average forgetting rate and boosting global accuracy.
关键词	federated learning class-incremental learning knowledge distillation
DOI	10.1007/s11390-025-5186-5
收录类别	SCI
语种	英语
WOS研究方向	Computer Science
WOS类目	Computer Science, Hardware & Architecture ; Computer Science, Software Engineering
WOS记录号	WOS:001621167000004
出版者	SPRINGER SINGAPORE PTE LTD
引用统计
文献类型	期刊论文
条目标识符	http://119.78.100.204/handle/2XEOYT63/43078
专题	中国科学院计算技术研究所
通讯作者	Wang, Yu-Wei
作者单位	1.Chinese Acad Sci, Inst Comp Technol, Beijing 100190, Peoples R China 2.Univ Chinese Acad Sci, Beijing 100190, Peoples R China 3.Zhongguancun Lab, Beijing 100190, Peoples R China 4.Beijing Jiaotong Univ, Sch Comp & Informat Technol, Beijing 100044, Peoples R China
推荐引用方式 GB/T 7714	Wu, Zhi-Yuan,Sun, Sheng,He, Tian-Liu,et al. Federated Class-Incremental Learning with New-Class Augmented Self-Distillation[J]. JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY,2025,40(5):1427-1437.
APA	Wu, Zhi-Yuan.,Sun, Sheng.,He, Tian-Liu.,Wang, Yu-Wei.,Liu, Min.,...&Jiang, Xue-Feng.(2025).Federated Class-Incremental Learning with New-Class Augmented Self-Distillation.JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY,40(5),1427-1437.
MLA	Wu, Zhi-Yuan,et al."Federated Class-Incremental Learning with New-Class Augmented Self-Distillation".JOURNAL OF COMPUTER SCIENCE AND TECHNOLOGY 40.5(2025):1427-1437.