Equivariant Differentially Private Deep Learning: Why DP-SGD Needs Sparser Models.

Florian A. Hlzl, Daniel Rueckert, Georgios Kaissis

Browse the full CCS paper archive.