Affine subspace constraints improve federated gradient inversion
The authors report better reconstruction than chosen baselines, including when batch size is moderately overestimated.
Stories per day · 30 days
Relatedmembership-inference2differential-privacy1dnn1federated-learning1gradient-inversion1on-policy1qnn1quantization1self-distillation1
4 stories, newest first · one card per story
Stories tagged #model-privacy
4-8 bit quantized DNNs stay exposed to membership inference
Eight attacks across six quantization methods find privacy gains mainly at 2 bits, with large utility loss.
OP-Print targets membership inference in on-policy self-distillation
Gray-box and black-box signals use sparse correction imprints, while the authors say baseline attacks stay near chance.
EIFL: Efficiently Protecting Global Model Privacy and Integrity Against an Untrusted Server in Federated Learning
EIFL lets federated-learning clients keep a global model private and verify it against an untrusted server.