Privacy-preserving knowledge transfer for intrusion detection with federated deep autoencoding gaussian mixture model

caused by the significant differences in the individual clients’ data distributions. Extensive experiments demonstrate the superiority of the proposed F-DAGMM.