Nouveau chat
Historique de recherche
Recherche académiqueRecherche d'articlesBibliothèqueDiscussions récentes
Fast Federated Machine Unlearning with Nonlinear Functional Theory
Fast Federated Machine Unlearning with Nonlinear Functional Theory
Tianshi Che,Yang Zhou,5 Auteurs,Jun Huan
2023 · DBLP: conf/icml/CheZZLL0DH23
International Conference on Machine Learning · 68 citations
TLDR
A fast FMU algorithm, FFMU, is proposed for improving the FMU efficiency while maintaining the unlearning quality and is proposed to employ nonlinear functional analysis techniques to re-employ the local MU models as output functions of a Nemytskii operator.